Loading...
HomeMy WebLinkAbout10-22-2002 Communication MEMORANDUM POLICE CITIZENS REVIEW BOARD A Board of the City of Iowa City 410 East Washington Street Iowa City IA 52240-1826 (319)356-5041 DATE: September 19, 2002 TO: PCRB FROM: Marian K. Karr, City Clerk RE: PCRB Video The City is exploring a video library within our City website. In the near future you may be able to access the PCRB video on the intemet using the City website, icgov.org DRAFT #1 POLICE CITIZENS REVIEW BOARD A Board of the City of Iowa City 410 East Washington Street Iowa City IA 52240-1826 (319)356-5041 October 8, 2002 Service Organization Contact Address City, State Zip Dear The Police Citizens Review Board now has an informational video available to interested persons or organizations. The primary focus of the video is to inform and engage the citizens of Iowa City regarding the origins, role, and function of the PCRB and the process by which complaints against the police are reviewed. The video is approximately 11 minutes long. There are two copies of the video available at the Iowa City Public Library for viewing or they can also be checked out for use in presentations to community and neighborhood groups, service clubs, City boards, commissions and employees. The video can be found on the New NF Video Shelf or someone from the Information Desk can assist you. Copies of the video may also be purchased through the City Clerk's office for $13.OO. Sincerely, John Stratton Chair, Police Citizens Review Board DRAFT Altrusa Club of Iowa City District 7 Ann Smothers, 2050 Prairie du Chien Road, NE, lC 52240 American Legion Walter Johnson Post 721 Dorothy M¢Cabe, 1015 - 8th Street, Coraville, 52241 Johnson County Bar Association Lois Cox, 112 South Dodge, lC 52240 Business and Professional Women Malinda Allen, 2756 Lake View Drive NE, Solon 52333 Business Women's Association Golden Hawks Chapter Linda Tompkins, 314 East Court Street Place, lC 52240 Business Women's Association Grant Wood Chapter Michele Branstatter, 1483 High Country Road, Coralville 52241 Downtown Association Lisa Barnes, P. O. Box 64, lC 52244 Fraternal Order of Eagles 695 Howard Cook, 2128 South Riverside Drive, lC 52246 Order of Eastern Star Jessamine Chapter Doris Thompson, 170 Paddock Circle, lC Elks Lodge # 590 Al Williamson, 637 Foster Road, lC 52245 Jaycees Jay Honohan, 2122 J Street, lC 52240 League of Women Voters Box 2251, lC 52244 Kiwanis, Golden K Jack Yanaugh, 110 Montrose, lC 52245 Kiwanis, Old Capitol Lee McCormick, 1725 E Street, lC 52240 DRAFT Kiwanis, Noon Al Seagren, 2949 Washington Street, lC 52240 Lions, Evening Dave Smith, 710 Brookside Drive, lC Lions, Noon Judy Terry, 3575 Hanks Drive, lC 52240 Lions, UIHC Mary Jo Piper, 1703 - 21st Avenue Place, Coraville 52241 Masons, # 4, AF & AM Robert Hibbs, 606 Reno Street, lC 52245 Masons, York Rite Robert Woodburn, 3109 Court Street, lC 52240 Loyal Order of Moose Mark Stimmel, 3151 Highway 6 East, lC 52240 Independent Order of Odd Fellows Eureka Lodge # 44 Jim Bigelow, 2427 Petsel Place, lC 52246 Veterans of Foreign Wars Van Eyck Post Sheryl Cook, 2128 South Riverside Drive, lC 52246 Optimists, Morning Club of Coralville Dave Wilderson, 950 Applewood Court, # 2, Coralville 52241 Optimists, Noon Club of Coralville Jeff Rubel, 1040 - 20th Avenue, Coralville 52241 Optimists, Noon Club of Iowa City Scott Means, 19 Ravencrest, lC 52246 Optimists, Sunrise Club of Iowa City Bill Hubbard, 1445 Oaklawn, lC 52246 Optimists Club of North Liberty Darlene Smith, P. O. Box 106, North Liberty 52317-0001 PEO, Chapter JF Janet Power, 1012 Estron Street, lC 52246 PEO, Chapter KP DRAFT Anne Tanner, 427 Elmridge Avenue, lC 52240 PEO, Chapter OD Ann H. Weber, 1731 Red Oak Drive, Coralville 52241 Pilot Club of Iowa City Betty Ketchum, 2929 Cornell Avenue, lC 52245 Noon Pilot Club Ann Weber, 1731 Red Oak Drive, Coralville 52241 Rotary, Noon of Iowa City K. Hughes, P. O. Box 684, lC 52244 Rotary, AM of Iowa City Elaine Shalla, P. O. Box 3166, lC 52244 Sertoma Club, Old Capitol Chuck Lindemann, 4306 Dane Road SW, lC 52246 Sertoma Club, U of I Maralee Dyson, 837 Kirkwood Avenue, lC 52240 Shrine Club John O. Cornelius, 1272 Oakes Drive, lC 52245 1 October 2002 TO: Members of the PCRB ! FROM: Loren Horton/~.~.~ RE: Police Traffic Stop Data Study As you know, I will missthe meeting scheduled for October 22, at which you probably will discuss the traffic stop study. Therefore I wanted you to be aware of my thoughts about some of the factors involved. You are aware that I attended the work session of the City Council at which the study was presented. There were 3 articles in the Iowa City Press- Citizen about the study. We might keep in mind that the study was done by and for the Police Department, and neither the Police Department nor the City Council have referred it to us for comment. In the study itself, I call yoL~attention particularly to four items. On page 10 it states that city census figures should not be ~ed as a baseline for comparison because of the number of commuters, visitors, and students here. On page 20 it lists factors that might be present during a traffic stop but are not measured (such as demeanor of driver, anything about passengers, weather, what is going on in town at the time, location, etc.). On page 24 it notes that the raciat distribution of stops does not necessarily have anything to do with the racial distribution in the general population. ~ will have more to say about that later. On page 31 it states that census data does not provide an appropriate point of comparison for traffic stops. Present at the work session and at the City Counci~ meeting the following night was a person who had much to say about the report. He tried to speak at length at the work session, but was not allowed to do so. He did speak for his allotted 5 minutes at the regular council meeting. One of his points at the work session was that the study was flawed because it did not test the data from traffic stops against the drivers who were not stopped. I seriously doubt if any one can know who was not stopped nor how many were not stopped. We have to work with data from drivers who are stopped. Who can know how many other drivers were on the streets that day at that particular time ? This study included data from April 1 to December 31. I believe that any comment should be delayed until a full year's data is available for study, whether that be April 1 to March 30, or January 1 to December 31. The census statistics for Iowa City and Johnson County are not particularly helpful in evaluating the data from traffic stops, UNLESS we know whether or not the stopped driver actually is a resident of Iowa City or Johnson County. Otherwise the number of males and females, the races of the drivers, etc. who are stopped are not comparable to percentages of such categories in the general population. CENSUS FIGURES COUNT EVERYONE (at least theoretically they do), WHETHER O~NOTTHEY ARE DRIVERS. A SIGNIFICANT NUMBER OF PEOPLE COUNTED IN THE CENSUS ARE TOO YOUNG TO DRIVE. IOWA CITY POLICE DEPARTMENT ~l~ 410 EAST WASHINGTON STREET, IOWA CITY, IA 52240 (319) 356-5275 ! FAX # (319) 356-5449 "An Accredited Police Department" Date: August 7, 2002 To: City Council From: RJ Winkelhake Ref: Police Traffic Stop Data Study The University of Louisville will present the results of the study of the Traffic stop Practices of the Iowa City Police Department at the work session of the City Council on the 19th of August 2002. A copy of the report is in the Council packet. Traffic Stop Department: Research Team Terry D. Edwards, J.D. Elizabeth L. G-rossi, Ph.D. Gennaro F. Vito, Ph.D. Angela D. West, Ph.D. University of Louisville Department of Justice Administration Brigman Hall, 2nd Floor Louisville, KY 40292 (502) 852-6567 June 13, 2002 *This report is confidential and is intended for the lowa City Police Department to use as it deems necessary. It is not to be distributed, quoted, or cited without the express written consent oftbe authors, of Chief R.J. Winkelhake, or others that the ICPD may designate. Executive Summary, This report summarizes the f'mdings of a study conducted using data collected by the Iowa City Police Department between April 1, 2001 and December 31, 2001. These data resulted from 9,702 interactions between law enforcement officers and citizens during traffic-reluted contacts. Information was collected about the driver, the officer, and the stop event. Driver demographics included race, sex, age, residency, and vehicle registration. The only information collected about the officer was officer badge number. Finally, data collected about the stop event include the date, time of day, "reason for stop," "search," "propemy seized," "force," and "outcome of the stop." Data analysis was conducted with the aid of SPSS-I 1.0 (Statistical Package for the Social Sciences). Analyses were conducted on two levels. First, descriptive analysis, using percentages, summarized stop patterns, stop characteristics, and driver demographics. This information is useful only to describe the existing state of affairs ("what is"), but not to explain them ("why") or to formulate predictions about future events ("what it"). To address the complex relationships that exist among different variables, a program called "chi-square automatic interaction detector" or CHAID was used to evaluate the variables in terms of their relatiopships with one another (mukivariate analysis). The greatest percentage of stops was made in the month of April (15%), with the fewest in June (9%). Interestingly, 41% of stops occurred between midnight and 3am, with the third shift (I lpm-7am) responsible for the greatest percentage (54%). Stopped drivers were mostly White (84%), male (65%), young (median age of 23), Iowa City residunts (62%), with Iowa vehicle registrations (86.5%). Drivers were mainly stopped for moving violations (69%), were not searched (95%), and were released w'rth a warning (58%). The descriptive analysis indicated some slight percentage differences among the races in certain events (e.g., stopped for equipment/registration violations). These percentage differences, however, cannot be used to infer correlation or causation ("racial profiling"). To make these types of inferences, multivariate analyses using CHAID were conducted. CHAID segments the sample of traftic stops and reveals the interrelationship between the potential predictors and the events involved in the stop. The CHAff) procedure generates a "decision tree" that identifies significaut predictors of each decision in question. In effect, the procedure "cross-references" each event with each potential predictor. Results from CHAID analyses resulted in only three events (moving violation, being warned, being cited) with significant predictors. Being stopped for a moving violation was significantly related to the age of the driver;, the youngest and oldest drivers were most likely to be stopped for this reason. Warned drivers were those least likely to have been searched, and cited drivers were those least likely to have been stopped for an equipment/registration violation. Race of the driver never appeared as an independent predictor of any event. These data provide no empirical evidence that the ICPD is systematically engaging in discrimirmtory stop practices. Stops conducted by the Iowa City Police Department, as a whole, during the study period, do not involve the race of the driver as a significant factor related to events and outcomes. This does not mean, however, that no individual citizen ever experienced discrimination. It is always possible that individual officers may engage in racially biased practices, both in determining which drivers they will or will not stop and in determining what steps to take after the initial contact. To detect discriminatory practices at this level, however, requires constant vigilance by the community, by all tho officers within the department, and by the departmental adininistration. Statistical analysis, while valuable, cannot substitute for community involvement and effective management. The full report notes some minor problems with the data entry process, provides a discussion of the "baseline dilemma," makes recommendations for continued study to obtain a full year of"clean" dam, and suggests modifications of the data collection instrument to include more variables (e.g., warrant check information). Table Of Contents Introduction ...................................................................................................... 1 - 3 Methods ..................................................................................................................... 3 - 7 Data Collection .............................................................................................. 3 Variables ........................................................................................................... 4 Collection and Measurement Concerns ............................................................. 5 Analyses and Results ......................................................................................... 6 Descriptive Analyses and Results ............................................................................ 7 - 16 Driver Demographics .................................................................. 7 Stop Event ............................................................................... 10 Summm3~ of Descriptive Analyses ................................................................. 16 CI-IAII~ Analyses and Results ................................................................................... 16 - 23 CHAID Results ............................................................................................. 20 Reason for Stop ................................................................. 20 Outcome ......................................................................... 22 Summary o£ CHAID Analyses ............ i .............................................................. 23 The "Baseline" Dilemma ............................................................................................. 23 Legal Issues Relating to Bias~Racial Profiling Data Collection and Analysis ........ 26 - 29 Overview ................................................................................. 26 Civil Liability ............................................................................ 26 Disclosure of Information/Records ................................................... 28 Conclusion ................................................................................ 29 Conclusion and Recommendations ............................................................................. 30 - 33 Bibliography ...................................................................................... 34 Appendices ....................................................................................... 35 Appendix A: Iowa City Police Department Policy on Racial Profiling Appendix B: Iowa City Police Contact Sheet Introduction Racial Profiling Accusations of discriminatory traffic stop practices ("racial profiling") have emerged as a critical issue facing law enforcement. According to a 1999 Gallup poll and research conducted by the American Institute of Public Opinion (2000), many believe that racial profiling is widespread and disapprove of the practice of stopping motorists simply because the driver fits a partic~dar profile (Newport, 1999). In response to this growing concern regarding traffic stops and a more general distrust ofiaw enforcement personnel, many police departments across the U.S. have begun to more closely examine their traffic stop policies and procedures. Further, some police departments have begun collecting traffic stop data The collection of traffic stop data initially may appear to be a rather straightforward process. In reality, however, the collection and analysis of traffic stop data is far from simplistic. A number of concerns must be addressed by any agency comemplating such an endeavor. These concerns range fi-om defining the issues, developing data collection iastn~ments and procedures, training personnel to collect data, and detcmfining the most appropriate means to analyze the data. Defining Racial Profiling The precise definition of racial profiling is a matter of debate. While no universal definition exists, racial profiling is generally regarded as any act by law enforcement, whether it involves motorists or pedestrian~, based solely on the race of the alleged violator (Ramirez, McDevitt & Farrell, 2000). In expanding on this broad definition, the U.S. Department of Justice considers racial profiling to be "any police action that relies upon the race, ethnieity or national origin of an individual rather than behavior of that individual that leads the police to a particular individual who has been engaged in or having been engaged in criminal activity" (Ramirez, McDevitt & Farrell, 2000). Accordingly, police may use race and ethnicity to determine if an individual matches a suspect description bat police may not use stereotypes when deciding who to stop, to search, or make subject to other stop - related actions. Further, as Withrow (2002) notes, profiling by police can be further defined based on specific factors used in profiling. MacDonald (2000) suggests that profiling can be considered hard or soft. Hard profiling occurs when race is the one and only factor used in police decisions to stop a particular motorist. Soft profiling occurs when race is one of several factors the police use in determining whom they stop. For this report, the Iowa City Police Department defines racial profiling as "the detention, interdiction, exercise of discretion or use of authority against any person on the basis of their racial or ethnic status or characteristics" (Racial Profiling, General Order 01-01). A copy of this policy is contained in Appendix A. Collecting Data Many departments have, independently or in collaboration with others, undertaken the task of analyzing traffic stop data. These agencies vary in te~ms of their structure and function, as well as in the type of data they collect. In addition, some data collection efforts involve sophisticated data analyses where others simply compare basic percentages. These differences, onthe surface, are not all that dramatic. When making conclusions about the practices ora department, however, these methodological considerations take on more importance. In fact, methodological considerations are considered paramount by prevailing judicial opinions (see following discussion on legal issues). It should be noted; however, just as there are no widely accepted standards for defining racial profiling, the methods of collecting and analyzing traffic stop data are not universal. 2 Police departments across the country collect a variety of data elements in their analysis of racial profiling. Some agencies collect a minimal amount of data such as the race, age, and gender oftbe driver, along with the reason for, and outcome of the traffic stop. Other agencies collect data pertaining to all passengers of the vehicle, key events that may occur during a traffic stop (e.g. warrant check, search), and police officer demographics. There appears to be no consensus regarding the most appropriate data collection elements across departments. The National Institute of Justice (NIJ), however, recommends certain data be collected on a "routine" basis (Ramirez, McDevitt & Farrell, 2000). These data elements include: date, time, and location of stop, license number and description of vehicle, length of stop, and name and idemification number of the officer initiating the stop. The NIJ also recommends that certain "study specific" variables be considered. These include the race, date of birth and sex of the driver, the reason for stop; the outcome of the stop, and whether or not a search was conducted. Methods Data Collection Data were collected about each traffic stop (lq = 9,702) made by officers of the Iowa City Police Department over the nine-month period between April and December, 2001. Officers were required to enter data into mobile data terminals (MDTs) after each traffic stop interactiom A copy of this form is contained in Appendix B. When an officer would initiate a traffic stop, he or she would call that stop into the dispatcher, who would document the contact. At~er the stop, the officer would fill out a screen on the MDT located in the vehicle. These data were centrally stored in a Microsot~ Excel spreadsheet. Each stop became a case for analysis. The Excel fde was subsequently transferred into SPSS for analysis. 3 Variables Information was collected about the driver, the officer, and the stop event. Driver demographics included race, sex, age, residency, and vehicle registration. The only information collected about the officer was officer badge number. More data about the officer, such as sex, race, age, time in service, etc., can he entered at a later date. Several items of interest pertain to the stop event, including the date and the time of day. One broad category related to the stop event involved the "reason for the stop." These were coded dichotomously (yes/no) and included the following: moving violation, equipment/registration violation, criminal offense, other violation, call for servicedsuspeet or vehicle description, pre-existing knowledge or information, special detail, and other. Information regarding any "search" that might have been requested or conducted also was collected and included the following dichotomous variables: consent search requested, consent search of vehicle requested, consent search of person requested, consent search conducted, officer safety search conducted, search incident to arrest conducted, and probable cause search conducted. Data pertaining to any "property seized" also was collected and included the following dichotomous variables: property seized, alcohol seized, weapons seized, money seized, narcotics seized, evidence seized, other seized. The "outcome of the stop" also was measured, and included the following dichotomous variables: no action, citation, arrest, warning, and field interview. Finally, information also was collected aborn whether any "force" was used during the stop and whether the force was against the driver or a passenger. 4 Originally, 38 variables were measured and entered for analysis. Some of these were recoded for analysis. For example, driver race was collected in 7 categories (Caucasian, Black, Asian, Spanish, Native American, Other, and Unknown). These were collapsed into 3 categories for the analyses (White, Black, Other). In addition, some new variables were created to obtain a clearer picture of the data. For example, it is logical to as,~me that a person stopped for multiple reasons might be more likely than a person stopped for only one reason to get a citation or to be arrested. The dichotomous variable "multiple reasons" was created by distinguishing between cases with "only one reason" for the stop and cases with "more than one reason." Collection and Measurement Concerns Glitches arise at the initial stages of any large data collection undertaking. This study was no exception~ Although the ICPD engaged in a series of training sessions to familiarize officers with the data collection form, the use of the MDT, and the procedure, difficulties and oversights still occurred. This most problematic of these became apparent when Excel entries (from the MDTs) were cross-referenced with CAD entries. A discrepancy was noted between the number of stops as indicated by the CAD system and the number of stops as indicated by the Excel entries. In addition, stops were referenced by officer badge number. A routine check of CAD and MDT entries appeared that some officers were calling stops into the CAD system and not entering them into the MDT, or entering them into the MDT without calling them into the CAD. ARer a series of inquiries and discussions with the officers, it was obvious that the problem was related to training and/or to data entry. For example, in situations where 2 officers were in the car during a traffic stop, one officer may call in the stop to the dispatcher, who enters the stop under that officers badge number. The second officer in the car may take responsibility for entering the data into the MDT, which then gets entered under his or her badge number. In this way, it may appear as ff an officer is either under- or over-reporting on the MDT system. In addition, some difficultias were noted with officers forgetting to "save" the data into the MDT after they had entered it, resulting in some lost information, These problems were quieldy caught and corrected, but it is recommended that any conclusions drawn f~om this data keep these difficulties in mind. A full year of "glitch-flee" data collection is recommended for use as a baseline for this department. Moreover, any close inspection of stopping behavior by individual officers should not be undertaken until a full year of corrected data collection is completed. Analyses & Results Analyses were conducted on two levels. First, descriptive analysis takes a broad look at stop patterns, stop characteristics, and driver demographics. This information is useful for descriptive purposes only. That is, this type of analysis is useful in understanding the existing state of affairs ("what is"). Descriptive analyses are neither predictive nor explanatory. They cannot explain "why" things are as they are and they cannot predict hew things might be in the future. Comparisons using descriptive analyses also are problematic given that descriptive statistics do not consider relationships ~mong different variables involved in any given situation. To address the complex relationships that exist among different variables, multivariate analyses also were conducted. Specifically, a program called "chi-square automatic interaction detector" or CHAID, was used to evaluate the variables in terms of their relationships with one another. For example, this type of analysis is able to determine whether the sex of a driver is related to the reason for stop, given all the other variables that might interact, such as race or age. A more detailed explanation of this process is contained below. Descriptive Analyses & Results Driver Demographics The variables related to the driver involved in the stop were the following: race, sex, age, residency, and vehicle registration. Drivers stopped were mostly White (84%), male (65%), younger, Iowa City residents (62%), and ~om Iowa (86.5%). Race. As indicated below, 84% of the drivers stopped were White, 9% were Black, and 7% were Other. The "Other" category includes Asian, Hispanic, Native American, and Other. There were 13 cases in which the race of the driver was coded as "unknown" and these were counted as "missing" in the analyses (See Table 1). Table 1: Percentage of Stops by Race of Driver 90O/o./ 80o/o./ ~ 70o/o./ 60%. / 50%. / 400/o./ 30%, / 20%, / 10°/° ~ ~ White Black Other 7 Sex: Sex by Race. Most (65%) stopped drivers were male (See Table 2). A higher percentage of Non-White males than White males were stopped, whereas a higher percentage of White females than Non-White females were stopped (See Table 3). Table 2: Percentage of Stops by Sex of Driver 50% / Female Hale Table 3: Percemage of Stops by Sex and Race of Driver 80O/o` 70o/o 60% [] White i1 Black 300/o ilOther 20% 10% 0% Female Hale Age: Age by Race. The median age of drivers stopped was 23, with most stopped drivers being 21. In fact, more than 7 in 10 drivers stopped were under the age of 30 (See Table 4). Higher percentages of Non-White drivers than White drivers between the ages of 25 and 44 were stopped. In general, younger (24 & under) and older (45 and over) White drivers than Non- White drivers were more likely to he stopped (See Table 5). Table 4: Age Categories of Drivers Stopped 50%. 