Jury Representativeness – Part 2

Last week’s blog, Jury Representativeness – Part 1, presented the argument for the use of the “representativeness” of the final juror pool as a measure of fundamental fairness, equality and integrity. This second part of the blog is an operational definition of the measure using the same straightforward format as that used to describe the ten measures of the CourTools.


The comparative parity (i.e., the absence of disparity) -- expressed as a percentage -- between the representation of minority groups in the population and the representation of the same groups in the final juror pool or venire.


How well juries mirror the community from which they are drawn is widely considered a reflection of the equality, fairness and integrity of our justice system. In the past, juries were often composed of individuals selected from certain strata of the population. Bias or discrimination, intentional or not, was usually the result. Verdicts reflected the community standards of these strata, and the viewpoints of juries rarely were those of the entire community. For example, it was only in 1975, in Taylor v. Louisiana, 419 U.S. 526, that the U.S. Supreme Court held that women could not be excluded simply because they were not men.

Courts cannot guarantee that juries will always reach decisions that are fair and equitable. Nor can they guarantee that juries chosen through voir dire are a fair cross-section of the community. Courts can, however, provide significant fairness, equality, and integrity by ensuring that the prospective jurors who actually report to the courthouse to make themselves available for jury duty -- the final juror pool -- represent the community. This measure gauges the representativeness of the final juror pool in terms of its racial or national origin composition -- the source referent -- and compares it to that of the community, i.e., the degree to which the race and national origin of the persons in the pool or venire represent that in the population of the jurisdiction -- the target referent. It is expressed as a ratio of the absolute race/national origin disparity divided by the by the percent of the racial/national origin representation in the community. For example, a jury pool would be perfectly representative (i.e., 100% comparative parity) if both the jury pool and the population of the jurisdiction contained 60% White, 20% Black or African American, and 20% Hispanic or Latino.

Arguably, identifying a combination of demographic characteristics as the source referent -- including gender, age, income level, and education -- may be better than looking at race and national origin alone to determine jury representativeness. However, because race and national origin are the most salient demographic characteristics in law, policy debate, and jury utilization and management, disparities in race and national origin may be seen as reflective of disparities in other demographic categories. That is, if the jury pool fairly represents the community in terms of racial and national origin, most other demographic characteristics are likely to be represented fairly as well.


Planning and Preparation

This measure requires tapping two sources of information for two data elements -- the racial makeup of the geographic area corresponding to the court's jurisdiction and the racial and national origin composition of the jury pools assembled by the court.The website of the U.S. Department of Commerce, Bureau of the Census, maintains "demographic profiles" for the most recent census data by geographic region, including race and national origin of persons 18 years of age or older (http://www.census.gov). The available data are extensive, including population estimates by age, race, gender, national origin for each county by Zip code. The Census Bureau identifies six racial categories and a category for two or more races. In addition, the Census Bureau tracks Hispanic origin (Hispanic or non-Hispanic). This is treated as a separate dimension from race. Extrapolations of these data for the period between the census years are prepared by the Bureau of the Census and by local units of government such as planning commissions. In addition to assistance with the website and its contents, a Census Data Center in each State (usually at one of the major universities) provides access to and assistance with census data and other statistics or data sources.From this source, courts can obtain the number and percent of juror-eligible individuals in eight categories of race and national origin: White, Black or African American, Hispanic or Latino (of any race), American Indian and Alaska Native, Asian, Native Hawaiian and other Pacific Islander, "mixed" and "some other race" in their jurisdiction. Courts with jurisdictions encompassing geographic areas that do not correspond to political boundaries or Zip codes may need to make some adjustments and estimates based on Bureau of the Census information.The second data source that needs to be tapped for this measure is the racial and national origin composition of the jury pool. (The group of prospective jurors reporting for jury duty in a given term who are awaiting assignment to a panel for voire dire and selection to sit on a jury. In a one-judge court, a pool and panel are likely to be the same.) This source of information may be recorded by prospective jurors on qualification questionnaires or summonses, or questionnaires distributed to the prospective jurors in the pool.(Note: An alternative approach to measuring jury representativeness uses Zip code representation as a proxy for representation by race and national origin. It should be considered by those courts wanting to measure jury representativeness that are unable or unwilling to gather race or national origin data for practical or political reasons (e.g., fear of providing ammunition to those challenging the jury selection process). This approach, which has diagnostic value but may not be dispositive in legal challenges, focuses on the difference in percent of the general population living in particular Zip codes and the percent of the pool living in those Zip codes. The difference in the percents of Zip code representation in the population versus the Zip code representation in the pool can be used as a proxy for comparative race disparity in the preferred measure of jury representativeness. For example, in a jurisdiction covering ten Zip codes with a minority population living mostly in three Zip codes, a "Zip code disparity" of 24 percent would mean that there is a disparity of 24 percent between the population in the three "minority" Zip codes and the percent of the prospective juror pool that lives in the three "minority" Zip codes.)

