The Quality of Our Jury System Depends on the Quality of Our Data

As someone with a commonly misspelled name, without a home of my own, who occasionally fails to register to vote and has not renewed her driver’s license and six years, I now understand why I have never been called to jury duty. According to the judiciary, I am off the grid!

In their book, American Juries: The Verdict, Hans & Vidmar demonstrate that a diverse and representative jury promotes better fact-finding, decreases the impact of bias, and offers more legitimacy for the resulting verdict. While the constitution prohibits excluding groups from the jury systematically (via the Sixth Amendment) or intentionally (via the Equal Protection Clause), juries have historically fallen short of representing a fair cross-section of the community. The judicial system has gradually adopted the idea that the best way to ensure a fair and impartial jury is to ensure a diverse jury pool from which to select juries. The first step in jury selection—compiling the master list (also known as a “jury wheel” or “master wheel”)—is thus key to assembling a diverse jury that represents a fair cross-section of the community. (See Principle 10(A) “The jury source list and the assembled jury pool should be representative and inclusive of the eligible population in the jurisdiction”).

The master jury list (also known as the "jury wheel" or the "master wheel") is often sourced from voter registration lists, supplemented with other lists. Voter registration lists are readily available, frequently updated, and collected in districts within jurisdictional boundaries. However, basing the master jury list on voter registration can lead to excluding the poor, the young, racial minorities, and the less educated. Because of these limitations, many jurisdictions use other sources, including telephone directories, utility customer lists, or driver's license lists, to compile a master list. Theoretically, additional lists find additional potential jurors not on the voter registration or supplemental list. A brief review of a fifty-state survey indicates that as these lists merge, duplicates are removed.

Merging databases will create duplicates. Additionally, all databases have duplicated entries, and sometimes they can be difficult to identify. Hans & Vidmar briefly note that “sometimes there are problems combining the lists and removing duplicate names.” Neil Vidmar & Valerie P. Hans, American Juries: The Verdict 76-77 (1st ed. 2007). For example, names may be spelled differently for different contexts (‘Elizabeth’ vs ‘Beth’) or misspelled (‘Sara’ vs. ‘Sarah’). If there were only a few hundred records, one might be able to match the names manually, comparing every string with the other strings and selecting the right one. However, with datasets intended to represent entire populations of counties or judicial districts, organizations will often purchase commercial or develop rudimentary algorithmic systems in order to cleanse and match data. An efficient system will identify acronyms, name reversal, name variations, phonetic spellings, deliberate misspellings, inadvertent misspellings, insertion and removal of punctuation / spaces / special characters, different spelling of names (e.g. ‘Sara’ instead of ‘Sarah,’ ‘Jon’ instead of ‘John’), shortened names (e.g. ‘Elizabeth’ matches with ‘Betty,’ ‘Beth,’ ‘Elisa,’ ‘Elsa,’ ‘Beth,’ etc.), and many other variations.

While additional research on specific methods may reveal that certain jurisdictions effectively maintain a source list that in fact represents a fair cross-section of the community, my personal experience indicates otherwise. Improving the source list by aggregating additional lists, as well as enhancing data quality and cleansing methods, should yield a more accurate representation of the community, thus, not only meeting, but living up to, our constitutional values.