44% 4o( 300/0- 23% 20%- ~3% 10%- 4% 0%. Under 18 18-20 21-30 31-40 Over 40 Table 5: Percentage of Drivers Stopped by Race and Age 60%, 50% 40% 30% ~White · Black 200/° · Other 10O/o 0% Under 18-20 21-30 31-40 Over 40 18 9 Residency and Vehicle Registration. Of all drivers stopped, less than two-thirds (62%) were city residents (See Table 6). Another 11% were Johnson County residents. An equal percentage either was from other Iowa locations or from out of state (13.5% respectively). This is probably characteristic of a city with commuters and a large college campus. Given this, city census figures should NOT be used as a baseline for comparison to the overall stop data. Table 6: Percentage of Drivers Stopped by Residency/Registration 70%, 62.0% 60% 50O/o. 40% 13.5% 10o/ 0% Towa City Johnson Other Iowa Non-Iowa County Stop Event Temporal Distribution. The most active month for stops was April (15%), followed by May (13%), and November (12%). June (9%) was the least active month (See Table 7). The time distribution of stops was unusual, with 41% of all stops occurring between midnight and 3:00 am. In fact, the most active time was between 1-2am (17%) (See Table 8). This is likely due to the fact that Iowa City is a college town with a high concentration of bars and restaurants that close around that general time. Drivers probably get stopped as they are leaving a bar or restaurant after closing time. Given this time distribution, it is no surprise that the third shift is responsible for the highest percentage of stops (54%), followed by the second shift (25%), and the first (21%) (See Table 9). Table 7: Percent of Stops bT' Month 20% 18°/o 13% 16°/o' 13% 14°/o txwo 11% 12% 12% Oo/o ~0O/o ~0~o 10% 8% §O/o~ 4% O% Table $: Percent of Stops by Hour of Day 11 Table 9: Percent of Stops by Shif~ 60%- 54% 40%- 30%- 21o/o 200/0. 10%. 0%' 1st Shift 2nd Shift 3rd Shift In general, ~ivcrs wcrc stopped for mo¥i~g ¥io]atious (69%) or equipment/registration violatious (26%), were not sc.ched (95%), ~md wcrc relcasod with a wm*nJng (56%). Only cases ~volYed usc of £o~cc, so tls Yariablc w~ not used i~ any ~malyses. LAcwisc, only ]4'7 (!.5%) c.cs i~vo]vcd a~y type ofp~opc.-y seizure (m~'m/y ~cotics) so this variable is couskle~ed further. Only 5% of the c~cs L~vol~¢d a sc~c~ ~d t~csc wc~c ~stly (7~%) ~c~dcm to ~c~o~ [o~ ~top b~ ~. TAc ~cc most-c~tc~ ~c~o~ ~or stops wc~c 1) viol~fio~ (~9%); 2) cq~pmc~cg~fio, v~o]~fio~ (2~%); ~d 3) otbc~ v~oAfio~ (~%). ~tops of %th~' ~iv~s (7 ]%) wc~e more likely ~n stops of white (69%) or black ~i~s (63%) to ~olv~ a mov~g v~o]afion. Twc~-s~ p~ccnt (26%) o[all stops w~c ~or ~uymc~Vregis~afio~ violation; stops of black ~vcrs (3 ] %) w~c more ]~]y t~ stops of white (25%) or otb~ ~iv~s (~4%) to ~vo[vc this ~son. ~ violations (6%) ~volvcd white ~iv~s (6%) mo~c o~c. ~ black (~%) or otb~ (~%) (g~ T.~le 10), 12 Table 10: Percentage of Stops by Reason and Race of Driver 80%' 70%' 60%* 400/0.50%' I. TOT..AL I[] Whtte 30%. 20%' *1 · Black :1.0%. I 0%. [mOther Searches by Race. Most searches (75%) were conducted incident to arrest; "Other" drivers (85%) and White drivers (77%) were more likely than Black drivers (64%) to be searched for this reason. Consent to search was given in 23% of cases, more so by Black drivers (28%) than by White (23%) or "Other" drivers (8%) (See Table l 1). There were 359 searches incident to arrest (< 4% of all stops) out of 479 total cases in which a search was conducted (75%). Of all drivers stopped, 3% of White drivers, 7% of Black drivers, and 3% of"Other" drivers were searched incident to arrest. Of all the drivers searched for this reason, 79% were White, 15% were Black, and 6% were "Other" (See Table 12). Out of all drivers stopped, there were 83 consent searches (less than 10% of all cases): 2.4% of all Black drivers were involved in this event, compared to .7% of all White drivers, and .4% of all "other" drivers. Of the consent searches (n = 83), white drivers comprised 72%, Blacks were 24%, and other drivers were 4% (See Table 13). 13 Table 11: Percentage of Searches (n = 479) by Typo and Race of Driver 60%* [] White 0%. · Black e~'~ ·Other Table 12: Percentage of Searches Incident to Arrest (n = 359) by Race of Driver 90%. / 70% / 60% '" 500/0/ 40O/o/ 30% / WhKe Black ~er 14 Table 13: Percentage of Consent Searches (n ~ 83) by Race of Driver Outcome of Stop by Race. Other drivers (65%) and black drivers (61%) were more likely than white drivers (57%) to be issued a warning as a stop outcome (See Table 14). White drivers (41%) were more likely than black (37%) or other drivers (34%) to be issued citations. A much higher percentage of black drivers (13%), however, had arrest as the outcome of their stop. Only 7% of white drivers and 6% of other drivers were arrested. These percentages do not include the 431 cases in which the outcome was "no action." Table 14: Percentage of Drivers Stopped By Outcome and Race 70%' 60%' 50%' 40%' · TOTAL 30%' DWh~e 20%' · Black 10%. · Other 0o/o, 15 Summary of Descriptive Analyses At first glance, one might be tempted to conclude that race is a factor in some evems. For example, higher percentages of"Other" drivers were stopped for moving violations while higher percentages of Black drivers were stopped for equipmentYregistration violations. Similarly, the sox and age of the driver also appear to be factors given that higher percentages of White females were stopped, as were higher percentages of Non-White males. Descriptive statistics are very superficial and only give the broadest picture of the d_ata: This type of analysis lacks inferential ability. One cannot use it to predict events or to describe the relationships among characteristics and events. Descriptive statistics only should be used to describe the state of affairs. They will not help to: 1) understand why the percentages are the way they are; 2) determine the relationships among the characteristics and events; 3) predict one outcome or event over some other outcome or event. Providing a description of the data should only be the first step in a thorough analysis. More comprehensive multivariate analysis is required to understand the relationships between and among variables, and to understand how theso variables interact with one another to produce a certain reality, as por~ayed by the descriptive statistics. In this case, a procedure called chi- square-automatic interaction detector (CHAID) was used to more fully explore the relationships between and among the various variables. CItAID Analysis & Results This portion of the report examines the relationship between three demographic predictors (age, race, sex), vehicle registration (Iowa/non-lowa) and soveral events related to the traffic stop. These events involve the following questions: 16 1) Reason for the Stop? (moving violation, equipment/registration violation, pre-existing knowledge, other violation, crime, special detail, other). 2) 'Search Conducted? (vehicle search or driver search). 3) Type of Search? (search incident to arrest or consent search). 4) Property Seized? 5) Outcome (warning, citation, or arrest). Some of these decision points also were examined as predictors of subsequent events. For example, whether property was seized might be related to whether a driver was warned, cited, or CHAID is based on an analytical technique called chi-square. Chi-square analysis demonstrates whether a particular observed proportion within a sample is statistically different from a particular expected proportion within that sample. The expected proportion is based on the premise that there is no relationship (i.e., one has no impact on the other) between the two variables in question within the population from which the sample under study was drawn. It is calculated using information fi'om the entire group. For example, if we were interested in whether race (White, Other) and being arrested (Yes, No) are related in a population, we would use chi-square analysis. The chi-square procedure would determine that 25% of all the persons (regardless of race) were arrested and 75% were not. Then, the chi-square procedure would determine that, of all the "Other" drivers, 30% were arrested. Chi-square analysis would then conclude whether the 5% difference between all persons arrested and "Other" persons arrested is attributable to chance, or whether it is likely that there is a tree difference in the population between White and Other drivers in being arrested. Iftbe chi-square value is "statistically significant," this 5% difference is not 17 attributable to chance and represents a true difference I~ween White and Other drivers in being arrested. By convention, statistical significance is reached when the probability of error in this conclusion is less than .05 (Le., only 5 times out of 100 would one reach this conclusion in error). Race, however, is just one factor that could be related to any event in a stop situation. Other variables may be more important. They may mediate, or even eliminate the influence of race. This is why we use a "measure of association" called the "phi coefficient" with chi-square analysis. The phi coefficient ranges in value bom 0 (no relationship) to 1 (,perfect or very strong relationship). If chi-square analysis indicates statistical significance (that the 2 variables are related), it is then necessary to detem,ine the strength of that relationship. In the previous example, the 5% difference in the proportion of Black drivers arrested and the proportion of ali drivers arrested was statistically significant. The question now relates to how slrong the relationship is between race and arrest. The chi-square analysis determines the phi coefficient for this relationship to be .03. This indicates an extremely weak, almost non-existent relationship between race and arrest because .03 is much closer to 0 than to 1. In fact, this means that very little variation in arrest is explained by the race of the driver. Another variable or set of variables is more influential in arrest than the race of the driver. This is where it becomes necessary to conduct multivariate analysis. CHAID is a multivariate technique that segments the sample of traffic stops and reveals the interrelationship between the potent'mi predictors and the events involved in the stop. The CHAID procedure generates a "decision tree" that idemifies significant predictors or each decision in question. In effect, the procedure "cross-references" each event with each potential predictor. 18 CHAID simultaneously considers the impact of several independent variables (age, race, sex) upon a particular event in question (arrest, in the example above). The CHAID results indicate the strongest predictor of the event, while taking the other variables into account. It may he that no variable or set of variables is a predictor of the event when the other variables are considered. This means that any original relationship (e.g., between race and arrest) is so weak that when other independent variables are considered (age, sex), nothing predicts the event. In the arrest example, the program examines all the cases in which individuals were arrested. It then examines all the factors associated with each case and determines the ones that keep oecm't'iag in conjunction with an arrest. Then, the program compares that state of affairs with the cases in which drivers were NOT arrested. In this way, it is possible to determine whether factors are really predictive of an event or whether observed differences between these arrested and those not arrested occurred purely by chance. For example, if descriptive analysis dctcJ ,~ines that 30% of the drivers arrested were White and 70% were Black, one might he tempted to conclude that there was a racial bias in arrests. However, the CHAID analysis would examine the cases and simultaneously consider all the other potential factors involved in an arrest. The decision tree that it generates might indicate that the most significant factor related to arrest is a stop for "pro-existing knowledge." The analyses demonstrate which of the potential predictors (if any) had the strongest and most important reiatiouship to the events or outcomes. In this case, the potential predictors were used to examine the five events listed above to determine if they were actually related or whether any observed differences occurred purely by chance. The advantage of multivariate analysis is that it reveals the strongest predictors of the event in question. In other words, if race is a factor, it will emerge independently of the other 19 factors. If race is not a factor, then the one or more of the other predictors will emerge, or none of the selected predictors will emerge as related to the events/outcomes. If no significant predictor emerges, it either means that the analyses did not include the most relevant predictors or that no measured factor is related to the event. This attribute is particularly relevant for a traffic stop situation in which many things go unmeasured. For example, one cannot measure the quality of the personal interactions between an officer and the individuals stopped. One cannot measure the demeanor of the driver. In this case, one cannot measure any information about the passengers in the vehicle. Finally, extraneous factors such as the weather, the time of year, the social environmem, and the location are not measured in this study. CHAID Results Results fi.om CHAID analyses using the 5 event categories and the potential predictors outlined above resulted in only tkme events that had significant predictors. Within "Reason for Stop," being stopped for a moving violation was si~ifieantly related to one or more of the potential predictors. Having a search conducted (vehicle search or driver search), type of search (search incident to arrest or consent search), and having property seized had no significant predictors. Two outcomes (warning and citation), however, did have significant predictors. Arrest, on the other hand, did not. Reason for Stop Reasons for stop included the following: 1) moving violation; 2) equipment/registration violation; 3) other violation; 4) pre-existing knowledge; 5) criminal offense; 6) special detail; and 7) other. A stop for a moving violation was the only variable that had significantly related predictors. 20 Moving_ Violation. For the entire group, 68.6% (6656/9702) of the drivers were stopped for a moving violation. This is the base rate for moving violations. The question is whether any subgroups formed based on the potential predictors had moving violation rates significantly different ~om (greater than or less than) that of the entire group. The most significant predictor of a moving violation was the "age" of the driver. Second- order predictors were "sex of driver" and "vehicle registration." Sex (being female) was a predictor for the 10-17 age group. Vehicle registration (non-Iowa) was a predictor for the 18-20, 21-30, and over 40 age groups. Finally, "race" and "sox" emerged as third-order predictors. Race (being an "other person of color") was a relevant factor among the 18-20 and 21-30 year old Iowa residents. Sex (being female) was relevant among the over 40 year old Iowa residents. Apart bom order of significance, the subgroups also can be described in t~alS of their proportion, remembering that 68.6% is the base to which comparisons are made. The subgroups receiving moving violations, in order of proportion are: 1. Persons over 40 with non-Iowa regis~'ation: 84.5% 2. Persons 10-17 who are female: 81.8% 3. Persons over 40 with Iowa registrations who are female: 77.8% 4. Persons 10-17: 75.0% 5. Persons over 40: 74.6% Thus, age (being in the youngest group or being in the oldest group) was the strongest predictor of a moving violation stop. Sex of the driver, registration, and race only emerge in combination with the other variables, not as independent predictors. There is no evidence of racial bias in drivers being stopped for a moving violation~ 21 Out~ome Warning. Nearly two-thirds (65.5%, or 5383/9702) of the entire group was given a warning. The must signifw, ant predictor of being given a warning was whether a search was conducted. Those who were not searched were significantly more likely to receive a warning. The second-order predictor was being stopped for an equipment/registration violation, and the third-order predictor was having a non-Iowa registration. Those most likely to receive warnings were: 1. Persons who were not searched, who were stopped for equipment/registration violations, and who were non-Iowa registrants: 77.1%. 2. Persons who were not searched, who were stopped for equipment/registration violations: 70.2% The strongest predictor of a warning was whether a search was conducted. These finding are logical in that drivers who were searched would he more likely to have been stopped for a more serious violation and would therefore not reCeive a ming. Also, drivers stopped for only having an equipment/registration violation and/or to be from "out of town" might be less likely to he issued citations. Overall, no bias was detected in the issuanCe of warnings. Citation. With thi~ event, the base rote for the entire group was 38.7% (3753/9702). The most significant predictor of a citation was whether the driver was stopped for an equipment/registration violation. Being stopped for something other thall an equipment/registration was the most significant predictor of receiving a citation. The second- order predictor for this group was age (10-17), and the third-order predictor was having an Iowa vehicle registration (among those over 30 years old not stopped for E/R violations). Describing the subgroups in terms of their proportion (compared to the 38.7% base), the subgroups receiving citations are: 22 1. Persons not stopped for equipment/registration violations who were over 30 years old, with Iowa registrations: 50.4% 2. Persons not stopped for equipment/registration violations who were between 10-17 years old: 50.2% Therefore, not having an equipment/registration violation was the strongest predictor of a citation. This is logical given that the other biggest category of stops involved moving violations. This is the type of situation in which an individual is more likely to be issued a citation. Age and registration were related to receiving a citation in combination with equipment/registration violation, not as independent predictors. There is no evidence of racial bias in drivers being issued a citation. Summary of CHAID ~,nalyses Only three events involved significant relationships to tested predictors. Receiving a moving violation (being in the youngest or the oldest age groups), receiving a citation (not being stopped for an equipment/registration violation), and receiving a warning (not being searched) were the only events that CHAID analysis determined to have significant relationships with predictor variables. Sex of the driver and the vehicle registration also were related in conjunction with the significant predictors in some situations, but not as independent predictors of any given event. Race of the driver (being an "other person of color") only appeared once, as a third-order predictor among certain age groups of Iowa residents in being stopped for a moving violation. The "Baseline" Dilemma The most problematic part of any study of this nature is determining the baseline to which collected data should be compared. We want to look at "what is" and compare that state of affairs to "what should be." However, determining ''what should be" is troublesome. In 23 theory, the racial distribution of drivers stopped should represent the racial distribution of drivers doing something that makes them eligible to be stopped. For example, if20% of the drivers doing something that makes them eligible to be stopped by the police are Black and 80% are White, one would expect that 20% of the drivers stopped are Black and 80% are White. This comparison has very little, if anything, to do with any racial distribution in the city or county population. It has everything to do with the racial distribution of drivers on the roadways and the driving behaviors or characteristics that they exhibit. Making decisions as to whether a department is engaging in discriminatory stop practices depends on the ability to identify the racial distribution of stops that would exist in the absence of discriminatory stop practices. That is, one must know the true racial distribution of drivers eligible to be stopped (i.e., doing anything that could get them warned, cited, or arrested anything that creates reasonable suspicion or probable muse). Stops in the absence of discriminatory practices, then, would be the "right" proportions. One could then compare the research findings to the "right" proportions to determine whether discrimination exists. Unfommately, we cannot measure this objective reality. Determining the "right" proportion of stops is impossible bemuse of the infinite variations in driving behaviors and police response within various locations at various times on various days in various months during various years. Also missing is a measure of the interactions between those stopped and the officers. Demeanor is thought to significantly eontn'bute to stop outcome as well as to other law enforcement outcomes such as warning, citation, and arrest. This reality, however, is extremely difficult, if not impossible, to measure. We cannot know the racial distribution of drivers doing something that makes them eligible to be stopped. Some research has attempted to measure this, but the methodology employed is oRen seriously 24 flawed. The most common method involves posting trained observers at strategic locations armed with stopwatches to determine the racial distn30ution of speeders. Obviously, this me~hod is extremely limited, relying on split seeoud judgment by observers as to the race of drivers. In addition, this method rests on the assumption that speeding is the only thing for which drivers get stopped. In the current study, moving violations were the most commonly cited reason for a stop, but equipment/registration violations and other violations aceoumed for about 3 in 10 stops. Given that co~,arison to population data is invalid, we suggest that the current cl~a become the baseline from which to evaluate future practices. The initial analysis ora law enforcement agency's traffic stops does establish a benchmark for that department. Once an initial study is completed, a department has an empirical basis for comparison in the future. If an initial study indicates the possibility of bias (race appears as a significant predictor of some event), future research will provide data for comparison to help determine whether the relationship previously observed between race and some outcome persists or whether it has disappeared. If an initial study shows no evidence of bias (race does not appear as a significant predictor of any outcome), the department in question should attempt to maintain this desirable result. These data, collected from traffic stops made by the Iowa City Police Department between April 1 and December 31,2001, provide no evidence that the ICPD is systematically engaging in discriminatory stop practices. Stops conducted by the Iowa City Police Depastment, as a whole, during the study period, do not involve the race of the driver as a significant factor related to events and outcomes (e.g., arrest, search, etc.). This does not mean, however, that no individual citizen was ever discriminated against. There is always the possibility that individual officers may he engaging in racially biased practices, both in determining which drivers they will 25 or will not stop and in determining what steps to take after the initial contact. This is a serious possibility that is not likely to be revealed with statistical analysis. To detect discriminatory practices at this level requires constant vigilance by the commllllity, by all the officers within the department, and by the departmental administration. Statistical analysis, while valuable, cannot substitute for cemmRtlity involvement and effective management. Legal Issues ReLating to Bias/Racial Profiling Data Collection and Analysis Overview The findings and conclusions of any study involving b'ms/racial profiling are often used, or interpreted, in a number of ways, for a variety of purposes, by ma~ly factions. These stlldies often raise issues related to the management nnd admini.~ration of the agency, issues relating to the recruiting, mining and attitude of the officers, and issues related to the community, just to name a few. This section focuses strictly on the legal issues involved with this, or any, study of bias/racial profiling. Civil Liability Without a doubt, the central legal issue relating to any study of bias/racial profiling by a law enforcement agency is the degree to which the agency, ortbe individual officers employed by the agency, may be subject to civil liability for their actions. While the terms "bias profiling" and "racial profiling" are of relatively recent origin, and neither are legal terms, the practice of bias/racial profiling, if substantiated, allows victims to pursue civil clam against an offending agency, or officer, under a variety of legal theories. Although each legal theory has its own strengths and weaknesses, for a number of reasons, the theory employed by most plaintiffs, and the one that is arguably the most difficuit for plaintiffs to obtain evidence and prove, is that of a Constitutional violation of the 14~ Amendment's Equal Protection Clause. Generally speaking, 26 the standard required for a plaintiffto win in an Equal Protection claim is that the plaintiff must prove that other similarly situated individuals, of a different race, were treated differently. Likewise, provlng~ or disproving, disparity of treatment based on race should also be the focus of any study of bias/racial profiling. Thus, the key importance of any study on bias/racial profiling, from a legal perspective, is that the study's findings and conclusions can become the evidentiary basis for supporting, or defending, such claims, In short, the data, and more importantly the findings and conclusions of the evaluators, of bias/racial profiling studies serve as the statistical evidence used by plaintiffs or defendants to support or defend the legal claims~ Several courts have addressed the issue of civil liability under the 14t~ Amendmem based on a claim of bias/racial profiling and the evidentiary requirements needed to support such a clairrc These courts repeatedly emphasize the need for both plaintiffs and defendants to introduce valid and reliable statistical evidence establishing, or disproving, disparate treatn~nt based on race. Evidence taking the form of statistics based on anecdotal sources, or data evaluated using unacceptable methodology, are universally rejected by the courts. In Chavez v. Illinois State Police, 251 F.3d 612 (7t~ Cir. 2001), atypical Equal Protection lawsuit, the court went to great lengths to outline the validity and reliability standards required of evidence relating to the collection and/or analysis of data regarding bias/racial profiling. The court noted that statistical evidence may be used to establish that other similarly situated individuals, ora different race, were treated differently; however, to be admissible slid of any relevance to the issues before the court, such statistical evidence must be collected and analyzed in a universally scientifically acceptable manner. Further, the court noted that the statistical evidence must be subject to rigorous methodological procedures and evaluated by persons with the academic credentials and practical experience to qualify as experts. The court specifically 27 noted the inherent problems with statistical evidence relating to bias/racial profiling with regard to the following: establishing base lines, determining the quantity and quality of the data being collected, sample groups, and interpretation. Accordingly, if the stat'~ical analysis and findings and conclusions of this, or any, study of biadracial profiling are to be of any value from a legal perspective, the study should comply with the evidentiary requirements currently being imposed by the courts. This study seems to satisfy the admi.