Data Collection

Statistical Population. First, identify the racial or national origin make-up in the jury-eligible community (e.g., 18 years of age and older). This is known as the statistical “population.” As noted above, the best source of this information is the Bureau of the Census. For example, the makeup of a jurisdiction may be 66.9 percent White, 28.5 percent Black or African American, 2.0 percent Hispanic or Latino, 1.8 percent Asian, 0.1 percent Native Hawaiian or Other Pacific Islander, 0.7 percent some other race, and 1.6 percent reporting two or more races.Second, to calculate a single value for the core measure, identify the three race-national origin categories with the greatest representation in the jurisdiction. Continuing with the preceding example, 28.5 percent of the jury-eligible population is Black or African American, 2.0 percent is Hispanic or Latino, and 1.8 is Asian. Defining the source referent in terms of three dominant minority groups -- instead of the single most dominant or all eight groups used by the Bureau of Census -- is to a large degree arbitrary based on considerations of practicality and utility (similar to those most likely governed the Bureau of Census' identification of only eight categories). Similarly, the method of picking the number of race and national origin categories is arbitrary. Rather than simply choosing the three most dominant groups, an alternative method is to choose, for example, the number of groups that constitute 90% of the minority population in the community, a method that would have yielded the three groups chosen in the example in the text. (Using 95% would have included the 1.6% identifying themselves in two or more categories.)

Statistical Sample. Next, collect data on the jury pool -- the statistical “sample”. The jury pool is composed of individuals who report for jury duty and who are qualified to serve. In smaller jurisdictions, the jury pool or venire is likely to be the same as the "panel" from which a specific jury is drawn, i.e., the court may be seeking jurors for one jury trial only. In larger jurisdiction, multiple panels are drawn from the pool to serve multiple jury trials. Whether through qualification questionnaires, summonses or pool questionnaires, the court must be able to identify the racial identity or national origin of individuals in the jury pool. The preferred method is collecting race and national origin data by a questionnaire distributed at the time the jury pool is assembled. The questionnaire may contain information about obtaining juror fees, parking cost reimbursement and other items, thereby increasing the likelihood that prospective jurors will disclose their race or national origin. Although demographic questions may be optional, a note or explanation such as the following adopted from the State of New York may elicit more responses:

Information on race is sought solely to allow the court system to monitor the jury selection process to ensure that no discrimination is occurring in that process and that jurors are being randomly selected from a fair cross-section of the community. This information has absolutely no bearing on qualification for jury service.

To obtain the statistical sample, first note the days when a new jury pool reports to the courthouse (e.g., first Monday of the month). Second, record the total number of qualified prospective jurors in the pool on those days. (Some prospective jurors will be disqualified or exempted only after they report to the court for jury duty.) Third, record the total number of qualified jurors in the jury pool who classified themselves in each of the three minority groups or, alternatively, the Census categories that make up at least 95 percent of the population. To extend the example, assume that a jury pool of 256 individuals includes 49 Black or African American (19.1 percent of the pool), four Hispanic or Latino (1.6 percent), and two Asians (0.8 percent).