~sibility requirements for evidence relating to disparate treatment based on race, currently being imposed by courts in bias/racial profiling cases. This study employed sound methodological techniques with regard to the collection and analysis of data and was performed by individuals with nationally recognized expertise in statistical analysis. Disclosure of Information/Records Although generally not risingto the level ofcoacem as civil liability, law enforcement agencies engaged in the collection of information and analysis of data, whether related to bias profiling or some other topic, must be familiar with the applicable statutes and/or ordinances governing the release of public records. Typically referred to as "Open Records Acts", virtually all jurisdictions have enacted laws requiring certain records in the possession of police agencies to be released to the public. These "Open Records Acts" vary tremendously l~om jurisdiction to jurisdiction; however, in all jurisdictions, to some degree, the data collected as part of a bias profiling project will be subject to disclosure to the public, and to the media. Ideally, agencies will address this legal issue before initiating any data collection to ensure they know, going into the project, what records, if any, will be subject to disclosure, and under what circumstances. 28 The fundamental questions to be resolved relating to the release of data and information collected as part ora bias profiling project are: 1) Who, exactly, is the cnstodian of the data and information relating to the project? [This can become yep/complex in situations where agencies contract all, or part, of the project out to a consultant.] 2) What records are, and are not, subject to disclosure? 3) Can any of the information collected be "masked" or otherwise shielded from disclosure? Must any information be shielded from disclosure? 4) If large data sets are subject to disclosure, what format is required? 5) Where disclosure of large, bulky, data sets is required, what costs, if any, may be recovered by the agency? 6) Is the analysis/interpretation of the data subject to disclosure also? 7) When must data/information be released? [This can pose difficulties in multi-year, on going, projects.] 8) How long must the data/information be retained and who had responsibility for archiving the materials? Conclusion It is imperative that agencies practice proactive risk management with regard to the collection and analysis of data relating to bias/racial profiling. In addition to serving as the basis for addressing a host of management, administration and personnel issues, bias/racial profiling studies can also serve as useful tools for developing statistical evidence for defending against lawsuits alleging civil rights violations. However, experts in statistical analysis must conduct any study using scientifically acceptable methodology. The statistical analyses involved in this study appear to satisfy the legal requirements currently being imposed by the courts and the f'mdings and recommendations should serve as valid evidence relating to allegations of bias/racial profiling. Finally, a determination should be ascertained as to what degree the information/records will be subject to disclosure under the applicable Open Records laws. 29 Conclusion and Recommendation~ TI~ Iowa City Police Departn~nt, as a whole, does not appear to be systematically stopping drivers based on their '~racial or ethnic status or characteristics" as defined by departmental policy (Racial Profiling, General Order 01-01). While the percentages of races were not always equal in some categories, the discrepancies are most likely explained by factors other than the driver's race. For example, the age and sex of the driver were important explanatory factors in many events. This makes sense given that we know driving behavior to be different among various ages and between the sexes; younger drivers drive differently than older drivers and males drive differently than females. This study used a fairly comprehensive set of data collected about a population of stops over an 8-month period. The data were collected in a consistent manner, with only minor problems pertaining to entry and recording that were addressed as they were discovered. The statistical analysis used to evaluate the data was rigorous, thorough, and conducted by academicians with expertise in the coliection, analysis, and interpretation of such data~ Further, this analysis was conducted on a contractual basis with researchers from the University of Louisville in Louisville, Kentucky, providing a level of objectivity that is necessary to avoid any conflicts of interest or appearances of impropriety. These factors have yielded valid data, making valid conclusions highly likely. The only caveat is that one full year's worth of data should be collected and analyzed to provide a baseline from which to evaluate future stop practices. Moreover, the legal considerations set forth by the courts have been met, making legal actions against the Iowa City Police Department based on accusations of "racial profiling" very unlikely. However, the Department must still recognize that this does not preclude the actions of 30 any one officer becoming suspect. Our findings do not conclude that such profiling might not be occurring against individual citizens by one or more individual officers. This type of discrimination on an individual level, however, is virt~mlly impossible to detect or to prove given the type and amount of discretion that officers must use in the completion of their duties. These matters exe more likely to be discovered through admini.~'ative ~ Sllp~'xdSOP] vigilance, and through community awareness, rather than through the collection and analysis oftraffic stop The Iowa City Police Department can enhance their collection of traffic stop d_ata~ The recommendations offered here involve both process and content elements of the project. First, it is suggested that a full year of data be collected and subsequently used as a baseline for analyzing fixture department practices. The d~a in the study covers only 8 months of the year 2001 and may not fully reflect the traffic stops practices of the department on an mmual basis. Second, census population data should not be used as a baseline. As previously discussed, census data does not provide for an appropriate point of comparison and should only be used when nothing else is available. Clearly, with the adoption of the recommendation for a full year of data collection, the use of census data can be avoided. Third, data collected for the year 2001 (April-December) should he viewed carefully as the depat~anem experienced considerable challenges in refining the data entry process. Throughout the eom-se of this project quality assawanoe cheeks were employed to ensure that the data collected was valid although it is suggested that the validity of the data may continue to be somewhat suspect. Continued monitoring of date entry and fine-tuning of the department's quality assurance mechanisms, however, must be a priority. A fourth recommendation involves the training of all officers in regard to departmental policy, data collection procedures, and the 31 results of the analysis. OITlcor$ collecting the data must have a thorough understanding of the project in order to ensure more accurate and complete data collection and entry. In a similar vein, supervisors must be proactive in ensuring line officers understand the policies and pmcodures related to the project. Supervisors also should identify officers who require additiollal trlttining or closer supervision to ensure adequate understanding of the data entry procedures as well as policy compliance. FiRh, it is imperative that the depadmunt establish clear, written guidelines regarding the entry of traffic stops into both the CAD and MDT database systems. These guidelines should be made available to all personnel involved in the data entry process and should be incorporated into depaxtmental training as required. Further, dispatchers should receive guidelines and training regarding recording calls when mom than one officer in involved in a traffic stop. This will allow for more timely and accurate quality assurance checks as well as enhance the validity of the data. In terms of the content of the data collection forms, several data elements could be added to the form. First, in attempt to control for variations in traffic stop practices by location, the quadrant in which the stop occurred could be added to the form allowing for traffic stop identification. Also, the form should contain information about warrant checks. First, there should be a question that asks whether a warrant check was performed during the stop. Secondly, the form should contain a section addressing the outcome of the warrant check. Currently, information about outstanding warrants is obtained through a plate and/or license check. These types of checks, however, are not performed routinely. Finally, the form should include an item that indicates whether the driver was asked to exit the vehicle. These additions are consistent with data collection efforts throughout the country, require minor modifications to 32 the form, and would aid in th~ development of a more accurate understanding oftbe key events that arc likely to occur during traffic stops. These recommendations are offered to improve the data collection process and to enhance the quality of the data. Several of these recommendations were communicated to the Department as the study progressed and have been addressed. Others are currently being implemented. Overall, the departmental arlmini~'afion has been receptive to recommendations for the improvement of their data collection and analysis, and seems genuinely concerned about the accurate measurement of traffic stop practices. Again, the only major concern is that this study is based on only 8 months of data with which some minor collection and entry problems were noted. Therefore, it is necessary that a full year's worth of"clean" data be collected and analyzed to provide the best baseline from which to evaluate the future stop practices of the depa~t~nent. Although no evidence of departmental discriminatory stop practices may be welcome news, the department now is faced with the responsibility of continual monitoring to maintain these practices for the continued benefit of both the department and the community. 33 Biblio~a~hv MacDonald, H. (2000). The burden of bad ideas: How modem intellectuals misshape our society. Chicago: Ivan R. Dee. Newport, F. (1999). Racial profiling is seen as widespread, particularly among young Black men...Gallup Polk December 1999, 8411, 18-23. Ramirez, D., McDevitt, J. & Farrell, A. (2000). A resource guide on racial profiling data collection systems: Promising practices and lessons learned. Simms, J. (2000). The Maryland 1-95 corridor study. University of Washington in Missouri. (http://www. artsci.wust l.edu/~focus205/supreme/stats_i95 .ht mi). Smith, M. & Petrocelli, P. (2000). Racial profiling: A multivariate analysis of police traffic stop datm Withrow, B.L. (2002). Race based policing: An initial analysis of the Wichita Stop Study. Paper presented at the meeting of the Academy of Criminal Justice Sciences, Anaheim: CA. Zingraff, M. Warren, P., Tomaskovic-Devey, D. Smith, W., McMurray, H., Mason, M & Fenlon, C. (2001). Evaluating North Carolina State Highway Patrol Data: Citations, warnings and searches in 1998. North Carolina Departmem of Crime Control and Public Safety. (On- line). Available: www.nccrimecontrol.org/shp/ncshreport.htm 34 APPENDIX A Iowa City Police Department Policy on Racial Profiling General Order # 01-01 Section Code OP$-17 0PS-17.1 RACIAL PROFILING IDate of Issue Janua~ 10~ 2001 I General Order Number 014)1 I Effective Date J Section Code February 1, 2001 0PS-17 Reevaluation Date December 2001 I Amends / Cancels New IC.A.~-E.A. 1.2.4,1.2.9,41.3.8,61.1.2.9 IReference INDEX AS: Racial Profiling Search and Seizure Complaints Traffic Stops Supervisor Responsibilities Arrests Warrants Discipline I. PURPOSE The purpose of this order is to unequivocally state that racial and ethnic profiling by members of this department in the discharge of their duties is totally unacceptable, to provide guidelines for officers to prevent such occurrences, and to protect officers from unfounded accusations when they act within the parameters of the law and departmental policy. II. POLICY it is the poliGy of the Iowa City Police Department to patrol in a proactive manner, to investigate suspicious persons and circumstances, and to actively enforce the laws, while insisting that citizens will only be detained when thers exists reasonable suspicion (i.e. articulable objective facts) to believe they have committed, are committing, or are about to commit an infraction of the law. Additionally, the seizure and request for forfeiture of property shall be based solely on the facts of the case and without regard to race, ethnicity or sex. III. DEFINITIONS Racial profiling - The detention, interdiction, exercise of discretion or use of authority against any person on the basis of their racial or ethnic status or characteristics. Reasonable suspicion - Suspicion that is more than a "mere hunch" or curiosity, but is based on a set of articulable facts and circumstances that would warrant a person of reasonable caution to believe that an infraction of the law has been committed, is about to be committed or is in the process of being committed, by the person or persons under suspicion. ("Specific and articulable cause to reasonably believe criminal activity is afoot.") OPS-17,2 IV. PROCEDURES The department's enforcement efforts will be directed toward assigning officere to those areas where there is the highest likelihood that vehicle crashes will be reduced, complaints effectively responded to, and/or crimes prevented through proactive patrol. A. In the absence of a specific, credible report containing a physical description, a person's race, ethnicity, or gender, or any combination of these shall not be a factor in determining probable cause for an arrest or reasonable suspicion for a stop. B. Motorists and pedestrians shall only be subjected to investigatory stops or brief detentions upon reasonable suspicion. C. Traffic enforcement shall be accompanied by consistent, ongoing supervisory oversight to ensure that officers do not go beyond the parameters of reasonableness in conducting such activities. 1. Officers shall cause accurate statistical information to be recorded in accordance with departmental guidelines. 2. The deliberate recording of any inaccurate information regarding a parson stopped for investigative or enforcement purposes is prohibited and a cause for disciplinary action, up to and including dismissal. D. Motorists and pedestrians shall only be subjected to investigatory stops or brief detentions upon reasonable suspicion that they have committed, are committing, or are about to commit an infraction of the law. Each time a motodst is stopped or detained, the officer shall radio to the dispatcher the location of the stop, the description of the parson detained, and the reason for the stop, and this information shall be recorded. E. If the police vehicle is equipped with a video camera, the video and sound shall be activated prior to the stop to record the circumstances surrounding the stop, and shall remain activated until the parson is released. F. No motorist, once cited or warned, shall be detained beyond the point where there exists no reasonable suspicion of further criminal activity. G. No parson or vehicle shall be searched in the absence of a warrant, a legally recognized exception to the warrant requirement as identified in General Order 00-01, Search and Seizure, or the parson's voluntary consent. 1. In each case where a search is conducted, information shall be recorded, including the legal basis for the search, and the results thereof. 2. A cursory "sniff" of the exterior of a vehicle stopped for a traffic violation by a police canine may be recorded on the department's canine action report form. TRAINING Officers shall receive initial and ongoing training in proactive enforcement tactics, including training in officer safety, courtesy, cultural diversity, the laws governing search and seizure, and interpersonal communications skills. 1. Training programs will emphasize the need to respect the rights of all citizens to be free from unreasonable government intrusion or police action. COMPLAINTS OF RACIAL/ETHNIC PROFILING Any person may file a complaint with the department if they feel they have been stopped or searched based on racial, ethnic, or gender-based profiling. No person shall be discouraged 0PS-'17.3 or intimidated from filing such a complaint, or discriminated against because they have filed such a complaint. 1. Any member of the department contacted by a person, who wishes to file such a complaint shall refer the complainant to a Watch Supervisor who shall provide them with a depad~nental or PCRB complaint form. The supervisor shall provide information on how to complete the departmental complaint fomn and shall record the complainants name, address and telephone number. 2. Any supervisor receiving a depaCanental complaint form regarding racial/ethnic profiling, shall forward it to the Commanding Officer Field Operations and all such complaints shall be reviewed and the complaint acknowledged in writing. The complainant shall be informed of the results of the department's review within a reasonable pedod of time. The report and the reviewer's conclusion shall be filed with the Chief of Police, and shall contain findings and any recommendations for disciplinary action or changes in policy, training, or tactics. 3. Supen/isom shall review profiling complaints, as well as pededically review a sample of in-car videotapes of stops of officers under their command. Additionally, supervisors shall review reports relating to stops by officers under their command, and respond at random to back officers on vehicle stops. 4. Supervisors shall take appropriate action whenever it appears that this policy is being violated. REVIEW 1. On an annual basis or as requested by the Chief of Police, the Commanding Officer Administrative Services, shall provide reports to the Chief of Police with a summary of the sex, race, and/or ethnicity of persons stopped. 2. If it reasonably appears that the number of self-initiated traffic contacts by officers has unduly resulted in disproportionate contacts with members of a racial or ethnic minority, a determination shall be made as to whether such disproportionality appears department wide, or is related to a specific unit, section, or individual. The commander of the affected unit, section, or officer shall provide written notice to the Chief of Police of any reasons or grounds for the disproportionate rate of contacts. 3. Upon review of the written notice, the Chief of Police may direct additional training towards the affected units/sections or to individual officers. 4. On an annual basis, the department may make public a statistical summary of the race, ethnicity, and sex of persons stopped for traffic violations. 5. On an annual basis, the department may make public a statistical summa~ of all profiling complaints for the year, including the findings as to whether they were sustained, not sustained, or exonerated. 6. If evidence supports a finding of a continued ongoing pattern of racial or ethnic profiling, the Chief of Police may institute disciplinary action up to and including termination of employment of any involved individual officer(s) and/or their supervisors. R. J. Winkelhake, Chief of Police WARNING OPS-I 7.4 IThis directive is for departmental use only and does not apply in any ndmlnal or civil proceeding. The~ department policy should not be construed as a creation of a higher legal standard of safety or cam in an evidentiary sense with respect to thirfl:party claims. Violations of this directive will only form the basisl for departmental administrative sanctions. I APPENDIX B Iowa City Police Contact Sheet IOWA CITY POLICE CONTACT SHEET DateofCoatact Time ~ ~_~ L~-riverlnfo R_es-- ~ehicle Registration -- Female _-- Johnson County Non-Iowa month lay ear hour minute -- Unknown _ Other County 0 0 010 010 0 0 0 0 0 0 0 0 Out of State Consent Search 2 212 2 2i2 2 2 2 2 2 2 2 2 [-~Yes j'~Vehicle 4 4 4 414 4 i4 414 4 4 4 4 --Caucasian 5 5 5 5!5 5 15 5i5 5 5 5 5 Blact~Negro/AfdcanAmerican 'l')peofSearch 6 6 6 6 6 6 6 6' 6 6 , 6 6 _-- Asian/Pacific Islander _ Consent 7 7 7 7 7 7 7 7 7 7 7 7 Spanish/Lafino/Hispan~c Officer Safety 8 8 8 8 8 8 ~' ~' 8 8 8 8 -- NatJveAmedcan Indian IncidonttoArmst 9 9 9 9 9 9 '~ ~' 9 91 19 9 _-- Otbe~ _-- Probable Cause Unknown Reason for Contact? Property Seized Z Moving V'relation Use of Force? Outcome None is Equipment or Registafion Vk~atJon ~ None I -- No Action -- -- Alcohol Criminal Offense Driver Citation IWeapons -- Other Viotafion -- Passenger -- A~Test _ Currency Call for Service-Suspect DescJVehicie Desc. Warning Narcotics -- Pre-existing knowledge or information -- Field Interview -- Evidence _ Special Deta, iOther Other Comments 0 If you add any commente to the area listed below, you must da~en the cimle to the left. IOWA CITY POLICE CONTACT SHEET Date of Contact Time Badge Age Driver Info Resident Vehicle Regis~afiorl IIIIII IIIIE~ FF] ::Male Zlowacity nlowa i Female _ Johnson County U NorNowa month day ear hour minute Unknown Obher County 0 0 0 0 01 0 01 0 0 01 0 0 010 -- _--Out of State ConsefltSearch 3 3 3 3 3 3 i3 3!3 3 3 3 3 R_~thnicity L.~No L_~person 4 4 4 4 4 4 4 4 4 4 ~1 4 4 Caucasian 5 5 5 5 5 5 5 5 5 ~ 5I 5 5 - Black/Negro/African American T~ pe of Search 6 61 6 6 6 6 6 6 6 6I 6 6 _-- ~JarVPacific Islander I_ Consent 7 7 7 7 7 7 7 7 7 7 [~, 7 _ SpanislVLafino/Hispanic _ Officer Safety 8 8 8 8 8 8 8 8 8 8 ;8 8 NativeAmedcanlndian Incident to Arrest 9 9 9 9 9 9 9 --9 9 9 9 9 _ Other _ Probable Cause Unknown R_eason/or Contact? P._roperty Seized _ Moving Violation Use of Force? Outcome None _ Equipment or Registafion Violation None No Action Alcohol Criminal Offense Driver Citation Weapons Other Violation Passenger ;Arrest _ Currency Call for Service-Suspect Desc. Nehicle Desc. ;Waming NarcolJcs _ Pm-e~dsting knowledge or infom~afion _ Field Interview _ Evidence _ Spedal Detail _ Other Other Comments (~ lf you add any comments to the area listed below, you must darken the ci.rcle to the left. ICpd contact sheet.x~ Mar~ll ICPD TRAFFIC STOP PRACTICES. Lehman/Mr. R.J. Sir. R.J. Winkelhake/ We're sony to keep you here this late. Lehman/You aren't keeping us. We have kept you and we apologize for that, but this is the people's republic. Winkelhake/I live here. I wouldn't mind staying. - Lehman/You take a month offtoo many things on the agenda. Champion/But we've had a month off. Lehman/Okay. This represents only a reasonable accurate transcription of the Iowa City Council Meeting of Auglts~ lq, 2002 August 19, 2002 Special Work Session Page 102 Winkelhake/What you're going to hear tonight is a report about traffic stop practices for the Iowa City Police Department. Just give you a little bit of reminder we started doing data collection quite a number of years ago. We were the only police department in the State that was doing that. We began doing this particular type of data collection. We're the only police department in the State doing it. The reason we started doing it was because we wanted to know how we were interacting within our traffic stops and the results of those. We had the opportunity to send the Supervisors to an extension administrative officer's course at the University of Louisville and Lieutenant Jackson who is in the back here when we was there he was asked to do a research project which I get to pick the topic and it was profiling. And he did quite an extensive research paper on that. I don't know if you've ever seen that or not. Vanderhoef/I have. Winkelhake/However, he made a lot of recommendations in there and the recommendations that were made were also some that we got fi.om different national symposiums where the same kind of recommendations for the data to be collected and the way to collect it. So that was done. We choose to send data from April 1 st thru the December 31 st of 2001 to the University of Louisville to be able to do a traffic analyzation of what we were doing. And what you have is a report - I think you got it about a week and a half ago - and we have Dr. Angela West from the University of Louisville here tonight to give you the report on the data that we have collected. And when she's done with it and you're done with answers I got just a couple little things that we're going to continue on to leave you know about after you're done with the report. So this is Dr. Angela West. Lehman/Thank you. Angela West/Well thank you for having me here. I'm having a little trouble with my voice this evening so I'm going to try to be brief. What I'm going to do is basically just kind of give you an overview of what we've done and you have the report and hopefully you've read the report and have had access to the report. This is just a summary really of what the full report says. We looked at - because I anticipate several questions so I'm just going to fly thru this - we looked at 38 variables - driver demographics, stop information, officer badge number and we can also put more information in later on the individual officer as that need arises based on the badge number we can obviously get the race, the sex, the age of the officer time and service and that type of thing if that becomes a need of the Department. We looked at two different types of analysis. We had descriptive analysis which is basically just percentages. Everybody's familiar with percentages. That's a very superficial look at what's going on. It's only used to describe events and representations. We also looked at multi-varied analysis to try to provide inferential ability to took in the why. Why things are happening. Why things are the way they are. And also predictive abilities to help us predict future outcomes based on what we see currently, These types of analysis als0 help us to understmad relationships and interactions among all the various events that occur during a stop. A stop does not happen in a vacuum. When someone is stopped there's several different things going on at the same time that impact what happens...that may impact the actual stop itself- the weather, the location, any events that are going on in the community, the officer's mood This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 103 at the time even though...and the reason that happens is because officer's are given a great deal of discretion primarily because they can't fully enforce every single law. So they have to pick and choose the most severe infractions. Excuse me. So and along with the officer and the time of day and the environment and the weather there's driver characteristics and auto characteristics that are going on at the same time. And the type of multi-varied analysis that we used is called CHAID or Chi-Square Automatic Interaction Detector and I attached a printout from one of those processes looking at the variable citation whether or not a citation was issues in the stop. And I'll get to that in just a second, but that'sjust a printout form the CHAID. And what the CHAID analysis does is it looks at each point, each decision point...now it can't look at the initial decision point because that's something that we will never know exactly why the officer stopped the person initially. We have the reason they put on the form, but they may, in some cases, have put that reason just to put down a reason. So the initial traffic stop we will never know whether an officer is discriminating on that very first contact because we have what the officer puts on the form so we cannot analyze the reason for the initial stop as far as whether there's racial discrimination going on there. So we look at each decision point after that and the reason that the officer gives for the stop - whether it was a moving violation or an equipment registration violation. And what CHAID does is it looks at each of those decision points and throws that decision point into a big pot with all the other things that are going on in that stop at that time - the demographics of the driver, the age, sex, race of the driver and any other events and characteristics that we can say might predict any outcome. And it results in the decision tree which you can see why it's called the decision tree on your printout there arranged in a tree-shaped format. Excuse me. And that tree orders these predictors in order of their strength or their importance in predicting whatever event we're looking at. In this case it's a citation. Okay. Excuse me. We looked at five outcomes of interest, the reason for the stop that the officer gave, moving violation, equipment registration violation were the two most prevalent, whether there was a search conducted, the type of search if there was a search conducted, property seized, and the outcome of the stop. And there were three potential outcomes/ warning, citation, or arrest. Our results - you've read those - basically nothing predicted equipment registration violation. There were no significant factors that came to the surface when you threw all that stuffinto a pot. Age was the most significant predictor of receiving a moving violation, but it had significant interactions with sex and residency. So... Excuse me. The base rate for moving violations was 68.6% meaning that out of all the stops 68.6% were because ora moving violation. So what the CHAID does is it takes that base rate - 68.6% - and compares the rate for other groups in other situations to that base rate. It should be fairly similar okay across characteristics. It was not in certain cases. Most likely to be stopped for equipment...or moving violation were those over 40 with non-Iowa registrations. So they had a rate of 84.5%. So 84.5% of all the people over 40 with non-Iowa registrations received a stop for a moving violation compared to that 68.6% it's significantly higher. That's why the CHAID says it's a predictor. Okay?. The next most likely were those under 18 who were female - almost 82% of those were stopped. Whether there was a search conducted had no significant predictors. The type of search conducted there were no significant predictors. Whether there was property seized there were no significant predictors meaning that no particular group or characteristic was higher than the base rate for the entire overall group. For This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 104 outcome of stop there were no factors that predicted arrest, but receiving a warning and receiving a citation did have significant predictors. Receiving a warning...you were more likely to get a warning if you were not searched which kind of makes sense from a practical standpoint. You're not going to get a warning if you're searched hopefully. And equipment registration violation. So most likely to be warned were those who were not searched who had equipment registration violations and who had non-Iowa registrations and that might be an out-of-towner phenomenon. We're going to give you a warning sense you're out of town, sense you're from out of town. Thank you. Whether a driver was stopped for equipment registration violation was the most significant predictor of getting a ticket or a citation. Again age and residency came up as related to that. The base rate was 38.7%. Those most likely to receive a citation were those not stopped for an equipment registration violation. So you go again with you're not going to receive a ticket just for equipment registration violation. Who are over 30 years old and who had Iowa registrations. Excuse me. Race was never the factor that was the most influential in any of the outcomes of the stop. Excuse me again. The next page outlines what we call the baseline dilemma. And what that refers to is that in prior studies of this issue the tendency is for the percentage of stopped drivers - the racial distribution of stopped drivers to be compared to the racial distribution in the community. That's the tendency. There are several problems with that that I outline here and I'll get to that in just a second. But the whole issue revolves around comparing what is to what should be and that's a problem. To determine what should be one has to get a measure of the racial distribution of drivers who are doing something that would make them eligible to be stopped. You have to know who's out there driving in a way that will make them eligible to be stopped. And I call these people the violators. And your proportion - you're racial distribution of stops should mirror the racial distribution of people doing something wrong. Does that make sense.'? So that's what the baseline should be. Actually measttring that is...nobody's been able to do that yet - to find out who's driving in a way that will make them eligible to be stopped. There've been attempts to measure this, but mostly that consists of posting observers either by the side of the road or driving on the road and counting the number of speeders going by and trying to document their race. Okay? So that's the way it's been tried...attempts have been made to measure that. One of the biggest problems I see with that is speeding is the only behavior that they're looking at in those types of studies. Well speeding is a minority reason for a stop. Okay? We did a study in Louisville, Kentucky and only 37% of all the stops were for speeding. So you're missing 63% of the reasons why somebody might be stopped jnst sitting there looking at speeders going by. And not to mention the. difficulty of measuring races of drivers as they're driving past - speeding past. Excuse me. Comparisons to the population - the census data - axe invalid for the reasons that are outlined here. Census figures include the entire population. The population of drivers to be stopped is generally over the age of 15. Driving populations and police stop practices fluctuate depending on several factors. It ignores the fact that a significant proportion ofdrivers stopped are not city residents. That's my biggest point. It's hard to compare a population is a city when...with the population of driver stopped when 38% of the drivers stopped aren't from the city. And there's also no theory to back the belief that the population of drivers stopped should reflect any resident population. Any there's no theory to back the belief that driving characteristics or events should be equally distributed among populations. This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 105 Different groups can have different driving patterns and behaviors. That's why young, male insurance rates are a lot higher than anyone else because we know younger people drive differently and males drive differently. Excuse me. And the conclusions and recommendations we found no evidence in the data that we had that there was systematic discriminatory stop practices. Again in the events that happen aRer the initial stop and in the reasons that the officers gave for the stop. Again this does not preclude the possibility that an individual officer could be individually discriminating, but that's another thing that's almost impossible to measure. You'd have to know exactly what things were going on in the officer's mind in any situation. The best way to do this type of thing is to use this type of information along with complaints from citizens and that's one of the biggest measures of what police are doing wrong or right is to look at have there been complaints that I was stopped for an unnecessary purpose. And also use of force reports to go along with this. The age and the sex of the dryers as I just mentioned young or males typically have the riskiest driving behaviors. And those were the two things that were most predictive of stop outcomes. We've made recommendations to ICPD and Chief Winkelhake is taking those to heart and there's been an on-going renovation process, revision process to the data collection and to improve the quality of the data. Any questions? Champion/You brought up a...I just want to ask (can't hear) or not? The percentage of out-of- towners that were stopped was higher than people with in-town registration, but did you take account of the fact that we have thousands of students here with out of town cars. Is that considered at all or that just random? West/ That would come into play in the analysis. Yes. Lehman/I think you said 38% were non-residents were they actually non-residents or people with out of town plates? West/ Well let's see. We did both looking at vehicle registration Iowa versus non-Iowa and then looking at city, county other county within Iowa. Lehman/So ifI were driving a car with an Illinois plate and I was a student at the University of Iowa I would be considered a resident? West/ No. Champion/No. You'd be an out-of-towner. Lehman/So the 38% many of those could be residents of Iowa City who are here to go to the University. Kanner/It doesn't look at address? West/ But they're really not counted in the census population for the City are they? Kanner/Yes they are. This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 106 Lehman/Well they are, but they would also certainly included in any measure of profiling because they are residents of Iowa City even though they may not be permanent resident - they live here. West/ Okay. That's one reason that we look at vehicle registration - Iowa versus non-Iowa and then looked at city resident versus county resident versus...there were two measures of residency that we used. Lehman/That probably would have skewed it anyway. Vanderhoef/And you were talking about, if I'm reading this correctly, over - I'm presuming - 40 when you say that you're talking age. West/ Yes. Vanderhoef/So that is not our typical student population. West/ Right. Vanderhoef/In the age 40 part of it. Lehman/No. Vanderhoef/I have been thinking the same thing that you were Connie about our out-of-town registrations that the students... Champion/I think some of my kids have been in school long enough to be 40. Vanderhoef/Going back for the second time. Lehman/Questions? Comments? O'Donnell/I think it's very good. Wilburn/Can you comment in general just - I realize we're one of the few it'not the only departments in Iowa doing...eolleeting this information - but from what you've seen with some other parts of the country can you compare (can't hear). We don't have or we haven't...this is our first go at this. We haven't standardized tested the reliability of the instrument, but can you just in general comment on information that may have appeared on other... West/On the data collection instrument? Wilbum/Yeah on the instnmaent itself. On other areas that they may have included. West/ Yes. Well we came in after the fact on this data collection form. Wilburn/Right. This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 107 West/ But it's fairly comprehensive compared to prior studies into current studies. Many, many studies only include the basic demographic information of the driver, some officer information, and maybe the time and location of the stop and the date that type of thing. This has several other factors on here. Wilbum/Is there other information that you might include? West/ Yes. One of the recommendations that we made was to include information on some other events that happened. Of course that's a departmental issue as far as what the need is. For example some other departments ask whether a driver was asked to exit the vehicle. And that might be an indication of discrimination is minority drivers are asked to exit more frequently than white drivers are. Or whether or not a warrant check was requested or conducted on the driver - that's also another one that other departments do. Wilbum/And have you had a chance a conversation with the Department here about perhaps including those types of things or the benefit? Have you had those conversations yet? West/ Yes. We've made recommendations. Wilbum/Okay. Vanderhoef/And have you looked at these in terms of what standard operation procedures are - whether all drivers are asked to get out of the car for instance? To put that into the data? West/ Right. Right. That's a...that's a very specific thing to the department that's doing the study. For example in Louisville we just finished a whole year of their data analysis. One of their standard procedures that their policy is to do is to conduct a warrant search on everybody they pull over just as a matter of practice. That's not standard procedure in every department. But then we found the rate of doing that was not 100% as the chief would have liked. It was only around 86% which is still fairly high, but not consistent Wilburn/I'm glad you pointed out the limitation in the fact that we given what we're doing and the way that we're going about it we cannot know...we can't get into the mind of an individual officer. And I've always been of the belief that one would hope not but any racism stereotypes in the general population, in my opinion, there's no reason to believe that they may not be, you know, close to or similar to what's on any city department and, you know, some folks that I've talked to about the issue and just with racism in general it's good to know that there's no systemic, you know, given the limitations or the constraints that we're aware of in the department. There may be anecdotal - and there will be anecdotal information. West/ Yes. This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 108 Wilbum/But with folks that I've talked to their concern it's, you know, we know that you can't get some of those ideas or thoughts out of people's heads, but what you do with that is a different story. And you can believe what you want to believe, but you're not going to comb at any act against because of your racist, sexist, etc., etc. beliefs and that's why it's important to have...I look at this and I think it's important that we continue to take a look at this because just with our participation in this for any individual officers that may, you know, may use some type of profile whatever their motivation for stopping someone if it's something other than legitimate, you know, police business that's it's being watched and they need to explain their stops, their behavior. I look at it as a tool. West/ Exactly. Wilburn/And a piece along with what we have with our, you know, our other complaint process - our Human Rights Commission complaint process. I don't know if there's...I don't know if we have...if our vehicles have video on them. Okay so that's another piece that adds to the puzzle so I look at this as another bit of information in terms of everything that we're doing. West/ And it's meant to be used as a tool - an administrative tool. Lehman/Since you're negotiating a second contract for a full year of data with Iowa City... West/ Yes. Lehman/...are we going to be a little...is it possible to get better or more sophisticated data than we have now? West/ Yes. They've already made some changes I believe to the...maybe Chief Winkelhake can speak more to that as to what they have done. Pfab/ I have a comment and kind of... Lehman/Well let R.J. answer first for the data question. Winkelhake/(Can't hear) what we're doing. Lehman/Microphone please. This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 109 Winkelhake/The question about what we're doing. At one point we tried to correlate between traffic stops and the data we were collecting we found like an 89% correlation rate and that's not good enough. We've taken some steps after taIking with the people at Louisville and Lieutenant Jackson particularly is reviewing all that data and we're running now at about a 99% to sometimes 100% correlation. And we found a number of glitches that were computer generated. For instance ifa community service officer went to a traffic stop to stand by while the ear is going to get towed, it would show that that stop was made by the community service officer which is not the policy of the Police Department. Lieutenant Jackson's been working with the people that take care of the computers to make sure that doesn't happen. So our correlation has gone from 89 something to very close to 100 and that's month after month after month. And the process is going to continue. IfI could just one thing that she talked about the warrant checks for instance in Louisville they...it was a policy to go ahead and do that. The way the computer is set up for the State of Iowa an~ime an officer runs a license plate or runs your driver's license it comes back whether there is any warrants on you. So that's an automatic. So we never have taken it as an issue to put into policy to check everybody for a warrant because you could do either one of those you're going to get it back automatically. So that's always being done. Pfab/ One of the points that I think Ross related to, but there's a possibly another side to that that is does the officer out there when he knows he's entering his data entry person does it ever cross his mind that I've stopped a lot of minorities here and maybe I should lay off. Is that ever...? West/ I'm not sure I'm...it may, it may not. Pfab/ So it might work against minorities too if they know they're being tracked just to say well. West/ Generally when a behavior like that is being studied when any kind of study is imposed on a law enforcement agency there's an initial adjustment period right after it goes into effect but then people revert back to their normal behaviors. Pfab/ Okay. That's a good point. Winkelhake/As soon as we started saying we're going to start collecting it obviously that puts a flag up say somebody's looking at this. We did take a look at the number of trafftc stops that were made last year. We took the eight month data the April till the end. We decided how many...on average how many and then multiplied it by 12 to get a number and it came out 14,212 and we did the same thing with the data we've collected so far year to date and projected that out to the end of December and that number comes out to 14, 050. So we're very close. There isn't really anybody playing games that we can see just from that number. Kanner/Dr. West thanks for coming up here. What's your recommendation on having a peer review of this study that you've done? This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 110 West/ Excuse me. Kanner/Especially you've been hired by a department and is there a common accepted practice in the statistical field, scientific field to have in this situation especially to have a peer review of your results? West/ Well we were contracted to do this study by the Police Department. It never was meant to be a research project that was publishable. Of course when we go... Karmer/I say peer review in a more general sense since I lack of a better word. But you get my sense of having someone else hired. Do you recommend that someone else be hired to look over the data to do a second review to see if they come up with the same results? West/ No. What normally is done is you do the review on the front end to make sure you're doing the right things in the first place so that that's not necessary on the back end. So what we did was reviewed every shop practice, publication to the...at the point at which we started doing this type of work collected a big library full of publicatious and studies from across the country and reviewed their practices, reviewed their data collection forms, reviewed the government publications from the National Institute of Justice recommendations and created our data collection forms when we had tho input to do that. We didn't have as much input in this situation in the data collection form, but as far as practices go. And then we've, of course, improved on the methodology that had been used in prior studies. Kanner/We got some information tonight - I'll give you a copy of this from Dr. Baldus. I don't know if you have a chance to take a look at that. But there...it did bring some concern to my mind about you mentioned how hard it was to compare the no stops and...but he's of the opinion, I believe, that it is doable. That it is essential actually to make that comparison to the best of our ability. And I would ask the Council that we give Dr. Professor Baldus a chance tonight and if not tonight then another time set aside to respond to this. We've gotten some written information. West/ Did you want me to answer that question? Kanner/Yeah I do, but I also just wanted to continue with the Council just a second. West/ Okay. Kanner/If...since I need some help interpreting this I think it would be good to have Dr. Baldus...Professor Baldus perhaps tonight respond to some of this since he has gone to a lot of effort. He has some expertise in this area. Lehman/Would you...I mean you've received a copy of this. Would you look at it at your leisure look this over and give us a response to it as well. West/ Sure. Absolutely. Karmer/Anyone else like to hear from Professor Baldus tonight? This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 111 Lehman/I think the time is such tonight that it probably...it's almost 10/30. Pfab/ Is it possible that maybe...that maybe tomorrow night? Lehman/Certainly during public discussion or if we wish to have it...if the Council wish to have it...I would like to get your response to this and then make a decision on that. West/ Would you like a written response? Lehman/Read it over and see what you think. West/ Okay. Kanner/Yeah. Personally I don't know what the Council majority feels I would like a written response. I think that would be helpful. Champion/Yeah. Lehman/Obviously this is something you understand far better than we and so if you would I would appreciate that for the rest of the Council. Kanner/And if you could respond I'd appreciate it. Another question I had that was raised by Professor Baldus is regards to people given, I think they were stopped and warning and outcome warning and correlation with race on that. West/ To what are you referring? Kanner/I think he might have it in a different form and again it would be better if he could address it to you directly so we could see a little bit of question and answer. West/ Is that addressed in his? Kanner/Yeah. I guess you have it. West/ Then I shall... Kanner/Figure 1. West/ ...address that in my written response tomorrow if that's... Lehman/It doesn't have to be that quick, but... West/ Oh, okay I measat for your meeting tomorrow. Lehman/This isn't on our agenda tomorrow. Champion/No. This represents only a reasonable accurate/ranscription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 112 Lehman/We're receiving the report. We appreciate your being here to go through it with us, but I do think...I have not read Professor Baldus' comments. I will take them home. My sense is that you will understand them a lot better than we are. West/ Okay. Sure. Lehman/So. Pfab/ A possibility since it's not on the agenda tomorrow Professor Baldus could have five minutes if he wants... O'Donnell/Emie we're not going to get into a debate tonight about this Irvin. Pfab/ No, no I said tomorrow. O'Doanell/I'd like to here her response... West/ I'd like to make a comment to the question that you asked about the people who are not stopped. I don't know how you measure that. And if someone would come up with a way to measure the characteristics of the people who are not stopped I'd be more than happy to do the analysis. Lehman/They're quicker. West/ Apparently. Lehman/Okay. Any other questions. Wilburn/Just a couple comments in terms of that and we'll certainly hear comments on the results of your study as we will from anyone and anyone with a research background can look. One way offthe surface is to observe the officers to have someone observe the officers and you know that depends on how much money and resoumes you want to put into that. But, you know, the other point about the peer review thing your point was well taken about whether or not you're looking at publishing an article something researched but the methodology is spelled out in the report if someone wants to take a look at it it's apparently has happened or is going to happen they can send us a response as to want or come to get their five minutes before us to respond. West/ Right. One thing that when you send something off for publication consideration or in this case it's a report for an agency that is not as in-depth as far as the statistical procedures and numbers and results are. Had it been for publication consideration we would have included a lot more in-depth on the statistical outcomes and probabilities and chi-square values and degrees of freedom and all those kinds of things. But it was for the use of the department, not necessarily for critique for that reason. Champion/R.J. was there any surprises in the outcomes? Did anything surprise you with the percentages? This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 113 Winkelhake/Thc thing that surprised mc most was the age - that 41 the likelihood of getting a ticket. I really didn't have any preconceived notions of what this was going to tell us other than we started a long time ago. I personally didn't think that we were stopping people because of race. It's nice to have this to be able to say there's some documentation about what we do. One of the things you talked about was the Iowa plates versus non-Iowa and we also looked at other Iowa. So you'd also have Iowa Johnson County, other counties, and then out of state. And we did that mainly because of student population. And a lot of the visitors with out of state plates that come in here. There was also another thing that we put on the form was about the whether use of force - whether there was any force that was used and whether it was the driver or the passenger. That was something we wanted to know and we put it in there that way. So there were some things that were done simply because we wanted to know what we were doing and get a better handle. When I talk to my colleagues across the State and they say why in the world are you doing this and my answer is at least why I can tell you what we're doing. l'm not too sure you can tell us what you're doing. All the time no matter what the data is going to be you're going to have people that are going to disagree with it. And that was one of the reasons we choose to find an organization outside of our State. Wilbum/And I think the other piece to add to that is to people do draw different conclusions about your methodology and your findings. It's what you do with that and we have other - again we're continuing to look at this. We have other pieces of data we collect whether it's looking at complaints filed against individual officers and/or video account and so it's pieces of the puzzle that we need to have to try and prevent it from happening. Winkelhake/One of the things we're continuing to use (can't hear) we do look at use of forth. We do look at citizen complaints. But what we're going to do is continue to gather data. We will have it analyzed again. We are going to continue to look at it. We're going to do more so we can get into individual officer's rather than just the Department as a whole and start looking at it from that standpoint. We're fairly well satisfied at this point that the data we're gathering is good data. I think the people who are doing the work agree the data we're collecting now is better than was done before. (End of Side 1, Tape 02-67, Beginning of Side 2) Winkelhake/...report you see a sheet. Well that is actually on a computer, you know a laptop in the car. And the officers can do this very quickly in the car now. And we did make some adjustments to that. We include a five digit number from the dispatch center so we that coalition coming almost 100%. Ross you had made reference to whether or not there are videos in the car, there are. So we have the videos in the car, we have citizen complaints, we have this data, we have a number of things that we can look at. What we are viewing it as is kinda of an early warning system to see once what we are doing so that we have some idea of how we have to respond to changing attitudes if that's necessary. Andwe will be doing this, its going to continue. We're looking to getting a contract signed very shortly and possibly in a year come back to you. Wilbum/Great. This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August 19, 2002 Special Work Session Page 114 Vanderhoef/Then you would make an assessment, I would hope, that this is a good use of public monies and does not create questions in the community that are not valid? Winkelhake/I think it's a good use of money simply because of the fact that I think it is a good thing to do. I think it is something we should be doing. You ask some other Police Chiefs and Sheriffs across the state they're going to tell you its probably not a good use. Any time you have data there's going to be different interpretations of it. The one thing I can do is say "Here this what we do". You're going to interpret it and you may have seven different opinions right here but least wise I can say this is what we do. Wilburn/Well not just the data itself but the methodology used. If someone wants to say what you're doing is flawed... Lehman/So be it. Wilburn/and then we can you're right or no we disagree. Dr. West/I've heard that a lot. Wilbum/Thank you. Lehman/Thank you. Vanderhoef/Thank you. This represents only a reasonable accurate transcription of the Iowa City Council Meeting of August 19, 2002 August18,2002 ~ ' ~ "' '~ ........25~ City Council of Iowa City 410 E. Washington St. C'.i-i i , ,, Iowa City, Iowa i.