Data Analysis, Interpretation and Use

Analysis of the race and national origin data drawn from the population and the pool is straightforward. First, calculate comparative disparity for each of the three identified "minority" categories using the following formula:

Comparative Disparity = [(% Population - % Pool) / % Population] x 100

Comparative disparity is the absolute disparity -- the difference between the representation in the population and the pool -- divided by the percent representation in the population, expressed as a percent by multiplying this number by 100. The indicator of comparative disparity compensates for the limitation in the indicator of absolute disparity by relating disparity to the size of the underrepresented group in the population.

Inserting the values of the examples noted above into this formula yields the following comparative disparities. For Black or African American, the comparative disparity is [(28.5 - 19.1) /28.5)] x 100 = 33.0 percent, for Hispanic and Latino it is [(2.0 - 1.6) /2.0] x 100 = 20.0 percent, and for Asians it is [(1.8 - 0.8) /1.8] x 100 = 55.6 percent.

Reduce the three disparities to a single value -- the aggregate comparative disparity -- by calculating the mean of the three comparative disparities. Reducing the measure of representativeness to a single number, the aggregate of the three comparative disparities, is consistent with the definition of core performance measure, i.e., a primary (or super ordinate) performance measure, preferably expressed as a single value, that is aligned with, but distinguishable from, subordinate measures in a hierarchy of measures. The larger the comparative disparity, the more unrepresentative is the jury pool. For the example, the aggregate comparative disparity is 36.2 percent or (33.0 + 20.0 + 55.6) /3.

Finally, as suggested by the operational definition of the measure, it may be advantageous to report and discuss Jury Representativeness in positive instead of negative terms, that is, in terms of parity instead of disparity. The value of comparative (race and national origin) disparity for the measure of Jury Representativeness -- which decreases as the court improves on the measure (i.e., less disparity is better) -- is changed to parity and converted into a number that increases as the court gets better. For our example, a comparative disparity of 36.2% is subtracted from 100% to yield 63.8% comparative parity (100% - 36.2% = 63.8%).

No Benchmarks, Yet

Of course, absent standards, benchmarks or norms, a single value of Jury Representativeness has limited utility beyond establishing a baseline. Is the comparative parity of 63.8 percent in our example good or bad? However, the measure gains considerable value for purposes of control, diagnosis, ascertaining trends, and planning with (1) repeated measurement (Jury Representativeness improves or deteriorates), (2) establishment of control levels (i.e., established levels of parity that will trigger corrective actions), and (3) comparisons against standards, benchmarks, or norms.

Unfortunately, no established standards of comparative disparity exist to set norms or benchmarks for court performance. One source suggests a maximum comparative disparity value of 15 percent, another suggests 43 percent and below. See D. Kairys et al., “Jury Representation, A Mandate for Multiple Source Lists,” 65 California Law Review 776-827 (1977) [15 percent]; Measter et al., “Getting a Fair Cross-Section of the Community,” Forum 14-21 (1989) [43 percent]. Numerous legal challenges of the composition of jury pools or venires -- i.e., that the composition of the pool is not fair and reasonable in relation to that of the community -- have explored the "constitutional limit of permissible disparity" and comparative disparities of at least as high as 45.2%. These challenges were not sustained, however, because they failed to meet the third element or prong of the Duren test (Duren v. Missouri (1979) 439 U.S. 353, 364) requiring that even these seemingly large comparative disparities resulted from systematic exclusions in the jury selection process that are the cause of the disparities. In other words, even large comparative disparities do not by themselves establish a prima facie case. They must be accompanied by a showing that the jury selection processes used by the court are by design not race-neutral. Evidence that a race-neutral selection process produces de facto exclusion of a class of jurors is insufficient. (See People v. Aldridge Currie, No. A084426 (Contra Costa County Super. Ct. No. 962407-3, filed January 30, 2001) (2001 Daily Journal D.A.R. 1963); Rhode Island v. Tremblay, P1 No. 97-1d816AB (Providence Super. Ct., filed March 19, 2003).

For the latest posts and archives of Made2Measure click here.

Copyright CourtMetrics 2006. All rights reserved.

Popular posts from this blog

A Logic Model of Performance Inputs, Outputs and Outcomes

Q & A: Outcome vs. Measure vs. Target vs. Standard

Top 10 Reasons for Performance Measurement