~.,',,. "'~;" ~."~ 52240 Re: Terry Edwards et al. Traffic Stop Practices of the Iowa City Police Department: April- December 31, 2001 (June 13, 2002) Dear City Council Members: The recently released study -- Terry Edwards et al. Traffic Stop Practices of the Iowa City Police Department: April- December 31, 2001 (June 13, 2002) (hereinafter the "study") - is of interest to us professionally because a primary area of our research is the use of statistical evidence to test the validity of claims of discrimination in the administration of the criminal law. Our resumes, which document our experience in this area, are attachedJ We routinely peer review for scholarly journals empirical studies that are comparable to the Edwards et al. study of Iowa City police stops. On the basis of its findings, the study states that the "data provide no empirical evidence that the ICPD is systematically engaging in discriminatory stop practices." (p.1) The study implies that it has affirmatively established that there is, in fact, no racial discrimination in Iowa City police stop practices. In addition, the study asserts that for the purposes of future research, the racial distribution of auto stop data presented in the report should "become the baseline from which to evaluate futura practices" (p. 25). It may well be that there is no systemic discrimination in Iowa City's auto stop practices and we hope that this is the case. However, the study fails to establish that fact and is incapable of answering the question one way or another. Obviously our critique of the study is merely the opinion of two people. One way to resolve the esoteric methodological issues that our critique raises is to get other opinions. We strongly recommend that you send this report out for peer review by other scholars working in this field? A. Two Decision Points In considering what the study does and does not prove, its is useful to distinguish between two parts of what the study refers to as "stop practices." They are shown in Figure 1 attached to this letter. The first pan, shown in Part I of Figure 1 is the decision ~ Professor Baldus' publications are on page 1-3 of his resume and Professor Woodworth's publications are on pages 3-9 of his resume. O a related issue, ffthe Cdy supports more research on this ~ssue, we suggest that it fol ow standard practice on sucli mat~ers and issue a request for proposals (RFP) and invite competitive bids from interested researchers. Those proposals can then be peer reviewed before a eonlract to do the work is awarded by the City. to stop a motorist. The second part, which is shown in Part II of Figure 1, consists ora series of decisions made after the stop is executed. .~i 'i ~. ~,.: I ~ Fi"~ 2:53 B. The Decision to Stop Ci~'~" ~ ,~.[ ':, The decision to stop is the principal decision of interest. T'b;{~§~ the extent to which race may be a systemic factor in the exercise of officer discretion to stop motorists, one would ideally have information on the racial characteristics of the people who were not stopped. This would enable us to compare the racial composition of those stopped with those who were not stopped.3 When a database contains no information on the persons who were not stopped, as is almost always the case, researchers look for a proxy population that will enable them "to identify the racial distribution of stops that would exist in the absence of discriminatory stop practices" (p.24). With such a population, one can compare, for example, the percentage of blacks among those stopped with the pementage of those who would have been stopped in the absence of discriminatory stop practices. The issue is whether and how this proxy population should be identified. The authors of the study (hereinafter the "authors") are of the opinion that the identification of this population is "extremely, if not impossible, to measure" (p. 24) and that a "comparison to population data is invalid" (p.25). Instead the study asks whether, among the drivers who were stopped, there is an association between the driver's race and whether they were stopped for a moving violation. The study concludes that there is no evidence of "racial bias in drivers being stopped for a moving violation"(p. 21).4 This is the sum and substance of the "multivariate" results bearing on the stop issue. The results fail to discount the possibility of racial discrimination in the stops. They do not even address the question of whether among all the persons stopped the proportion of minorities is higher or lower than the proportion of minorities among all the drivers who could have been stopped but were not. In short, the analysis is irrelevant to the issue of racial bias in the decision to stop a motorist. We agree that the identification of a comparison population raises a number of interpretive issues and that such results must be viewed with caution. However, the use of comparison populations is essential and unavoidable in this kind of research. Researchers routinely compare the racial distribution of the stopped drivers with the racial distribution of comparison populations and compute the disparities. For example, in 1998, we conducted such an analysis of the Iowa City stop data that were available at that time. Specifically, we compared the racial distribution of Iowa City and Johnson 3 Note that with respect to the post-stop decisions in Part II of Figure 1, we do have information that will enable one to compare the racial composition of the two relevant groups, for example, those who were searched and those who were not. This permits analyses of those decisions that is much more powerful than what can be applied to the initial decision to stop a motorist. 4 The study reports that in this analysis race emerged as a third-order predictor" {among 18-30 year old Iowa residents) (p. 21 ) but the details and possible implications of this finding are not reported. Good practice calls for the presentation of the statistical results for key analyses. 2 County residents with the racial distribution of persons stopped in Iowa City.~ -The results are shown in Tables 1 and lA attached to this letter. Table 1 presents the racial distribution of Iowa City residents t.g, ow A)~ and.the racial distribution of people stopped during the eight months in 1,9~tad.2(J0,6:i relevant comparison ts between Row A, which presents the racial ~"r~tbu~tmh of City citizens, and the racial composition of the motorists stopped by the police, which is shown in Rows B.1 and B. 2. The only column, in which the citizens stopped are over- represented in Rows B. 1. and B.2., compared to Row A, is Column D. It indicates that blacks constitute 2.5% of the population of the city, but represent 8-9% of the people stopped. Table lA focuses on the "rates" that citizens are stopped given their. representation in the City (Row B) and the County (Row C). Stop rates provide a more sharply focused measure of the comparative risk that different members of the community face of being stopped while driving in Iowa City. For example, Row A, Column A indicates that 5,028 motorists were stopped during the period of this study. Row B indicates that among the population of Iowa City, the stop rate was .08 (5028/60272). Column D indicates that blacks are the group most at risk of being stopped. Specifically, the black-motorist stop rote is .30 if the comparison population is Iowa City and .19 if the comparison population is Johnson County. These rates are from 3 to 4 times higher than the rates experienced by the other racial groups. We do not suggest that these data constitute definitive evidence of discrimination since the police practices may have changed since 1999. Moreover, the introduction of controls for other motorist characteristics may reduce the magnitude of the black motorist effects documented in Figures 1 and IA. However, we believe that the results presented in these two figures clearly indicate that the authors, in spite of their skepticism about the validity of analyses that involve comparison populations, should have included comparison data in their report. They could then have explained the limitations of this methodology and lef~ it to City Council and the people of Iowa City to assess the validity of the comparisons? Fumhermore, on the issue of stops, the raw data presented in the study raise several questions that deserve exploration: a. "Black and other" males are over-represented by about 25% among drivers stopped in the 21- to 40-year old age range (Table 5, p.9). We would expect this disparity to be even stronger among the males who were stopped. If racial profiling were practiced, minorities in this age group, especially males, would be likely targets and would be over-represented among the motorists who were stopped. s Indeed a comparison of the data in thc study with comparable results from the earlier period may give the community a sense of whether the system is improving. 3 Fl! account for a disproportionate proportion of the stops. Thc racial distribFtigq.[9.:[ people stopped in this period deserves additional attention. C~'; :' (, c. Table 14, page 15 indicates that "black and other" likely than whites to receive warnings. Other research on this issue indicates that when racial profiling occurs and traffic stops of minorities are pretextual, the minorities stopped are less likely to be cited or charged because there was no good cause for the stop in the first place. In this regard, the racial distribution of the 431 persons against whom "no action" was taken would be relevant. C. The Post-Stop Decisions Part II of Figure 1, which is attached to this letter, indicates the range ofdeeisious that may be made after a motorist has been stopped. The Edwards et al. study analyzes thoroughly only two of those decisions - who received warnings and who received citations (pp.22-23). It concludes that those inquiries detected no evidence of bias. Because these are core findings, good practice calls for a report of the statistical models and results of these analyses, which were not included in the study. It is not clear from the report, why it does not focus as well on the following post- stop decisions: a. Other outcomes - such as arrest, field interview, and no action. b. Searches - requested and conducted, as well as search type. c. Property seizures. d. Use of force. The raw data raise questions about the post-stop decisions that, in our judgment, are not answered satisfactorily. The text on page 13 indicates that black drivers consented to searches at a higher rate than whites (28% v. 23%). This resulted in blacks being over-represented (24%) among thc 83 consent searches compared to their representation (9%) among all persons stopped (p.9). Blacks were also over-represented (15%) among those searched "incident to arrest" (Table 12, p.14). In addition, other research indicates that the key decision is not the consensual search itself but the request to conduct the search. When racial profiling is in place, racial minorities are more likely to be asked if they will consent to a search. That information is not presented in the report. Blacks were also more likely to be arrested (13%) v. a 7% rate for whites and 6% for others (p. 15). 4 Conclusion First, the study's conclusion on the core issue of whether there is racial bias in traffic stops is not supported by the evidence. It may well be that there is no bias and we sincerely hope that this is the case. However, the limited scope of the methodology used in this study cannot support the conclusion that there is no racial profiling in Iowa City traffic stops. Moreover, the raw data presented in the report suggest further analysis is needed on the basic stop issue. Second, as for the issue of bias in post-stop decisions, the report only scratches the surface of the issues that should have been addressed concerning post-stop searches, arrests, use of force, property seizures, and final outcomes. In short, far more analysis is required before this study can validly support a judgment about racial profiling in Iowa City traffic stops. If we were reviewing this study for a scholarly journal, we would recommend that it be returned to the authors for a more thorough consideration of their methodology and a more systematic analysis of all of the issues that the data permit them to present. It might also be advisable in any areas that appear to be problematic to conduct follow up research involving a consideration of the evidence and final disposition of specific cases and the collection of questionnaires from a sample of motorists who were stopped. Sincerely yours, David C. Baldus George Woodworth 34 Seventh Ave, North 14 West View Acres 5 DAVID C. BALDUS Joseph B, Tye Professor. University of Iowa College of Law · Iowa City, Iowa 52242-1113 Ph: 319/3 3 5-9012 - Fax: 319/3 3 5-9098 .1nternet : david-baldus~uiowa, edu ACADEMIC EMPLOYMENT UNIVERSITY OF IOWA COLLEGE OF LAW, IOWA CITY, IOWA Joseph B. Tye Professor, 1983 - Present Professor, 1972-83 ~4ssociate Professor, 1969-71 Subjects: Criminal Law, Anti-discrimination Law, Capital Punishment, Federal Cr/minal Law, and Admiralty SYRACUSE UNIVERSITY COLLEGE OF LAW Center for Interdisciplinary Legal Studies Professor and Director, ]981-82 NATIONAL SCIENCE FOUNDATION Director, Law and Social Sciences Program, 1975-76 PRE-ACADEMIC EMPLOYMENT PENNSYLVANIA CONSTITUTIONAL CONVENTION Delegate, 1967-68 GENERAL PRACTICE OF LAW, Pittsburgh, Pennsylvania 1964-68 U.S. ARMY/ARMY SECURITY AGENCY (ASA) Lieutenant, 1958-59 EDUCATION YALE LAW SCHOOL LL.M., 1969 - LL.B., 1964 UNIVERSITY OF PITI'SBURGH M.,4., 1962 (Political Science) DARTMOUTH COLLEGE /I.B., 1957 (Government Major) BOOKS AND MONOGRAPHS Statistical hoof of Discrimination, 386 pages, Shepards-McGraw Hill (1980) (with James W. Cole). Annual Suvvlement, Statistical Proof of Discrimination (1981), (1982), (1983), (1984), (1985), (1986), and (1987) (with James W. Cole). Equal Justice and the Death Penalty: A Legal and Err~irical Ana]vs:is~ 698 pages, Northeastern University Press (1990) (with G. Woodworth & C. ~u'l~sl~0r'! ~ i ~: [ ~ ~r;~'~, 2: ~ ARTICLES~ BOOK CHAPTERS & REPORTS . ~., "State Competence to Ternunate Concession Agreements with A! ~,'i. 03'~nVaclt¥~ Ig~ 56-97 (1964). "Pennsylvania's Proposed Film Censorship Law - House Bill I098," 4 Duauesne L. Rev. 429-40 (1966). "Weffare As A Loan: An Empirical Study of the Recovery of Public Assistance Payments in the United States," 25 Stanford L. Rev. 123-250 (1973). "A Model Statute for the Regulation of Abandoned Railroad Rights of Way" in Re-Use Planning for Abandoned Transttortation Proverties, Final Report to DOT. 109-25 (K. Deuker and R. Zimmerman eds. 1975) (with S, Grow). "A Comparison of the Work of Thorsten Sellin and Isaac Ehrlich on the Deterrent Effect of Capital Punishment," 85 Yale. L. J. 170-86 (t976) (with J. Cole). "Quantitative Proof of Intentional Discrimination," I Evaluation Quarterly 53-85 (1977) (with J. Cole), "Statistical Modeling to Support a Claim of Intentional Discrimination," Am. Statistical Assn., Proceedings nf the Soc. Stat. See. Part I pp. 465-70 (1977) (junior author with J. Cole). "Quantitative Methods for Judging the Comparative Excessiveness of Death Sentences" in The Use/Nonnse/Misues of ADvlied Social Research in the Court: Conference Proceedings, 83-94 (M. Saks & C. Baron eds. 1980). "Identifying Comparatively Excessive Sentences of Death," 33 Stan. L. Rev. 601-77 (1980) (with C. Pulaski, G. Woodworth, and F. Kyle). "Comparative Review of Death Sentences: An Empirical Study of the Georgia Experience," 74 J. Crim. L. & Criminology 661-753 (1983) (with C. Pulaski & G. Woodworth). "Monitoring and Evaluating Contemporary Death Sentencing Systems: Lessons Prom Georgia," 18 U.C. Davis L. Rev. 1375-1407 (1985) (with C. Pulaski & G. Woodworth). "Arbitrariness and Discrimination in the Administration of the Death Penalty: A Challenge to State Supreme Courts," 15 Stetson L. Rev. 133-261 (1986) (with C. Pulaski and G. Woodworth). "Law and Statistics in Conflict: Reflections on McCleskev v. Kemn," ia Handbook on P~vchololtv and Law 251-73 (D. Kagehiro & W. Lanfer eds. 1991) (with G. Woodworth & C. Pulaski). "Race Discrimination and the Death Penalty," in Oxford Companion to the Sut~reme Court of the United States 705-07 (K. Hall ed. 1991) (with C. Pulaski and G. Woodwoxth). Death Penalty P~:ovortionalitv Review Project: Final Report to The New Jersey Supreme Court~ 120 pages plus 200+ pages of tables and appendices, (September 24, 199 I) State v. Robert Marshal[: Reoort to the New Jersey Supreme Court. 80 pa~gel ~S6ptember 24, 1991). "Proportionality Review of Death Sentences: The View of the Special Master," 6 Chance 18-27 (Summer 1993) (with G. Woodworth). (~ii "Reflections on the 'Inevitability' of Racial D~scrmunatmn m Capi~l Sentericlffg and 'Impossibility' of its Prevention, Detection, and Correction," 51 Wash & Lee L. Rev. 419- 79 (1994) (with G. Woodworth and C. Pulaski). "Improving Judicial Oversight of Jury Damage Assessments: A Proposal for the Comparative Additur/Remittitur Review of Awards for Nonpecuniary Harms and Punitive Damages," 80 Iowa L. Rgv. 1109-1267 (1995) (with J. MacQueen & G. Woodworth). Keynote Address: "The Death Penalty Dialogue Between Law and Social Science." 70 Ind. U. L. Rev. 1033- 41 (1995). "Additur/Remitdtur Review: An En'ri~trieally Based Methodology for the Comparative Review of General Damages Awards for Pain, Suffering, and Loss of Enjoyment of Life," (with G. Woodworth and J. MacQueeu) in Reforming the Civil J'ustice System, 386-415 (Likamer, ed. 1996). "When Symbols Clash: Reflections on the Future of the Comparative Proportionality Review of Death Sentences," 26 Seton Hall L. Rev. 1582-1606 (1996). "Race Discrimination in America's Capital Punishment System Since Furman v. Georgia (1972): the evidence of race disparities and the record of our courts and legislature in addressing the issue," Report to A.B.A. Section of Individual Riehts and Respun~ibflities (7/25/97) (19 pages) (with G. Woodworth). "Pediatric Traumatic Brain Injury and Bum Patients in the Civil Justice System: The Prevalence and Impact of Psychiatric Symptomatology," 26 J .Am. Acad. Psvchiala3, L. 247-58 (1998) (junior author with J. Max et al.) "Race Discrimination and the Death Penalty: An Empirical and Legal Overview" (with G. Woodworth) in America's Exneriment with Caoital Punishment) 385-416 (J. Ackar el al, eds. 1998). "Race Discrimination and the Death Penalty in the Post Furman Era: An Empirical and Legal Overview, With Recent Findings From Philadelphia," 83 Comell L. Rev. 1638-1770 (1998) (with G. Woodworth et al.). "The Use of Peremptory Challenges in Capital Murder Trials: A Legal and Empirical Analysis,' 3 U. Penn. J. of Constitutional Law 3-170 (2000) (with G. Woodworth et al,) "Disposition Of Nebraska Capital and Non-Capita[ Homicide Cases (1973-1999): A Legal and Empirical Analysis: Report to the Nebraska Commission on Criminal lustice and Law Enforcement" (October 10, 2001), 120 pages (with G. Woodworth et al.) (forthcoming in the University of Nebraska Law Review). BOOK REVIEW8 'D~ Chambers, Making Fathers Pay," 78 Mich. L. Rev. 750 (1980). M. O. Finkelstein, Ouantitative Methods in Law & W. Fairley & F. Mosteller, Statistics and Public Policy, 1980 Am. Bar. Found. R. $. 409. "W, White, The Death Penaliv in the Einhties" & "H. Bedau, Death is Different." 1 Crim. L. Forum. 185 (1989) (with G. Woodworth & C. Pulaski).....~, ,..~ P^?I~R8 PP-~SENTED SINCE 1985 "Arhih-arinass and Discrimination in Capital Sentencing: A Challenge For Pre~ntad State Supreme Courts," Stetson Law School, March 1985. ">d'bitmrioess and Discrimination in Capital Sentencing: The Georgia Experience," Formnoff Criminal Jnatic¢ Colloquium, N.Y.U. Law School, May 1985. "Statistical Proof in Employment Discrimination Litigation: An Overview", State of Washington Judicial Conference, Tacoma, Washington, August, 1985. "A~bitxaffmess and Discrimination in Capital Sentencing" Symposium on Capital pmushinant, Columbia Law School, December 1985. "Capital Punishment -- A Tragic Choice?" Mount Mercy College, Cedar Rapid~, Iowa, April 1986. "Consistency and Evenhandedness in Federal Death Sentencing Under Proposed Legislation," testimony before House Criminal Justice Subcommittee, Washington, D.C., May 1986. "The Impact of Prnsecutional Discretion on Arbitrariness und Discrimination," American Criminology Society, Atlanta, GA, November 1986. "Death Penalty Cases: The Role of Empirical Data," National Judicial College of San Diego, February 10, 1987. "Individual Right~ and the Constitution: Issues and Trends in the Death Penalty," Controversy & The Constitution Conference, Ames, Iowa, February 12, 1987. "Equal lustice in Proposed Federal Death-Sentencing Legislation: lessor~ fi`om the states," Testimony before the United Slates Sentencing Commission, Hearing on the Commission's responsibility regarding promulgation of sentencing guidelines for federal capital offenses, Washington, D.C., February 17, 1987. "Usable Knowledge from the Social Sciences: A Lawyer's Perspective," University of Nebraska College of Law, April 10, 1987. "Equal Justice and the Death Penalty: Some Empirical Evidence," University of Nebraska College of Law, April 10, 1987. "McCleskey v. Kemp: A methodological critique," Law and Society Association, Washington, D.C., 1un¢ 12, 1987. "Law and Statistics in Conflict: Reflections on McCl¢$key v. Kemp," University of Bristol (March 4, 1988), University of Durham (March 16, 1988), Hebrew University (April 17, 1988), lJniversity of Reading (May 6, 1988), University of Oxford (May 27, 1988). 4 "Arbitrariness and Discrimination in the Imposition of the Deatl~nal~yi~ Testimony before Senate Judiciaw Committee, Washington, D.C., October 2, 1989. "Arbitrariness and Racial Discrimination in Post-Furman D~all~ Sentencing: Imp icat ons for the Racial Just ce Act and Proposed Federal Death-PerialtycLegislaficlni"'T~timony before the Constitutional and C vx Rights Jubcomnultee Washington, D.C., May 3, 1990. ~,~ .... ' "The Proportionality Review of Death Sentence: New Jersey's Options," New Jersey Bar Assembly, Headquarters, New Bmuswlck, New Jersey, April 23, 1992. "Proportionality Review of Death Sentences: New Jersey's Options," Law and Society Association, PhiladeIphia, May 24, 1992. "Regulating the Quunmm of Damages for Personal Injuries through Enhanced Additur- Remittitur Review," Law and Society Association, Philadelphia, May 28, 1992. "Proportionality Review of Death Sentences" & "Race Discrimination in the Use of the Death Penalty," University of Michigan Law School, January 1993. "Reflections on the Reinstatement of the Death Penalty in Iowa," Public Lecture, Cee College, April 1993. "Discretion and Disparity in the Administxation of the Death Penalty" & "Racial and Ethnic Bias in the Criminal Law: Some Trends and Prospects," AALS Workshop on Criminal Law, Washington, D.C., October 29 & 30, 1993. "Improving Judicial Oversight of Jury Damages Assessments: A Proposal for the Comparative AdditurfRemittitur Review of Awards for non-pecuniary harms and punitive damages," Conference of Chief Justices, Williamsburg, Virginia, January 1993; Department of Pediatrics, University of Iowa Medical School, February, 1993; Conference on Civil Justice Reform, NYU Law School, October 1993. "Racial Discrimination in Capital Sentencing: Reflections on its Inevitability and the Imposs~tlity of its Prevention and Cure," Symposium on Racism in the Criminal Law, Washington and Lee Law School, March 11, 1994. "Racial Discrimination in Mortgage Lending," Department of Housing and Urban Development, January 19, 1994. "The Death Penalty Dialogue Between Law and Social Science," Keynote Address, Symposium, Capital Jury Project, Indiana Law School, February 24, 1995. "Reflections on the Failure to Reinstate the Death Penalty in Iowa" & "Claims of Arbitrariness and Discrimination UMer State Law; recent t~ends." Legal Defense Fund Annual Conference on the Death Penalty, Airlie House, Virginia, July 28 & 29, 1995. "Statistical Approaches to Title VII Discrimination Claims" Defense Lawyers Association, Des Moines, September 1995. "The Marshall Hypothesis Revisited," University of Pittsburgh Law School, October 1995. "When Symbols Clash, Reflections of Proportionality Review, Death Sentences," Luncheon speaker, Death Penalty Conference, Seton Hall Law School, Nov. 2, 1995. "Law As Symbol: explaining the uses of the death penalty in Amefica, f DIPiffl ~Ldw School, Chicago, January t996; Northwestern Law School, March 1996. . !.l t9 I:i; "Post-McCleskey Discrimination Claims: Law, Proof and Poss~b~ht~es, Plenary Sess]o~ Legal Defense Fund Annual Conference on the Death Penalty, Georg~tg~,n Universipg, luly26, 1996. L..[] I I i' iO',"b.'-, i, ,. "Preliminary Finding from the Pennsylvania Capital Charging and Sentencing Study" and "Law As Symbol," American Criminology Society, November 1996. "The Death Penalty and How It Might Affect the Iowa Practitioner," Iowa Bar Association Criminal Law Seminar, Des Moines, March 21, 1997. "Race Discrimination and the Death Penalty: Recent Findings from Philadelphia" Plenary Session, Legal Defense Fund Annual Conference on the Death Penalty, Airlie House, Virginia, July 1997; Death Penalty Symposium; Cornell Law School March 1998; American Society of Criminology, Washington D.C. November 1998. "The Death Penalty for Iowa: What Would It Bring," testimony before the Iowa House Judiciary Comrmttee, March 1998. "Race Discrimination and the Proportionality Review of Death Sentences," Yale Law School, March 1998; St. John's Law School, March 1999. "The Use of Peremptory Challenges in Capital Murder Trials: A Legal and Empirical Analysis," Research Club, University of Iowa, December 17, 1999; Center for Socio- Legal Studies, University of Iowa, January 21, 2000; "Race, Crime, and the Constitution Symposium," University of Pennsylvania Law School, January 29, 2000; Law Dept., Erlangnn University, Erlangen, Germany, July 18, 2000. "Race Discrimination in the Administration of the Death Penalty," Senate Judiciary Committee, Pcrmsylvania Legislature, Harrisburg, Pa., J'anuary 22, 2000; The Governor's Race and the Death Penalty Task Force, Tallahassee, Florida, March 30, 2000. "Reflections on the Use of Capital Punishment in Europe and the United States," Political Science Dept., Erlangan University, Erlangcn, Germany, J'niy 17, 2000. "Race Discrimination in the Administration of the Death Penalty: Current Concerns and Possible Strategies for Addressing the Issue During a Moratorium on Executions," ABA's Call to Action: A Moratorium on Executions, ABA ComCerence, Carter Center, Atlanta, Georgia, October ] 2, 2001. "Race and Gender Disparities in the Administration of the Death Penalty: Recent Finding From Philadelphia and Legislative and Judicial Strategies to Reduce Race and Gender Effects," Pennsylvania Supreme Court CommiRee on Racial and Gender Bias in the Justice System, Philadelphia, Pa. December 6, 2000. "Race Discrimination in the Administration of the Death Penalty," Death Penalty Symposium, NYI3 Law School, March 29, 2001. "Reflections on the Use of the Death Penalty in Europe and the United States," Capital Punishment Symposium, Ohio State Law School, March 31, 2001. "Arbitrariness and Discriminatinn in the Administration of the Death Penalty: the Nebraska Experience," Judiciary Committee, Nebraska Legislature, October lC, 2001; University of Nebraska Law School, February 22, 2002. 6 MISCELLANEOUS Member: American Bar Association; American Law Institute; American Society of Criminology; Law and Society Association. Board of Editors: Evaluation Quarterly (1976-79); Law & Policy Quarterly (1978-79); Law and Human Behavior (1984- ); Psychology, Public Policy and Law (1994-). Board of Trustees, Law and Society Association (1992-94). Grant Recipient, N.S.F. Law and Social Science Program 1974-75--"Quantitative Proof of Discrimination." Invited Participant, N.S.F. Sponsored Conference on the Use of Scientific Evidence in Judicial Proceedings, November 1977. Invited Participant, ABA--AAAS Conference on Cross Education of Lawyers and Scientists, Airlie Home, Virginia, May 1978. Reporter, Roscoe Pound Am. Tr. Lawyers Foundation Cone On Capital Punishment, Harvard University, June 1980. Grant Recipient, National Institute of Jmtice, 1980-81, "The Impact of Procedural Reform on Capital Sentencing: the Georgia Experience." Consuhant, Delaware Supreme Court, April 1981 and South Dakota Supreme Court, November 1981, on the proportionality review of death sentences. Member, Special Committee of the Association of the Bar of New York on Empirical Data in Legal Decision Making and the Judicial Management of Large Data Sets (1980-82). Grant Recipient, NSF Law & Social Science Program. "A Longitudinal Study of Homicide Case Processing" (1983). Consultant, National Center for State Courts project on the proportionality review of death sentences (1982-84). Expert witness in McCleskey v. Kemp, 105 S.Ct. 1756 (1987), a capital case challenging the constitutionality of Georgia's capital sentence process. Recipient, Law and Society Association's Harry Kalven Prize for Distinguished Scholarship in Law and Society (with G. Woodworth & C. Pulaski) for our capital panishment reseamh ( June 11, 1987). Grant recipient, State Justice Institute, 1988-1992, "Judicial Management of Judicial Awards for Noneconomic and Punitive Damages" (with Dr. J. MacQueen & J. Gittler). Special Master for Proportionality Review of Death Sentences for the New Jersey Supreme Court: 1988-91. Member, AALS Committee on CUmculum and Research (1994-97). 7 Recipient, "Michael J. Brody Award for Faculty Excellence in Service to the University of Iowa", October 1996. Recipient, "Award For Faculty Excellence," Board of Regents, State of Iowa, October 18, 2000. Grant recipient, Nebraska Crim~ Commission, "The Disposition of Nebraska Homicide Cases (1973-1999)" (2000) Member, AAUP, Iowa Chapter (1969-___.), Member, Executive Board (1992- __), Member Committee A ( 1985-~ GEORGE WOODWORTH CURRICULUM VITAE August, 2002 Address: George Woodworth Department of Statistics FAX: 319-335-3017 and Actuarial Science Voice: 319-335-0816 241 SH Home: 319-337-2000 University of Iowa Intemet: George-Woodworth@uiowa.edu Iowa City, IA 52242 Personal Data: Born: May 29, 1940, Oklahoma City, Oklahoma Marital Status: Married with two children Education: B.A. Carleton College, Northfield, Minnesota, 1962 Ph.D. University of Minnesota, 1966 Employment: Instructor, Department of Statistics, University of Minnesota, 1965-66. Assistant Professor, Department of Statistics, Stanford University, 1966-71. Assistent (Visiting Assistant Professor), Department of Mathematical Statistics, Lurid Institute of Technology, Lund, Sweden, 1970-71 (on leave from Stanford). Associate Professor, Department of Statistics, The University &Iowa, Iowa City, Iowa, 1971- 1996. Associate Director, Director (1973-1980), Acting Director (1982-3), Adviser (1984-present): University of Iowa Statistical Consulting Center. Associate Professor, Department of Preventive Medicine, Division of Biostatistics, University of Iowa, 1990-1996. Professor, Department of Statistics and Actuarial Science, University of Iowa, 1996-. Professor, Department of Preventive Medicine, Division of Biostatistics, University of Iowa 1996-. Research Interests: Bayesian Inference Statistical Computing Applications of Statistics in Biomedical Science, Behavioral Science, and Law and Justice Multivariate Analysis and Discrete Multivariate Analysis Choice Modeling Longitudinal Data Stanford University Ph.D.: '-i~i I. Reading, James (1970). "A Multiple Comparison Procedure for Classifying AIl Pairs out of k Means as Close or Distant". :-' 2. Withers, Christopher Stroude (1971). "Power and Efficiency of a Class ~f2~o~dness o~Fit,'~' Tests." 3. Rogers, Warren ( 1971). "Exact Null Distributions and Asymptotic Expansions for Rank Test Statistics." University of Iowa, Ph.D.: 4. Huang, Yih-Min(1974). "Statistical Methods for Analyzing the Effect ofWork-Group Size Upon Performance." 5. Scott, Robert C. (1975). "Smear and Sweep: a Me,od of Forming Indices for Use in Testing in Non-Linear Systems." 6. Hoffman, Lorrie Lawrence (1981). "Missing Data in Growth Curves." 7. Patterson, David Austin (1984). "Three-Population Partial Discrimination.*' 8. Mot/, Motomi (1989). "Analysis of Incomplete Longitudinal Data in the Presence of Informative Righ! Censoring." (Biostatisties, joint with Robert Woolson) 9. Galbiati-Riesco, Jorge Matrricio (1990). "Estimation of Chnice Models Under Endogenous/Exogenous Stratification." 10. Shin, Mi-Young (1993). "Consistent Covariance Estimation for Stratified Prospective and Casc-Con~xol Logistic Regression." 11. Lian, Ie-Bin (1993). "The Impact of Variable Selection Procedures on Inference for a Forced-in Variable in Linear and Logistic Regression." 12. Nunez Anton, Vicentc A. (1993). "Analysis of Longitudinal Data with Unequally Spaced Observations and Time Dependent Correlated Errors." 13. Bosch, Ronald J. (1993). "Quantile Regression with Smoothing Splines." 14. Samawi, Hani Michel (1994). "Power Estimation for Two-Sample Tests Using Importance and Antithetic Resampling." (Biostatistics, joint with Jun Lemke) 15. Chen, Hungta (1995). "Analysis of Irregularly Spaced Longitudinal Data Using a Kernel Smoothing Approach." (Biostatistics) 16. Nichols, Sara (2000). "Logistic Ridge Regression." (Biostatistics) 17. Dehkordi, Farideh Hosseini (2001). "Smoothness Priors for Longitudinal Covariance Functions." (Biostatisfics) University of Iowa, MS: 18. Juang, Chifei (I 993). "A Comparison of Ordinary Least Squares and Missing Information Estimates for Incomplete Block Data." 19. Wu, Chia-Chen (1993). "Time Series Methods in the Analysis of Automatically Recorded Behavioral Data." 20. Peng, Ying (1995). "A Comparison of Chi-Square and Normal Confidence Intervals for Variance Components Estimated by Maximum Likelihood." 21. Wu, Li-Wei (1996). "CART Analysis of the Georgia Charging and Sentencing Study." 22. Meyers,Troy (2000) "Bias Correction for Single-Subject Information Transfer ir{ Audiological Testing." ~ ~- , r .... :~.', 19 Publications d P bli ,, Referee u cations: ,..: , 1. Savage, I.R., Sobel, M., Woodworth, G.G. (1966), Fine Sh'ucture ofth~Ord~rmg Of. !:i.,7,'; l?robabilities of Rank Orders in the Two Sample Case," Annals of Mathernatical Statistics, 37, 98-112. 2. Basu, A.P., Woodworth, G.G. (1967), "A Note on Nonparametric Tests for Scale," Annals of Mathematical Statistics, 38, 274-277. 3. Rizvi, M.M., Sobel, M., Woodworth, G.G. (1968), "Non-parametric Ranking Procedures for Comparison with a Control," Annals of Mathematical Statistics, 39, 2075-2093. 4.Woodworth, G.G. (1970), "Large Deviations, Bahadur Efficiency of Linear Rank Statistics," Annals of Mathematieal Statistics, 41,251-183. 5. Rizvi, M.H., Woodworth, G.G. (1970), "On Selection Procedures Based on Ranks: Counterexamples Concerning Least Favorable Configurations," Annals of Mathematical Statistics, 41, 1942-1951. 6. Woodworth, (3.0. (1976), "t for Two: Preposterior Analysis for Two Decision Makers: Interval Estimates for the Mean," The American Statistician, 30, 168-171. 7. Hay, I.G., Wilson, B.D., Dapena, J., Woodworth, G,G. (1977), "A Computational Technique to Determine the Angular Momentum ora Human Body," d. Biomechanics, 10, 269-277. 8. Woodworth, G.G. (1979), "Bayesian Full Rank MANOVAJMANCOVA: An Intermediate Exposition with Interactive Computer Examples," dournal of Educational Statistics, 4(4), 357-404. 9. Baldus, DC., Pulaski, C.A., Woodworth, G.G., Kyle, F. (1980), "Identif,fing Comparatively Excessive Sentences of Death: A Quantitative Approach," Stanford Law Review, 33(1),1-74. 10. Louviere, J.J., Henley, D.H., Woodworth, G.G., Meyer, J.R., Levin, I. P., Stoner, J.W., Curry, D., Anderson D.A. (1981), "Laboratory Simulation vs. Revealed Preference Methods for Estimating Travel Demand Models: An Empirical Comparison," Transportation Research Record, 797, 42-50. 11. Baldus, D.C., Pulaski, C.A., Woodworth, G.G. (1983), "Comparative Review of Death Sentences: An Empirical Study of the (3eorgia Experience," The Journal of CriminalLaw and Criminology, 74(3), 661-753. 12. Louviere, J.J., Woodworth, (3.G. (1983), "Design and Analysis of Simulated Consumer Choice of Allocation Experiments: An Approach Based on Aggregate Data," Journal of Marketing Research, XX, 350-367. 13. Baldus, D.C., Pulaski, C.A., Woodworth, G.G. (1986), "Monitoring and Evaluating Contemporary Death Sentencing Systems: Lessons from (3eorgia," U.C. Davis Law Review, 18(4), 1375-1407. 14. Baldus, D.C., Pulaski, C.A., Woodworth, G.G. (1986), "Arbitrariness and Discrimination in the Administration of the Death Penalty: A Challenge to State Supreme Courts," Stetson Law Review, XV(2), 133 -261. 15. Bober, T., Putnam, C.A., Woodworth, G.G. (1987), "Factors Influencing the Angular Velocity ora Human Limb Segment," Journal of Biomechanics, 20(5), 511-521. 16. Gantz, B.J., Tyler, R.S., Knutson, J.F., Woodworth, G.G., Abbas, P., McCabe, I~.FJ,[ Hinrichs, J., Tye-Murray, N., Lansing, C., Kuk, F., Brown, C. (1988), "Evaluation of Five Different Cochlear Implant Designs: Audiologic Assessment and Predic[~c/ie i ~. Performance," Laryngoscope, 98(I0), 1100-6. 17. Tye-Murray, N, Woodworth, G.G. (1989), "The Influence of Final Syllable~s~tion.o:~., the Vowel and Word Duration of Deaf Talkers," Journal of the AcousttcdLT]d~iety_o~ America, 85, 313-321. 18. Baker, R.G., Van Nest, J., Woodworth, G.G. (1989), "Dissimilarity Coefficients for Fossil Pollen Spectra from Iowa and Western Illinois During the Last 30,000 Years," Palynology, 13, 63-77. 19. Shymansky, $.A., Hedges, L.V., Woodworth, G.G. (1990), "A Reassessment of the Effects of 60's Science Curricula on Student Performance," Journal of Research in Science Teaching, 27(2), 127-144. 20. Tye-Murray, N., Purdy, S., Woodworth, G.G., Tyler, R.S. (1990), "Effect of Repair Strategies on Visual Identification of Sentences," Journal of Speech andHearing Disorders, 55, 621-627. 21. Cadoret, R.C., Troughton, E.P., Bagford, J.A., Woodworth, G.G. (1990), "Genetic and Environmental Factors in Adoptce Antisocial Personality," European Archives of Psychiatry and Neurological Sciences, 239(4), 231-240. 22. Chakraburty, G., Woodwurth, G.G., Gaeth, G.J., Ettenson, R. (1991), "Screening for Interactions Between Design Factors and Demographics in Choice-Based Conjoint," Journal of Business Research, 23(3), 219-238. 23. Koehar, S.C., Woodworth, G.G. (1991). "Rank order Probabilities for the Dispersion Problem," Statistics & ProbabiliO, Letters, 14(4), 203-208. 24. Knutson, J.F., Hinrichs, J.V., Tyler, R.S., Gantz, B.J., Schartz, H.A., Woodworth, G.G. ( 1991 ), "Psychological Predictors of Audiological Outcomes of Multichannel Cochlear Implants: Preliminary Findings." Annals of Otology, Rhinology & Laryngology, I00(10), 817-822. 25. Knutson, J.F., Schartz, H.A., Gantz, BJ., Tyler, R.S., Hinrichs, J.V., Woodworth, G.G. (1991), "Psychological Change Following 18 Months of Cochlear Implant Use," Annals of Otology, Rhinology & Laryngology, 100(11), 877-882. 26. Kirby, R.F., Woodworth, C.H., Woodworth G.G., Johnson, A.K. (1991), "Beta-2 Adrenoceptor Mediated Vasodilation: Role in Cardiovascular Responses to Acute Stressors in Spontaneously Hypertensive Rats," Clin. and Exper. Hypertension.- Part A, Theory and Practice, 13(5), 1059-1068. 27. Tye-Murray, N., Tyler, R.S., Woodworth, G.G., Gantz, B.J. (1992), "Performance over Time with a Nucleus or Ineraid Cochlear Implant," Ear and Hearing, 13,200-209. 28. Tye-Murray, N., Purdy, S.C., Woodworth, G.G. (1992), "Reported Use of Communication Strategies by SHHH Members: Client, Talker, and Situational Variables," Journal of Speech & HearingResearch, 35(3), 708-717. 29. Mot/, M., Woodworth, G.G., Woolson, R.F. (1992), "Application of Empirical Bayes Inference to Estimation of Rate of Change in the Presence of Informative Right Censoring," Statistics in Medicine, 11, 621-631. !.--II FT."~ 30. Shymansky, J.A., Woodworth, G.G., Norman, O., Dunkhase, J., Matthews, C., Lift, t2.Cl'.; ~' ' (1993), "A Study of Changes in Middle School Teachers' Understanding of Selected Ideas in Science as a Function of an In-Service Program Focusing on Student Pr~66§t~3t~o[ag" Res. in Science Teaching, 30, 737-755. 31. Wallace R.B. Ross J.E., Huston J.C. Kundel, C., Woodworth, O.O. (199~3)~ ;.r. ,~,a FICSIT Trial The Feasibility of Elderly Wearing a Hip Joint Protective G~:to'~ed6ce,'"i;',;'~fi, bIip Fractures," .t Am. Gedatr. Soc., 41(3), 338-340. 32. Gantz, B.J., Woodworth, G.G., Knutson, J. F., Abhas, P.J., Tyler, R.S. (1993), "Multivariate Predictors of Success with Cochlear Implants," Advances in Oto-Rhino- Laryngology, 48, 153-67. 33. Moil, M., Woolson, R.F., Woodworth, G.G. (1994), "Slope Estimation in the Presence of Informative Right Censoring: Modeling the Number of Observations as a Geometric Random Variable," Biometrics, 50(1), 39-50. 34. Nunez-Anton, V., Woodworth, G.G. (1994), "Analysis of Longitudinal Data with Unequally Spaced Observations and Time Dependent Correlated Errors," Biometrics, 50(2), 445-456. 35. Baldus, D.C., Woodworth, G.G, Pulaski, C.A. (1994), "Reflections on the Inevitability of Racial Discrimination in Capital Sentencing and the Impossibility of Its Prevention, Detection, and Correction," Washington and £ee Law Review, 51(2), 359-430. 36. Cutrona, C.E., Cadoret, R.J., Sub. r, J.A., Richards, C.C., Troughton, E. Schutte, K., Woodworth, G. G. (1994), "Interpersonal Variables in the Prediction of Alcoholism Among Adoptees: Evidence for Gene-Environment Interactions," Comprehensive Psychiatry, 35(3), 171-9. 37. De Fillippo, C.L., Lansing, C.R., Elfenbein, J.L., Kallaus-Gay, A., Woodworth, G.G. (1994), "Adjusting Tracking Rates for Text Difficulty via the Cloze Technique," Journal of the American Academy of Audiology, 5(6), 366-78 38. Gantz, B.J'., Tyler, R.S., Woodworth, G.G., Tye-Murray, N. Fryauf-Bertschy, H. (1994), "Results of Multichannel Cochlear Implants in Congenital and Acquired Prelingually Deafened Children: Five Year Follow-Up," Am. ~ Otol., 15 (Supplement 2), 1-7. 39. Cadoret, R.I., Troughton, E., Woodworth, G.G. (1994), "Evidence of Heterogeneity of Genetic Effect in Iowa Adoption Studies," Annals of the New York Academy of Sciences, 708, 59-71. 40. Bosch, R., Ye, Y., Woodworth, G.G. (1995), "An Interior Point Quadratic Programming Algorithm Useful for Quantile Regression with Smoothing Splines,' Computational Statistics and Data Analysis, 19, 613-613. 41. Cadoret, R.J., Yates, W.R., Troughton, E., Woodworth, G.G., Stuart, M.A. (1995), "Adoption Study Demonstrating Two Genetic Pathways to Drug Abuse," Archives of GeneralPsychiatry, 52(I), 42-52. 42. Tye-Murray, N., Spencer, L., Woodworth, G.G. (1995), "Acquisition of Speech by Children who have Prolonged Cochlear Implant Experience," Journal of Speech & Hearing Research, 38(2), 327-37. 43. Cadoret, R.J., Yates, W.R., Troughton, E., Woodwor*.h, G.G., Stewart, M.A. (1995), "Genetic-Environmental Interaction in the Genesis of Aggressivity and Conduct Disorders," Archives of General Psychiatry, 52(I 1), 916-924. 44. Tyler, R.S., Low&r, M.W., Parkinson, A.J., Woodworth, G.G., Gantz, B.J. (1995), "Performance of Adult Ineraid and Nucleus Cochlear Implant Patients after 3.5 Years of Use," Audiology, 34(3), 135-144. 45. Baldus, D, MacQueen, JC, and Woodworth GG. (1995) Improwng Judicial Oversight.of. Jury Damages Assessments: A Proposal for the Comparative Additur/Remittitur Review of Awards for Nonpeeuniary Harms and Punitive Damages," with John C. M~i~..~)~.~i~ ~ ~:; i ,~: r2 0 George Woodworth, 80 Iowa Law Review 1109 (1995), 159 pages. 46. Parkinson, A.J., Tyler, R.S., Woodworth, G.G., Lowder, M., Gantz, B.J., (199~6~) hA Vqihl[~.irl~ ( Subject Comparison of Adult Patients Using the Nucleus FOF1F2 and F0~t~B~IB5' ?~?~,';',~ Speech Processing Strategies," Journal of Speech & Hearing Research, Volume 39, 261- 277. 47. Baldus, D., MacQuecn, J.C., Woodworth, G.O., (1996) "Improving Judicial Oversight of Jury Damages Assessments: A Proposal for the Comparative Additur/R.emittitur Review of Awards for Nonpecuniary Harms and Punitive Damages," lowa Law Review, (80) 1109- 1267. 48. Cadoret, Remi J., Yates, William R., Troughton, E., Woodworth, G.G. (1996) "An Adoption Study of Drug Abuse/Dependency in Females," Comprehensive Psychiatry, Vol. 37, No, 2, 88-94. 49. Tripp-Reimer, T., Woodworth, G.G., McCluskey, J.C., Bulechek, G. (1996), "The Dimensional StructUre of Nursing Intervention," Nursing Research 45(1) 10-17. 50. Tyler RS. Fryauf-Bertschy H. Gantz BJ. Kelsay DM. Woodworth GO. (1997) "Speech perception in prelingually implanted children at'er four years," Advances in Oto-Rhino- Laryngology. 52:187-92. 51. Tyler RS, Gantz BJ, Woodworth GG, Fryauf-Bertschy H, and Kelsay DM. (1997) "Performance of 2- and 3-year-old children and prediction of 4-year fi.om 1-year performance. American Journal of Otology. 18(6 Suppl):SI57-9, 1997. 52. Miller CA, Abbas PJ, Rubinstein JT, Robinson BK, Matsuoka AJ, and Woodworth G. (1998) "Electrically evoked compound action potentials of guinea pig and cat: responses to monopolar, monophasic stimulation." HearingRe*earch. 119(1-2):142-54, 1998 May. 53. Knutson IF, Murray KT, Husarek S, Westerhouse K, Woodworth G, Gantz BJ, and Tyler RS. (1998) "Psychological change over 54 months ofcochlear implant use." Ear & Hearing, 19(3):191-201, 1998. 54. Gfeller K, Knutson IF, Wondworth G, Witt S, and DeBus B. (1998) "Timbral recognition and appraisal by adult cochlear implant users and normal-hearing adults." Journal of the American Academy of Audiology, 9(1):1-19, 1998. 55. Baldus D, Woodworth G, Zuckerman D, Weiner NA, Broffitt B. { 1998) "Racial Discrimination and the Death Penalty in the Post-Furman Era: An Empirical and Legal Overview with Recent Findings from Philadelphia," Comell Law Review, 88:6, 1998. 56. Green GE. Scott DA. McDonald JM. Woodworth GG. Sheffield VC. Smith RJ. Carrier rates in the midwestem United States for GJB2 mutations causing inherited deafness. JAMA. 281(23):2211-6, 1999 Jun 16. 57. Gantz BJ. Rubinstein JT. Gidley P. Woodworth GG. Surgical management of Bell's palsy. Laryngoscope. 109(8):1177-88, 1999 Aug 58. Featherstone KA. Bloomfield JR. Lang AJ. Miller-Meeks MJ. Woodworth G. Steinert RF. Driving simulation study: bilateral array multifocal versus bilateral AMO monofocal intraocular lenses. Journal of Cataract & Refractive Surgery. 25(9): 1254-62, 1999 Sep. 59. Weiler JM. Bloomfield JR. Woodworth GG. Grant AR. Layton TA. Brown TL. McKenzie DR. Baker TW. Watson GS. Efl¥cts of f~xofenadine, diphenhydramine, and alcohol on driving performance. A randomized, placebo-controlled trial in the Iowa driving simulator. Annals of Internal Medicine. 132(5):354- 63. 2000 Mar 7 · 60. Tyler RS. Teagle HF. Kelsay DM. Gantz BJ. Woodworth GG. Parkinson AJ. S~edh perception by prelingually deaf children after six years of Cochlear implant use: effects of age at implantation. Annals of Otology, Rhinology, & Laryngology - 2000 Dec. 61 Bal ard KJ Robin DA Woodworth G Zimba LD Age-re ated changes in m~t during artlcu ator visuomotor tracking. Journal of Speech Language cE H~tatI~g Reseanch 4z1(4):763-77, 2001 Aug. 62. Gfeller K. Witt S. Woodworth/3. Mehr MA. Knutson $. Effects of frequency, instrumental family, and cochlear implant type on timbre recognition and appraisal. Annals of Otology, Rhinology & Laryngology. 1,I ] (4): 349-56, 2002 Apr. Books, Chapters: 63. Bober, T., Hay, J.G., Woodworth, G./3. (1979), "Muscle Pre-Stretch and Performance," in Science in Athletics, eds. Jut'is Terauds and/3eorge G. Dales, Del Mar CA: Academic Publishers, pp. 155-166. 6~. Hay, J.G., Dapena, ~'., Wilson, B.D., Andrews, J.G., Woodworth, G.G. (1979), "An Analysis of Joint Contributions to the Performance of a Gross Motor Skill," in International Series on Biomechanics, Vol. AB, Biomechanics VI-B, eds. Erling Asmussen and Kuert Jorgensen, Baltimore: University Park Press, pp. 64-70. 65. Hay, $./3., Vaughan, C.L., WoodwortAh/3.G. (1980). "Technique and Performance: Identifying the Limiting Factors," in Biomechanics VII-B, eds. Adam Morecki, Kazimarz Fid¢lus, Krzysztof Kedzior, Andrzej Wit, Baltimore: University Park Press, pp. 511-520. 66. Woodworth, G.G. (1980). '%lumerical Evaluation of Preposterior Expectations in the Two- Parameter Normal Model, with an Application to Prepostefior Consensus Analysis," in Bayesian Analysis in Econometrics and Statistics, ed. Arnold Zellner, Amsterdam: North- Holland Publishing Co., pp. 133-140. 67. Hedges, L.V., Shymansky, J.A., Woodworth, G.G. (1989), Modern Methods of Meta- Analysis: an NSTA Handbook, Washington, D.C.: National Science Teachers Association. 68. Baldus, D.C., Woodworth, G.G., Pulaski, C.A. (1990), Equal Justice and the Death Penalty: A Legal and Empirical Analysis, Boston: Northeastern University Press. 69. Baldus, D., Pulaski, C., Woodworth GG (1992) "Law and Statutes in Conflict: Reflections on McCleskey v. Kemp," in Handbook of Psychnlogy and Law, edited by Dorothy K. Kagehiro and William S. Lanfer. New York: Springer-Verlag, 1992. 70. Baldus, D., Pulaski, C., Woodworth GG (1992) "Race Discrimination and the Death Penalty," with Charles J. Pulaski, Jr. and George Woodworth, in The Oxford Companion to the Supreme Court of the United States. New York: Oxford University Press, 1992, p 705-7. 71. Woodworth, G.G. (1994). "Managing Meta-Analytic Databases," in The Handbook of Research Synthesis, eds. Harris Cooper and Larry V. Hedges, New York: Russell Sage Foundation, pp. 177-189. 72. Lovelace, D. Cryer, J., Woedworth, G.G. (1994), Minitab Handbook to Accompany Statistics for Business Data Analysis and Modelling, 2nd edition, Belmont, CA: Wadsworth Publishing Company. 73. Tye-Murray, N. Kirk, K.L., Woodworth, G.G. (1994). "Speaking with the Cochlear Implant Turned On and Turned Off," in Datenknovertierung, Reproduktion undDrick, eds. I.J. Hochmair-Desoyer and E.S. Hochmair, Wien, Manz, pp. 552-556. 74. Baldus, D. MacQueen, JC, Woodworth GG. (1996) "Additur/Remittitur Review: An Empirically Based Methodology for the Comparative Review of General Damages Awards for Pain, Suffering, and Loss of Enjoyment of Life," with John C. MacQuean and Woodworth, in Reforming the Civil Justice System, edited by Larry Kramer. New York: New York University Press, 1996, p 386, 30 pages. 75. Baldus, D, and Woodworth, GG. (1998) "Race Discrimination and the Death Penalty: An Empirical and Legal Overview," with George Woodworth, in America's Experirffje~b,with. Capital Punishment, edited by James C. Acker, Robert M. Bohm, and Charl0~,l:,,L'ani~'i"' Durham, NC: Carolina Academic Press, 1998, page 385, 32 pages. Unrefereed Articles, Reviews: 76. Libby, D.L., Novick, M.R., Chen, $.A., Woodworth, G.G., Hamer, R.M. (1981), "The Computer-Assisted Data Analysis (CAI)A) Monitor," The American Statistician, 35(3), 165-166. 77. Woodworth, G.G. (1987), "STATMATE/PLUS, Version 1.2," TheAmeriean Statistieian, 41(3), 231-233. 78. Hoffmaster, D., Woodworth, G.G. (1987), "A FORTRAN Version of the Super Duper Pseudorandom Number Generator," Science Software Quarterly, 3(2), 100-102. 79. Baldus, D.C., Woodworth, G.G., Pulaski, C.A. (1987) "Death penalty in Georgia remains racially suspect," Atlanta Journal and Constitution, September 6, 1987. 80. Hawkins, D., Conaway, M., HaeM, P., Kovacevic, M., Sedransk, J., Woodwor&, G.G., Bosch, R, Breen, C. (1989) "Report on Statistical Quality of Endocrine Society Journals,' Endocrinology, 125(4), 1749-53. 81. Woodworth, G.G. (1989). "Statistics and the Death Penalty," Stats. The Magazine for Students of Statistics, 2, 9-12. 82. Baldus, D.C., Pulaski, C.A., Woodworth, G.G. (1989), "Reflections on 'Modern'Death Sentencing Systems," Book review, CriminalLaw Forum, 1, 190-197. 83. Baldus, D., Woodworth, G.G. (1993). "Proportionality: The View of the Special Master," Chance, New Directions for Statistics and Computers, 6(3), 9-17. 84. 'Race Discrimination in America's Capital Punishment System since Furman v. Georgia (1972): The Evidence of Race Disparities and the Record of Our Cou~ and Legislatures in Addressing the Issue," with George Woodworth, Report to the A.B.A. Section of Individual Rights and Responsibilities (1997), 19 pages. 85. Baldus, David C., George Woodworth, David Zuckerman, Neil Alan Weiner, and Barbara Broffitt (2001). '"'fhe Use of Peremptory Challenges in Capital Murder Trials: A legal and Empirical Analysis," University of Pennsylvania Journal of ConstitutionaI Law, February, 2001. 86. "Complement to Chapter 6. The WinBUGS Program," in Bayesian Statistics: Principles, Models, and Applications, SecondEd#ion, by S. James Press, John Wiley and Sons, Inc., New York, 2002. Convention Papers, other Oral Presentations: 87. Woodworth, G.G. (1983), "Analysis ora Y-Stratified Sample: The Georgia Charging and Sentencing Study," in Proceedings of the Second Workshop on Law and Justice Statistics, ed. Alan E. Gelfand, U.S. Department of Justice, Bureau of Justice Statistics, pp. 18-22. 88. Woodworth, G.G., Louviere, J.J. (1985), "Simplified Estimation of the MNL Choice Model using IRLS," Contributed talk at T1MS/ORSA Marketing Science Conference at Vanderbilt University. 89. Woodworth, G.G. (1985), "Recent Studies of Race- and Victim Effects/n Capital' ~ ' Sentencing," Proceedings of the Third Y~orkshop on Law and Justice Statisli~s, .~.d,. Woodworth, U.S. Department of Justice, Bureau oflusfiee Statistics, pp. ~5~$8iI i~ I JG~'ii 90. Woodworth, G.G., Louviere, J.J. (1988), "Nested Multinoraial Logistic Choicq~-Models Under Exogenous and M~xed Endogenous-Exogenous Stratfficatmn," AS~4 p~.~,~ed ~t~s'~f, the Bustness and Econorntc$ Stattsttcs Sectton, American Statistical Assoc~t~/)n . p~. 121- 129. 91. Woodworth, G.G. (1989), "Trials of an Expert Wimess," .4S,4 Proceedings of the Social Science Section, American Statistical Association, pp. 143-146. 92. Kirby, R.F., Woodworth, C.H., Woodworth, G.G., Johnson A.K., (1989), "Differential Cardiovascular Effects of Footshock and Airpuff Stressors in Wistar-Kyoto and Spontaneously Hypertensive Rats," Society for Neuroscience.4bstracts, 15, 274. 93. Woodworth, C.H., Kirby, R.F., Woodworth, G.G., Johnson, A.K. (1989), "Spontaneously Hypertensive and Wistar-Kyoto Rats Show Behavioral Differences but Cardiovascular Similarities in Tactile Startle," Society for Neuroscience Abstracts, 15, 274. Unpublished Technical Reports and Manuscripts under Review: 94. Kadane, JB and Woodworth, GG. (1998) "Hierarchical Models for Employment Decisions," Submitted to,lournal of the 2merican Statistical/Issociation. Archived Data: 95. Baldus, D.C., Woodworth, G.G., Pulaski C.A. (1989). "Procedural Reform Study," Inter- University Consortium for Political and Social Research: Criminal Justice ,trchive. 96~ Baldus, D.C., Woodworth, G.G., Pulaski C.A. (1989). "Charg/ng and Sentencing Study," Inter-University Consortium for Political and Social Research: Criminal Justice.4rchive. Professional Awards: 1987 Harry Kalven pr/ze of the Law and Society Association (w/th David Baldus and Charles Pulaski). 1987 Iowa Educational Research and Evaluation Association, anaual award "For Excellence in the Field of Educational Research and Evaluation for Best Educational Evaluation Study," (w/th Larry Hedges and James Shymansky). 1991 Gustavus Myers Center for the Study of Human Rights in the United States, selection of Equal Justice and the Death Penalty as an outstanding book on the subject of human fights (with David Baldus and Charles Pulaski). 9 Service Activities Departmental Service: · University of Iowa Statistical Consulting Center: Founder, Associate Director, Director (1973-1980) Acting Director (1982-3) Member of Steering Committee and Adviser (1984-present). University Service: Outside member of over thirty Ph.D. dissertation committees, 1973-present. Woodworth, G.G., Lenth, R.V.L. (1982) "A Stratified Sampling Plan for Estimating Departmental and University-Wide Administration Effort." University of Iowa, Basic Mathematics Committee, January 1983-84. Statistics Advisor to the University ~f Iowa Journal of Corporation Law, 1984-85. University of Iowa, Research Council, 1984-87, Chairman 1986-87. University House Advisory Committee, 1986-87. Chairman, Political Science Review Committee, 1988-89. Interdisciplinary Ph.D. Program in Applied Mathematical Sciences, 1988-present. University of Iowa, Judicial Commission, 1979-81, 1990-93. University of Iowa, Liberal Arts Faculty Assembly, 1985 -87, 1995-6. Professional Service: NAACP Legal Defense and Education Fund, 1980-3: Statistical Analysis of the Georgia Charging and Sentencing Study, Expert testimony in McCleskey vs. Zant (decided in the U.S. Supreme Court). ASA Law and Justice Statistics ComnUttee, 1982-1987: Member of two methodological review panels in Washington, DC. Organizer of two-day Workshop on Law and Justice Statistics, August 1985. ASA Visiting Lecturer Program, 1984-1988. 1984 Invited talk at COver-Stockton College 1986 Invited talk at Moorhead State University 1988 Invited talk at Grinnell College Invited Participant, 1984, Planning Session for Florida Capital Charging and Sentencing Study, Florida Office of Public Defender, Richard H. Burr, Esq. Editor, Proceedings of the Third Workshop on Law and Justice Statistics, American Statistical Association, 1985. Invited Panelist, 1986 Law and Society Association Annual Meeting, Panel discussion of currant state of capital sentencing research. Invited Speaker, 1987 Seminar-Workshop on Meta-.4nalysis in Research, University of Puerto Rico, San Juan, Faculty of Education, Department of Graduate Studies. Associate Editor, Evaluation Review, 1983-1986. Baldus, D., Woodworth, G.G., Pulaski, C.A. (1989). Oral Testimony before the U.S. Senate Judiciary Committee (presented by D. Baldus). Invited Participant, ASA Media Experts Program (1989). 10 Statistical Consultant to Special Master, David Baldus. State of New Jersey, Administrative Office of Courts -- Proportionality Review System. 1989-present. ASA Law and Justice Statistics Committee, second appointment, 1993-95. Baldus, D., Woodworth, G.G. (1993), "An Iowa Death Penalty System in the 1990's and Beyond: What Would it Bring?" Report submitted to the Senate Judiciary Committee Iowa Legislature, February 24, 1993. Baldus, D., MaeQueen, J.C., Woodworth, G.G. (1993), "An Empirically-Based Methodology for Additur/Remittitur Review and Alternative Slrategies for Rationalizing Jury Verdicts," Report prepared for the Research Conference on Civil Justice Reform in the 1990's. Baldus, D.C., Woodworth G.G. (1995), "Proportionality Review and Capital Charging and Sentencing: A Proposal for a Pilot Study," Commonwealth of Pennsylvania, Administrative Office of Courts. Refereeing (since 1980): 1980: Journal of the American Statistical Association 1982: Journal of Educational Statistics 1983: Journal of Statistical Computation and Simulation ~ Annals of Mathematical Statistics <~.~-'~ Evaluation Review (associate editor) . 1984: Transporlation Research . ..- _ Law and Society Review ': American Journal of Mathematical and Management Sciences : ' .~ Journal of Educational Statistics Evaluation Review (associate editor) -~:..,. ;-.: ca.. 1985: Edited Proceedings of 3rd Workshop on Law and Justice Statistics Evaluation Review (associate editor) 1986: Psychological Bulletin National Science Foundation Evaluation Review (associate editor) 1987: J. Amer. Statist. Assoc. 1988: Science (ca. 1988) 1990: Annals ofOtology, Rhinology & LaEmgology American Speech-Language-Hearing Association Macmillan Publishing Company Survey Methodology Journal 1991: International Journal of Methods in Psychiatric Research 1993: Multivariate Behavioral Research 1994: International Journal of Methods in Psychiatric Research 1995: SIAM Review Duxbury Press Acta Applicandae Mathemaficae 1996: American Journal of Speech-Language Pathology 1998: Duxbury Press 2001: John Wiley and Sons, Inc. 2002: Addison-Wesley 11 Extramural Consulting: American College Testing Kaiser Aluminum ':' :~' ~.~ I '~ f::i~, 3: Allergan Electric Power Research Institiife- Beling Consultants, Moline IL NAACP Legal Defense and Educa;tion Fund .:.,' Bettendorflowa AEA National Research Council ~ I[: : ~,', Coerr Environmental, Chapel Hill Supreme Court of Nebraska ILJ; .~.-',~ . ~ Defender Association of Philadelphia Pittsburgh Plate Glass Death Penalty Information Center Rhone-Poullenc Florida State Public Defender's Office Stanford Law School Gas Research Institute. StarForms Hoechst Marion Roussel / Aventis Supreme Court of New Jersey HON Corporahon Vigertone Ag Products Legal Services Corporation of Iowa Westinghouse Learning Corporation Iowa State Attorney General's Office WMT news department Intramural Consulting: I consult almost on a weekly basis with colleagues and students throughout the University, including at one time or another (but not limited to): Audiology, Biology, Exercise Physiology, Geology, Law, Marketing, Nursing, Otolaryagology, Physics, Psychology, Psychiatry, Science Education, the Iowa Driving Simulator, and the National Advanced Driving Simulator. Expert testimony / depositions: Robert R. Lang, Esq. (Legal Services Corporation of lows) 1982 Ruby vs. Deere (gender discrimination) Mark R. Schuling, Iowa Assistant Attorney General. 1984 Burlington Northern Railroad Co. vs. Gerald D. Bait, Director (taxation) Teresa Baustian (Iowa Asst. Atty. General - Civil Rights Division) 1988 Howard vs. Van Diest Supply Co. (age discrimination) Walter Braud, Esq. 1988 Hollars et. al. vs. Deere & Co. ct. al. (gender discrimination) Mark W. Schwickerath, Esq. 1988 Schwickerath vs. Dome Pipeline, Inc. (effects of chemical spill) Richard Burr, Esq. 1990 Selvage vs. State of Florida (capital sen~ncing) Amanda Potterfield, Esq. 1990 Reed vs. Fox Pool Corporation (product liability) 1994 State of Iowa vs. Dalley (forensic identification via DNA) Jerry Zircanerman, Esq. 1991 George Volk Case (age discrimination) 1993 Rasmussen vs. Rockwell (age discrimination) 1994 Hans vs. Courtaulds (age discrimination) Thomas Diehl, Esq. 1992 State of Iowa vs. William Albert Harris (jury composition) Diane Kutzko, Esq. (Iowa State Bar Association) 1995 Consultation on the validity of the Iowa bar exam. John Allen, Esq. 1995 Buchholz vs. Rockwell (age discrimination) Michael M. Lindeman, Esq. 1995 Beck vs. Koehring (age discrimination) Timothy C. Boller, Esq. 12 1995 Larh vs. Koehring (age discrimination) Thomas C. Verhulst 1995 Cart vs. $.C. Penny (racial discrimination) J. Nick Badgerow, Esq. 1995 Zapata et. al., vs. IBP, Inc. (racial/national origin discrimination) David J. Goldstein, Esq., Faegre and Benson, Minneapolis 1999 Payless Cashways, Inc. Partners v. Payless Cashways (age discrimination) Catherine Ankenbamdt, Deputy First Assistant Wisconsin State Public Defender 2001 Civil commitment hearing of Keith Rivas (Prediction of Sexual Recidivism) Michael B. McDonald, Assistant Florida Public Defender 2001 Frye hearing in re Actuarial Prediction of Sexual Reeidivisim Greg Bal, Assistant Iowa Public Defender 2001 Civil commitment hearing of Lanny Taute (Prediction of Sexual Recidivism) Harley C. Erbe, Esq. Walker Law Firm, Des Moines 2002 Campbell et al. v. Amana Company (Age Discrimination) Texas State Counsel for Offenders, Huntsville, TX 2002 Daubert bearing in re Actuarial Prediction of Sexual Recidivisim Michael H. Bloom, Assistant Wisconsin Public Defender 2002 Detention of Morris F. Clement, Forest County Case No. 00 C101 (Prediction of Sexual Recidivism) cZ2, Federal Court Division, Defender Association of Philadelphia, Capital Habeas ~s 2002 Petitioner Reginald Lewis (racial discrimination) ..... 57'1 13 Iowa City Police Department Traffic Stop Data Analysis: aoox Presented to The City Council of Iowa City, IA By Terry D. Edwards, J.D. Elizabeth L. Grossi, Ph.D. Gennaro F. Vito, Ph.D. Angela D. West, Ph.D. University of l.ouisville Department of Justice Administration Louisville, KY August 19, 2oo2 Introduction & Overview , Analysis of 9,7o2 contacts occurring over the 9 month period from April l- December 3~, 2oo~ Contract between ICPD and researchers at the University of Louisville · ICPD contact sheet/MDT screens · Formatted into an Excel spreadsheet, then transferred into SPSS for analysis · 38 variables (driver demographics, stop information, officer badge number) · Minor glitches with the data collection that were addressed as they arose, or when they became known Data Analyses · Two levels of analysis: descriptive and multivariate · Descriptive analyses provide percentages and give only a very superficial look at the data--they describe the current state of affairs / Lack inferential ability: cannot answer "why" ~ Cannotpredictevents:cannotaddress"whatif' */ Cannot describe relationships among variables ~ Only a "first step" in a thorough analysis · Multivariate analyses provide an in-depth examination of the data ~ Inferential: can help to answer "why" ~ Predictive: can help to predict future outcomes ~ Can help to understand relationships and interactions between and among variables that lead to a certain reality (as portrayed by the percentages) · CHAID: Chi-Square Automatic Interaction Detector (see attached) / Examines each decision point (e.g., arrest, citation, moving violation) ,/ Determines the ability of driver demographics (age, sex, race) and other events and characteristies to predict any particular outcome ,~ Results in a "decision tree" that orders factors related to the outcome in order of their strength (predictive power) ~ Outcomes of interest: ~) Reason for the stop (moving violation, equipment/registration violation)?; 2) Search conducted?; 3) Type of search (incident to arrest, consent)?; 4) Property seized?; 5) Outcome of stop (warning, citation, arrest)? Results · Reason for stop? ~ No factor (race, sex, age, residency) was a significant predictor of an equipment/registration violation ~ Age was the most significant predictor of a moving violation. It had significant interactions with sex, and residency. The "base rate" for moving violations was 68.6%. Most likely to be stopped for this reason were those over 40 with non-Iowa registrations (84.5%). Next most likely were those under 18 who were female (8~.8%). · Search conducted of driver or vehicle? ,/ No factor (race, sex, age, residency) was a significant predictor of whether a driver or a vehicle was searched. · Type of search conducted (consent or incident to arrest)? ~ No factor (race, sex, age, residency) was a significant predictor of whether a consent search or a search incident to arrest was conducted. * Property seized? ~ No factor (race, sex, age, residency) was a significant predictor of whether property was seized. · Outcome of stop (warning, citation, arrest)? ~ No factor (race, sex, age, residency, reasonfor stop, multiple reasons for stop, search conducted, property seized) was a significant predictor of arrest. ~ Whether a search was conducted was the most significant predictor of receiving a warning. It had significant interactions with reason for stop and residency. The "base rate" for receiving a warning was 55-5%. Most likely to be warned were those who were NOTsearehed, who had equipment/registration violations, and who had non-Iowa registrations (77.1%]. / Whether a driver was stopped for an equipment/registration violation was the most significant predictor of receiving a citation. It had significant interactions with age and residency. The "bedse rate" for receiving a citation was 38.7%. Most likely to receive a citation were those l~OTstopped for an equipment/registration violation, who were over 30 years old and who had Iowa registrations (50.4%). Next most likely were those NOT stopped for an equipment~registratlon violation who were under 18 years old (50.2%). ,/ Race was never "the factor" influential enough to be predictive of any outcome. The Baseline Dilemma · Comparing "what is" to "what should be" is problematic. · To determine "what should be" one would have to get a measure of the racial distribution of d~vem who are doing something that would make them eligible to be stopped ("violators"). This should mirror the racial distribution of drivers who actually are stopped ("stopped violators"). * Research that has attempted to measure this i~ flawed. Usually use speeding as the only detectable behavior-speeding stops are a minority. · Comparisons to population census data are invalid. ~ Census figures include the entire population and the population of drivers to be stopped is generally only of driving age (over 15); ~ Driving populations and police stop practices fluctuate depending on several factors (measurable and unmeasurable); ,/ Ignores the fact that a signh~cant proportion of drivers stopped are not city residents (38% in the current study); e' No theory to back the belief that the population of drivers stopped should · reflect any resident population; ,/ No theory to back the belief that driving characteristics/events should be equally distributed among populations--different groups can have different driving patterns/characteristics (males, younger persons, etc.). Conclusion and Recommendations · These data provide no evidence that the ICPD is systematically engaging in discriminatory stop practices. This does not preclude the possibility that any individual officer could be using race as a factor in any individual contact situation. This possibility is not measurable using these data and is unlikely to be measurable in any situation. · Age and sex of the drivers, along with other events related to the stop were more predictive of stop outcomes. · Recommendations have been communicated to the ICPD on an on-going basis and steps have been taken to improve the data collection process and the quality of the data. We are currently negotiating a second contract for a full year of data collection (2oo2). Page 7 David Baldus: Good evening Mayor Lehman and member of Council. I appreciate the opportunity to appear here tonight. (Reads statement). Lehman: Before you...you did,..we did give to Dr. West, I believe, a copy of the information you gave us last night and she has agreed to respond in writing. Did you...and I haven't had an opportunity to study what you gave us last night but basically what you said tonight in the report that you gave us last night. Baldus: This is an elaboration on what I gave you last night. Lehman: We will receive from her a written response. Baldus: Very well, Lehman: And we will see to it that you get a copy. Baldus: Okay. Very good. Thank you. Pfab: Motion to accept enrrespondenee. Champion: Second. This represents only a reasonably accurate transcription of the Iowa City City Council meeting of August 20, 2002 #4 Page 8 Lehman: Moved by Pfab, seconded by Champion to accept correspondepce; Baldus: Thank you. Lehman: All in favor? Opposed? Motion carries. Kanner: So, Emie this looks like this is an issue that probably needs some more discussion. Lehman: I think it will wait until we get the response and then we'll find out whether we think it warrants more discussion. How does the Council feel? O'Donnell: Exactly Emie we'll wait until we get a response. Champion: I think we should wait for the response. Lehman: And we'll see what happens. Kanner: I would say let's have a work session after the response so we can hear both of these peoples give and take and obviously there's some different perspectives. Lehman: We'll have that oppommi~, when we get the response I would say. Kanner: We'll have that. This represents only a reasonably accurate transcription of the Iowa City City Council meeting of August 20, 2002 REMARKS TO THE CITY COUNCIL OF IOWA CITY David C. Baldus George Woodworth August 20, 2002 My name is David Baldus, 34 7th Ave. N. and I teach at the University of Iowa College of Law. Joining me in these remarks is George Woodworth, who teaches in the University of Iowa Department of Statistics and Actuarial Science. We are here this evening to address the validity and accuracy of the principal conclusion of the empirical study on racial profiling in Iowa City traffic stops that was recently presented to city council. The study was prepared by criminologists from the University of Louisville. We have an interest in the methodology and validity of this study because we have spent much of our professional lives conducting empirical studies of the impact of race in the criminal justice system. The bottom line conclusion of the Louisville study is that the data "provide no empirical evidence that the Iowa City Police Department is systematically engaging in discriminatory stop practices." In considering these claims, it is important to distinguish between two very different points of decision in the process of stopping and charging motorists. First is the threshold decision to stop a motorist. This is the core decision about which there is the greatest public interest concerning racial profiling. This threshold decision is followed by a series of post-stop decisions involving searches, citations, warnings and arrests. These decisions raise important, but distinctly secondary issues. The "no bias" claim of the study has led many people to believe that it definitively established that race is not a factor in the initial stop decisions. For example, the Iowa City Press Citizen stated in an editorial this morning that the study shows that "Iowa City police did not systematically engage in a practice of pulling over drivers based on their skin color." The Iowa City police chief has stated on several occasions that he is pleased with the study because it shows that race plays no role in Iowa City stop decisions. However, as much as we hope that the police chief's belief is true, the traffic stop study contains absolutely no data to support either that belief or a belief that racial bias does play a systemic role in the process. The study simply does not address the issue. The only analysis in the study that purports to address the stop issue merely establishes that among the motorists who are stopped, blacks are no more likely than whites to be charged with a moving violation. This conclusion has nothing whatever to do with whether race played a role in the initial stop decisions. On this issue it is important to note that one co-author of the study, Dr. Angela West, agrees with us and has stated: "On the basis of our study, one simply cannot tell if race is a factor in the initial decision to stop motorists." The reason the study provides no basis for answering this question is a fundamental flaw in its research design. It has good racial information on the motorists who were stopped. They were 9% black. However, it contains no racial information on the people who were not stopped. Nor does it contain racial information on the population of stopped motorists that one would likely see if there were no racial bias in the system. Without such a comparison population nothing definitive can be said about the stop issue. To understand the significance of this omission from the study's research design, imagine that we were studying the impact of immaturity, i.e, being 16-18 years of age, on auto accident rates and we only had information on the age distribution of the drivers actually involved in auto accidents. Further, imagine that these data showed that 16-18 year olds were involved in 25% of the accidents. With only that information, we could say nothing at all about the impact of driver immaturity on the risk of being in an auto accident. To make any judgment about that issue, one would need information on the age distribution of all drivers. If the data showed that 16-18 year olds constituted only 10% of the licensed drivers, but were involved in 25% of the accidents, that comparison would provide relevant evidence on the influence of immaturity on auto accident rates. Therefore, to support an inference about the role of race in traffic stops we need racial data on a comparison population of citizens or drivers that could then be compared to the 9% of blacks among the motorists who were stopped. We have no preconceived belief about what the results of a properly conducted study would show. What we do believe it that the citizens and police force of Iowa City deserve to have the best study possible, one that cannot be significantly challenged on methodological grounds. Given the flaws in the Louisville study, it is surprising that so many people misinterpret its meaning regarding the stop issue. The reason is that the study is profoundly misleading. This arises from the weakness of its research design, a lack of clarity in its analysis, and a confounding of its findings about the role of race in the post- stop decisions with the role of race in the initial stop decisions. Because these are technical issues, we suggest that you submit the stop report to peer review by scholars who conduct empirical studies of this type. This study has been subjected to no peer review. Nor was the original research proposal. When co-author Angela West was asked about a peer review of the traffic study, she replied that peer review was not needed because the study was not going to be published in a scholarly journal. In our judgment, there is a far greater need for peer review of a study that is offered up, like the traffic stop study has been, as a basis for public action on an 2 important and sensitive political issue, than there is for peer review of a study published in a scholarly journal, which is unlikely to have any impact on important issues. We thank you for the opportunity to appear this evening and will be pleased to answer any questions you may have. C~ef ~J. Iowa City Police Depmmcm 410 E. W~gon St. Iowa Cky, IA 52240 Dear Chief Winkelhake, Enclosed is a hard copy of the fax that I sent you earlier containing our response to the Baldus and Woodworth critique. Please make copies and distribute as you see fit. However, please note that the actual model (of abiders and violators) may only be used with my permission. I am actually publishing a revised version of this letter in an upcoming issue of the Journal of Foreasic Psychology Practice, as part of a debate on measurement issues related to traffic stop practices. I look forward to hearing what the City Council thinks of the response. Of course, I also am interested in Baldus and Woodworth's reaction, as well. It has been a pleasure working with you and we are looking forward to the second year, as well. If you have any questions or concerns, please do not hesitate to contact me or any other member of the research team. Sincerely, August 22, 2002 City Council of Iowa City 410 E. Washington St. Iowa City. IA 52240 Re: 1) Letter from David Baldus and George Woodworth dated August 18, 2002 and presented to the City Council work group meeting on August 19, 2002; and 2) written memo from Baldus' appearance before the City Council formal meeting on August 20 Dear City Council Members: I am writing as requested to address the concerns raised by Professor Baldus and Dr. Woodworth regarding the methodology and conclusions in our study of the ICPD traffic stops. The critique from Baldus and Woodwotth claims that: 1) Our study "fails to establish that there is no systemic discrimination in ICPD stop practices;" and 2) that our study is "incapable of answering that question one way or another' (p. 1, para. 3). Their critique is based primarily on 'esoteric methodological issues' (p. 1). The "Decision" to Stop/Reason for the Stop They correctly assert that our study examined two principal points of interest. However, the authors have incon'ectlyident'~ied one of those two points as "the decision to stop.' The correct delineation would be "the reason for the stop." The difference is subtle but crucial We cannot measure an officer's decision to stop any particular vehicle except by the reason that he or she provides. That is, to measure why an officer stops a vehicle, we must rely on the reason that the officer provides on the contact report. That may or may not be the real reason that the officer decided to stop the vehicle. Decision-making is an internal process that is not available for measurement on a form; the reason for the stop, on the other hand, is measurable. Officers may engage in biased decision-making processes, but be able to translate those biased decisions into valid reasons for a particular stop. If Baldus and Woodworth have devised a mind-reading method to determine why an officer decides to stop any particular vehicle, or if they find officers willing to indicate on a data collection form that the reason for the stop was "color of driver's skin,' we will be mom than happy to employ either of those methods in any future studies. Short of that, we can only go by what the officer indicates on the form. Reasons for the stop primarily involved moving violations and equipment/registration violations. Again, Baldus and Woodworth Incorrect/y state that our analysis of being stopped for a moving violation was the Usum and substance of the 'multivariate' results bearing on the stop issue' (p. 2). We also analyzed being stopped for equipment/registration violations, other violations, pre-existing knowledge, criminal offense, special detail, and other (see p. 20 of the full report). However, being stopped for a moving violation was the only event that had significantly related predictors (age was the prireary predictor). Post-Sto~ Dectsion~ The second area to which Baldus and Wcodworth refer involves the 'post-stop decisions' (p. 4). Again, the authors incorrectly state that our study 'analyzes thoroughly only two of those decisions - who received warnings and who received citations' (p. 4, para. 3). On page 20 of our full report, we explain that we conducted full CHAID analysis on all the 'post-stop decisions' (having a vehicle or driver search conducted, being searched incident to arrest or by consent, having property seized, and stop outcom.= v;gming, citation, an'est). Only two of those events (receiving a warning and receiving a citation) had significant predictors. For the events with no significant predictors, no further discussion was necessary. The C0mparb0n Gr(;,up/Basellna Dilemma It seems that the primary argument Baldus and Woodworth have pertains to ~ lack of a coreparison group. They emphasize the importance of knowing the 'proportion of minorities among all the drivers who could have been stopped but were not' (p.2, para. 3). This corement is qualitatively and quantitatively different from their earlier claire (p. 2, para. 1) that 'to test the extent to which mca may he a systemic factor in the exercise of officer discretion to stop motorists, one would ideally have information on the racial characteristics of the people who ware not stopped. This would enable us to compare the racial composition of those stopped with those who were not stopped' (p. 2). Drivers who were not stopped are different frorn drivers who couldhave been stopped but were not. It is here where I must distinguish between two subpopulations of drivers. As indicated by the graph attached as Figure 1, at any given time on any given day at any given location under any given set of circumstances, them is a population of drivers ('Ail Drivers"). That population of drivers can be divided into two mutually exclusive categories: Subpopulation #1: 'Ablder~' Abiders are drivers genecally 'nol eligible to be stopped' because they are not doing anything illegal or anything that would other, vise bring them to the attention of law enforcement. ^biders should not be stopped by the police. Suboooulati~rt i~: "V/o/atom" Violators am drivers 'eligible to be stopped' because they are doing something that brings them to the attention of law enforcement (weaving, improper lane changes, failure to signal, speeding, reckless driving, expired plates, inoperable equipment, etc.). V'~ators should be stopped by the police. By all accounts, this subpopuletion consists of the majority of drivers (most drivers could be considered violators, prirearily for speeding). In fact, a survey by the National Highway Traffic Safety Administration (NHSTA) found that 84% of su~eyed drivem reported seeing speeding or other unsafe driving all or most of the time (NHSTA Traffic Tecfl 186, 1999). One can divide abidere and violators into drivers 'stopped' and ~not stopped." Theoretically, *stopped" ddvars should consist only of violators. 'Not stopped" drivers, on the other hand, include both abiders and violatom; abiders should not be stopped and it is impossible to stop all violatom. The job of law enforcement is to stop the violators (and conversely, not stop the abidem). But given that the violators are so numerous, law enforcement oflicars use their discretionary powem in determining who to stop and who to not stop. As a result, many violators am not stopped and abiders sometimes are (sos Figure 2). In this model, them are four possible combinations of outcomes--two of which involve the possibility of discriminatory practices: 1) Abidere who am not stopped Abidera should not be stopped and are not = No discrimination. 2) Abiders who am stopped Abiders should not be stopped and are = Possible discrimination 3) V/o/afore who are ,stopped Violators should be stopped and are = No discrimination 4) V/o/atom who are not Violators should be stopped and are not: Possible discrimination Therefore, there are two potential sources of discrimination in police stop practices, stopping drivers who should not be stopped, and not stopping drivels who should be stopped (outcomes #2 & #4 above). It is interesting to note that research efforts to date have focused on scenario #3--violators who are stopped--given that the only information we have is from tmf~ stops. Theoretically, this is not a potential source of discrimination since an officer must give a valid reason for stopping any vehicle (and so the ddver, per se, is a violator and should have been stopped}. Whether the driver believes the stop is valid is another question. The real focus of inquiry should be on outcomes #2 and #4. Those who say that police "racially profile' would seem to have two main contentions: 1) minority abidem are wrongfully stopped at higher rates than white abiders (and convarsely, white abiders are rightfully not stopped at higher rates than minority abiders). In this case, the minority driver is alleging that he or she is stopped ONLY because of their skin color. They were not doing anything that would make them eligible to be stopped as a violator; 2) minority violators are rightfully stopped at higher rates than white violators (and conversely, white violators am wrongfully not stopped at higher rates than minority violators). In this case, the minority driver acknowledges that he or she was doing something that would make them eligible to be stopped (they are a violator), but law enforcement disproportionately targets them rather than white violators. For the first contention, the most appropdata comparison group would be the subpopulafion of abiders who were not stopped. The racial dis~'ibufion of abiders not stopped should approximate the racial distribution of abidars who were stopped. That would mean that law enforcement was stopping people of both racos equally when they should not have been stopped. Stopping people who should not be stopped is a problem in itself and would need to be addressed by management. Proving that a person should not have been stopped, however, is problematic, given that law enforcement officers have at their disposal a wide variety of reasons for making a traffic stop. For the second contention, the most appropriate comparison group would be the subpoputation of violators who were not stopped. The racial distribution of violators not stopped should approximate the racial distribution of violators who were stopped. This would mean that law enforcement wes stopping people of both races equally when they should have been stopped. For both of these situations, the problem Ile~ in determinln9 the racial dist~fbution of abiders and violators who were not sa~pped. Curranfly, there is no measure of the racial dis~fbution of who is not stopped by the police. Obtaining data on who is not stopped is similar to obtaining data on crimes that are not reported, what c~iminologlats call the 'dark figure of crime." Although victimization and self-report studies attempt to measure this, each method has serious methodological issues that severely limit the validity of crime data it measures. Again, if Baldus and Woodworth have devised a method to do this, we would be more than happy to use it in any future analyses. Finally, these populations and subpopulations are constartfly changing depending on the time of day, the day of week, the week, the month, the season, the weather, social events, location, and many more factors. Wlfh such constantly changing populations, how can one devise valid measures of their characteristics? For example, think of the population driving on Sunday at 9:30 am. Are those drivers different fi.om ddvers on Sunday at 1:30 am? In Iowa City, 41% of the stops were between midnight and 3:00 am. Baldus and Woodworth must recognize that drivers on the road at this time am unlikely to reflect the population of Iowa City, in general. What if we are examining a stretch of roadway that cuts through an Hispanic neighborhood and there is a Latino festival being held? There will be more Hispanic ddvere than normal in that particular population. As to the validity of making conclusions without comparison populations, we believe it preferable to use the cu~ent data as a basis for comparison rather than make invalid comparisons to poorly devised proxy measures. The study is to be used as a management tool, in conjunction with other measures of police performance, including citizen satisfaction surveys, complaints, reports of excessive force, eto. The data and statistical analyses cannot substitute for good police-community relations, but primarily 4 serve as an additional way for law enforcement agencies to measure their performance in this area. Moreover, we believe that the CHAID analysis provides the beselina from w~ich the data can be evaluated. We have data pertaining to the population of stops. Each event can be computed as to its likelihood for the entire group, then sub-group comparisons can be made to that figure. For example, the base rate for receiving a moving violation was 68.6% for the population. This means that, in the entire population of atope, 68.6% of the stops ware made for a moving violation. CHAID determines whether any group (determined by mca, sex, age, residency, etc.) received moving violaUons at signif~antly higher rates than the base rate for the entire population. In this case, certain age groups of drivers (younger) were at significantly higher risk than others. This means that age is a significant factor in whether a person receives a moving violation. The NHSTA has found that, indeed, younger drivers report the highest levels of driving through stop signs without slowing, waaving back and forth between lanes, tailgating, driving through red lights, making an angry or obscene gesture or comment, cutting off another car, and driving under the influence (NHSTA Traffic Tech 186, 1999). These are all things that could result in moving violations. The Baldus Study of Iowa Cit~ Traffic Sto~)~ I prefer not to address the study to which Baldus and Woodworth refer in their critique, specifically because I do not know how that data was collected, for what purpose it was collected, or even from what time period it was collected. Although Baldus and Woodworth state that, "in 1998, we conducted such an analysis of the Iowa City stop data that were available at that time' (p. 2, para. 4), the attached charts refer to a time period from 8/1/99 - 4/10/00. I am not clear how they could analyze in 1998 data from 1999-2000. Moreover, they used very basic descriptive statistics failing to control for other variables that may have impacted the stop, examined only 2 variables (race and time of stop), used comparison data nearly 10 years old, and compared to the city/county population without having a measure of the proportion of stopped drivers actually from the city/county. This last probiem remains one of the most compelling reasons for researchers not to compare the demographics of stopped drivers to the demographics of any resident population--in Iowa City, 38% of the stopped drivers ware not residents of Iowa City, and 27% ware not residents of either the city or Johnson County. This has held true in other studies wa have conducted and makes any type of locally based, demographic proxy totally invalid. Descriptive versus Inferential Data As for Baldus and Woodworth's discussion of our 'raw data' (p. 3), one cannot examine the percentages and make inferences about relationships among variables. Dascrip~ive statistics are not inferential. Regarding point c. on page 4., Baldus and Woodworth make a good argument. It may be that drivers are being stopped for pre~.extuat reasons, resulting in an increased likelihood of warning and a reduced likelihood of citation. It also may be that officers are aware that their outcome decisions am under scrutiny and fear being accused of profiling, so they ara more likely to release Black drivers with a waming. Officer intewiews, however, indicate that decisions to cite or to warn may depend to some degree on the perceived socioeconomic status of the driver. Interviewed officers said that they were reluctant to issue a citation to a driver who they thought might have a problem paying the cost of the ticket (i.e., poorer drivers). One officer said that he considers whether the ticket might result in the person failing to pay, having a bench warrant issued for their an'est, being arrested, missing work, and enduring even greater financial hardship (personal interview, August 20, 2002). On the other hand, officers may be more likely to issue citations to ddvers who look like the cost of a traffic citation would pose no undue financial hardship (i.e., wealthier drivers). In the case of many cities, those who are in positions of greatest socioeconomic need are the minority populations. This may result in minority drivers being warned more often and white drivers being cited more often. Of course, there are several other possible explanations. In their "remarks to the city council' dated August 20, 2002, Baldus and Woodworth provide an example to illustrate the signific, ant 'flaw' in our research design. They say that, to deten'nine the impact of immaturity on auto accident rates, one would need to have the rate at which immature persons wore involved in accidents ('16-18 year olds were involved in 25% of the accidents,' page 2, para. 2) AND the proportion of immature persons among all licensed drivers ('16-18 year olds constituted only 10% of the licensed drivers'). They claim that, because '16-18 year olds constituted only 10% of the licensed drivers, but were involved in 25% of the accidents, that comparison would provide relevant evidence on the influence of immaturity on auto accident rates' (p. 2, para. 2). Them are several problems with this argument. First, one would need to know what percentage of the 16-18 year old licensed drivers is actually driving? If a greater proportion of 16-18 year olds is driving on the highways, one would expect their accident rates to be higher. How often were these drivers driving? If 16-18 year old drivers spent more time on the roadways than drivers of other ages, one would expect their accident rates to be higher. How far are these ddvem driving? If 16-18 year old drivers are driving greater distances than drivers of other ages, one would expect their accident rotes to be higher. This is the impact of mu/t/p/e var/ab/es on an outcome. Using Baldus and Woodworth's own argument, one should know the proportJon of 16-18 year old drivers in the population having accidents compared to the proportion not having accidents. Then, this proportion would be compared to figures from other age groups. If 16-18 year olds represent only 10% of licensed drivers, the implication is that they should only be involved in 10% of the accidents. This is faulty logic at its best. The proper comparison should be to compare the percentage of 16-f8 year old licensed drivers having accidents to the percentag~ of other age groups having accidents, not what proportion of the driving population they are, or in what percentage of accidents they are involved. We also would be interested in the inversc ~he percentage of licensed 16-18 year olds not having accidents compared to the percentages of other ages not having accidents. If comparing percentages is all the analysis that is required to make conclusions, then much effort is wasted by rasearchers who conduct multivariate analyses to make inferential conclusions about the relationships among variables. In fact, Professor 8aidus has wasted a great deal of time and effort in his own research on racial disparities in capital punishment. All he needed to have said was that, since Blacks comprised only 10% of a state's population, but were 25% of those sentenced to death, there was racial disparity in the sentencing. This, of course, fails to account for a defendant's prior criminal history, severity of the crime, and several other important factors that might have an impact on a parson's sentence. Current research efforts have recognized the inadequacies of population comparisons. In fact, in a review of 13 published studies on the traffic stop practices of various law enforcement agencies across the country between 1996 and 2001, Engel, Calnon, and Bernard (2002) argue that 'the mere presence of disparity in the aggregate rate of stops does not, in itself, demonstrate racial prejudice, any more than racial disparity in prison populations demonstrMes racial prejudice by sentencing judges' (p. 250). In addition, a recent publication by the Bureau of Justice Statistics concluded that racial differences in percentages of drivers stopped by the police 'am not necessarily evidence that police use race as a factor in deciding whether to make a traffic Mop--that is, not necessarily evidence of 'racial profiling" (Langan, Greenfeid, Smith, Durose, and Levin, 2001, p. 13). Although a national survey indicated that black drivers in 1999 "had higher chances than whites of being stopped at least ortce and higher chances than whites of being stopped more than once...to form evidence of racial profiling, the survey would have to show that (all other things being equal), blacks were no more likely than whites to violate traffic laws, and police pulled over blacks at a higher rate than whites' (p. 13, italics in the original). Currently, we are unable to measure racial differences in the breaking of traffic laws (or the racial distribution of violators on the roadways). It is apparent that Baldus and Woodworth realize the inadequacies of merely comparing percentages. Professor Beldus has published numerous studies examining the impact of race on various events and always has conducted fairly sophisticated multivariate analyses to reach his conclusions. It is curious that, when it came to their study of Iowa City traffic stop data, however, they were reluctant to do so. Perhaps they realized that the impact of race would he negated by other factors, such as driver sex and age. A conclusion of no racial bias could be damaging to researchers who have made careers and political allies by finding racial disparities. Finally, in response to the rest of the critique on page 4, I am not clear as to the implications of their arguments. For example, they state that blacks am 'over- represented" among consent searches, searches incident to arrest, and arrests. I am not sure what they mean by over-reprasented. Why, for example, would a group of people be arrested in proportion to their representation in the population? What theory can Baldu$ and Woodwodh present that supports their implication that arrests should be distributed in propoW~on to any representation? Many events dudng a traffic stop result fi-om preceding events. For example, property is often seized APII=R a search. These types of events are not independent, and they are not evenly distributed, but occur based on events that have previously transpired. Some events may be related to things that were not measured in the current study. For exampte, a driver may be an'ested after the officer conducts a warrant check that results in notification of an outstanding warrant for that driver. In conclusion, I am perplexed as to how Baldcs and Woodworth can critique our study on the "limited scope of the methodology' (p. 5, para. 1), especially given that their own study was purely desc~p~Jve and only included 2 variables. We used multivariate techniques and 38+ variables. We are left to wonder if Baldus and Woodworth, or any other critics of our methodology, would be leveling the same criticisms if mca had emerged as a significant predictor of events and we had concluded that ~ Iowa City Police ware engaged in discriminatory stop practices. We believe wa would not have heard a single objection to the methodology if that had been the conclusion. Peer Review This was a technical report, prepared for agency use. It was not prepared for publication or peer review in its current form. We recognize that a scholarly publication of this study would require significant revision to its format and to its content. We also recognize that the publication process involves peer review, and when wa gat to that point, wa will certainly ask for it and use whatever recommendations are made to improve the presentation of the study. Moreover, we have presented our research design and methodology at several professional conforenoes (Academy of Criminal Justice Sciences, American Society of Criminology, Southam Criminal Justice Association) where peers have had the opportunity to review and provide feedback. Our overall experience at these conferences is that others conducting research in this area are impressed with the scope of our inquiry and the depth of our analyses. We understand and value the role of peer review. Unsolicited peer review, however, of the type that Baldus and Woodworth provide not only is unprofessional, but also is insulting. I, or any of my colleagues at the University of Louisville who worked on this study, would have been more than happy to speak with Baldus and Woodworth and to address any concerns in a mom private forum. The type of last minute presentation that occurred at the Council's work group meeting on August 19 is mom akin to guerilla warfare in defense of personal and pol~ agendas than to legitimate professional I am happy to have been able to represent my research team and the University of Louisville. I appreciate the willingness of the City Council to allow me to present and clarify the study that we conducted, and to present this written response to Baldus and Woodworth's critique. However, this shall be the last I speak or write as far as responses are concerned. You all should be proud that Iowa City has a proactive police department that was willing to examine its stop practices before them ware any problems. This is not often ~ case. I have continued to be impressed with the professionalism and the dedication to community service that I have seen exhibited by Chief Winkalhake and the other members of the Iowa City Police Department. If I can be of further assistance to you or I can clmify any of the material presented in this letter, please do not hesitate to call me. I have made business cards with my contact information available for your convenience. I also look forward to continuing our work with the ICPD and with the City of Iowa City in studying another year of traffic stops. University of Louisville DepaJt~ent of Justice Administration B~igman Hall, 2'~ Floor Louisville, KY 40292 References Engel, R~S., Calnon, J.M., and Bernard, T.J. (2002). Theory and racial profiling: Sho[rtcominga and future dire~on$ in research. Jusl;ice. C~uarterlv, 19 (2), 249- 274. Langan, P.A., Greenfeld, L.A., Smith, S.K., Durose, M.R., and Levin, D.J. (2001). Contacts between oolite and the oublic: Findings fi.om the 1999 National ~urvey (NCJ 184957). Washington, D.C.: Bureau of Justice Statistics. National Highway Traffic Safety Administration (1999). Speeding and aggressive driving documented in national telephone suntey. T~;; Te~h Number 186. Washington, D.C.: U.S. Department of Transportation. l0 Fi~e 1: All Drivers Figure 2: Stopped Drivers Among Alt Drivers [~ Abidcr$ not stopped (no discrim.) ~ Abiders stopped (~ossibl¢ rli~rin~) I Violators stopped (no discrim.) ~ Violators not stopped (possible diserim.) October 8, 2002 To: City Council of Iowa City From: David Baldus and George Woodworth Re: Iowa City Police Stop Study by Angela West, University of Louisville We appreciate your sending us Angela D. West's August 27, 2002 letter to Chief Winkelhake responding to our August 18 and 20, 2002 critiques of her police stop study. Dr. West's letter presents a number of reasons why the conduct of police stop studies is difficult. However, her letter does not in anyway question our assertion that her June 13, 2002 study ~ "Traffic Stop Practices of the Iowa City Police Department: April 1- December 31,2001 "- simply does not address the issue of whether race was a factor in the initial traffic stops. This is also what she said in her remarks to city council at its August 19, 2002 work session, i.e., "On the basis of our study, one simply cannot tell if race is a factor in the initial decision to stop motorists." It is for this reason that the claim of the police chief and others that the Louisville study demonstrates that race is not a factor in the initial decisions to stop motorists has no support whatever in that study and is a complete misrepresentation of what it does contain. Submitted by Vice-Chair Loren Horton Profiling po' ce Datashow ~owam~ officers don't traffic stops discriminate ~' ~' ~' aa, 2~s By Vanessa Miller Iowa City ~ss-Citi,~ The Iowa City Po.ce mle~ ~. ~e study~b~d on eve~ ~c-~lat~ con~t ~om Apffi 1 m Dec. 31, 2ffil. Male Ferule p~ the ~m of ~e ~ fl~e driver, police officer ~d ~phics nord ~ ~e and vehicle re~stmtion. Offiy ~e badge nm~r Stop event infomafion h~cludes d~e, time, rein R.J. ~e~e. du~, de~ on ~y pm~ one ~du~ co~d e~ se~, force ~ . O~e ou~ome of ~e stop. ~e d~ show~ ~ no cen ~ we n~ W ~n~ue by the police dep~ent ~d on ~ ~o~ ~1 on r~e or much of from 9,702 in.motions anylhing for that matter," Iowa City Press-Citizen: Opinion Page 1 of 2 Tuesday, August 20, 2002 Submitted by Vice-Chair Loren Horton Continue traffic-stop monitoring We're glad the Iowa City Police Department plans to continue studying its traffic-stop data to determine if race plays a role in pulling over drivers. Although studies like this cost money, and initial data indicates there is no THE ISSUE: correlation between lowa City traffic Iowa City Police get stops and a driver's race, it's an an initial thumbs-up in exercise worth having and continuing, preliminary analysis of traffic-stop data, An initial study of eight-months' worth which was analyzed of'traffic data, from April 1 to Dec. 31, for possible racial 2001, shows that Iowa City police did profiling. not systematically engage in a practice of pulling over drivers based on their WE SUGGEST: skin color. Continued data collection and The details of that analysis can be analysis, and better tricky to interpret, however, record-keeping, will make the data more For instance, of more than 9,000 useful. recorded cases in which police were involved in a traffic-related contacts, 84 percent of people pulled over were white. In comparison, Iowa City's population is 87.3 percent white; Johnson County's is 90.1 percent white. According to the study's authors, from the University of Louisville, it means very little to compare the number of blacks cited in traffic stops to the number of blacks in the local population, however. The study's writers say we should instead be comparing the citation statistics to the number of blacks or non-white people in the pool of drivers eligible to be pulled over. That's what makes the continuation of this data-collection analysis so What do important. The first round of data collection creates a baseline to yOU think? which we can compare data yet to be collected. · Should the police continue to track Police Chief R.J. Winkelhake says trafficstop data? he is pleased by the findings, as he ° Send your comments should be. to Opinion Page, P.O. Box 2480, Iowa City, "The data showed that no systematic Iowa 52244; fax to (319) action was taken by the police 834-1083; e-mail to department based on race or much opinion~ press- of anything for that matter," he said. citizen.com. "That does not mean, however, that http://www.press-citizen.com/opinion/pceditorials/staffedit082002.htm 9/11/02 Iowa City Press-Citizen: Opinion Page 2 of 2 any one individual could not have been subject to this activity (profiling) by an individual officer. So we need to be conscious." We agree. If these statistics are a true reflection of reality, the department should keep up the good work. However, the next step is to collect a year's worth of "clean data" (without inputting errors), as the study writers suggest. As Winkelhake and the study's authors both suggest statistical analysis alone will not reveal whether any individual officer is engaging in racial profiling. Administrative supervision and community oversight is best suited to ferret that out. in 1999, Iowa City Police were the first in the state, and among the first in the country, to voluntarily begin collecting this data. Let's not lose that momentum. This is one of those cases where good enough is not going to be enough. http://www.press-citizen.com/opinion/pceditorials/staffedit082002.htm 9/11/02