Guest Post: Dr. Klumpp evaluates the March primary

by Albert J. Klumpp

For most of the period leading up to the March 17 primary, much of the buzz surrounding our judicial contests was about the large group of female candidates with Irish names, and the proven tendencies of some voters to favor females and Irish names. In the final days, the focus turned to coronavirus and how much of an impact it would have on voter turnout. But in the end, neither was much of an issue, and the primary was mostly business as usual. There was no Bonnie Lass Invasion, and the virus had much less of an impact than expected---although with an unexpected twist.

Voter turnout

Taking on faith the estimate of 80,000 mail votes still to be counted, turnout for the primary will end up at roughly 33% countywide. Figure 1 puts this figure into historical perspective—and shows that the turnout was very respectable compared to past primaries.

While there was a competitive majority-party presidential contest, the contest had a feel of inevitability to it by the time it reached Illinois, and its intensity was nowhere near that of the 2008 or 2016 primaries when presidential candidates with local ties pushed interest and turnout to very high levels. And while there were certain local and congressional races of significance, overall the contests and issues were not such a compelling group that a lower turnout would have been a surprise. Throw in a pandemic scare and 33% is actually an impressive number.

But for our purposes, there was no particular characteristic of the turnout that had any unusual impact on the judicial contests. For instance, the dropoff-versus-completion rate for the judicial section of the ballot was consistent with previous primaries:

And let’s put to rest the claim that the turnout was racially disproportionate. Figures 3 and 4 cover the 2012-2020 primaries and show the relative proportions of the Democratic vote and total vote, respectively, coming from black-majority, Hispanic-majority, white-majority, and mixed-race wards and townships. Both figures show that the 2020 proportions were virtually unchanged relative to 2016 and 2018, and that no one group turned out in disproportionately high or low numbers this year.

No more one-trick pony

For some time now the “name game” in judicial contests has been evolving as the county’s demographics have changed. Irish names have slightly less value than in past decades, and names common to African-Americans (based on US census data) have gradually become nearly as advantageous as Irish names. And while in the past there had been too few Hispanic names on the ballot to evaluate statistically, the 2018 primary showed for the first time that Hispanic names can generate a substantial advantage of their own.

The March 17 results show these trends continuing. In fact, statistical estimates of name values produced for the first time ever a higher value for African-American names (8.9 points per contest) than for Irish names (7.2 points). These estimates have margins of error and in this instance cannot be called statistically different, but they do add to the evidence of a gradually shifting electorate. And Hispanic names did even better, with a per-contest bump of more than 12 percentage points.

The Irish-named candidates also hurt their own cause by virtue of their sheer numbers. In many contests multiple Irish names competed against one another and subdivided the name advantage rather than monopolizing it. This worked to the benefit of opponents--black, Hispanic, even Jewish—who were not competing against similar opponents and did not subdivide whatever advantage(s) they held.

Other factors

Sorting through the other major variables:

• Gender remains a high priority among Democratic voters. The pro-female vote was roughly 15 points per contest, consistent with its long-term average. However, more than half of the judicial candidates on the 2020 ballot were female, and in many contests the gender advantage was subdivided. Several male candidates benefitted from this.
• Recommendations from the Chicago Tribune, Chicago Bar Association and Chicago Council of Lawyers were worth a combined 9.5 percentage points—a lower amount than in most years. It’s likely that the sheer volume of coronavirus coverage in the media had a crowding-out effect against the dissemination of bar rating information and against its reception by the electorate. This would explain the smaller number. (The Chicago Sun-Times, for instance, gave less attention to bar ratings than usual.)
• On the other hand, candidates slated by the county Democratic party received a bump of roughly 11 percentage points. This is slightly higher than the long-term average, and considering the severe drubbing of the party establishment’s mayoral candidate just last year, might seem surprising. But the party’s two slate-card mailings were sent with their usual impeccable timing, and with alternative information being less publicized, it may be that slating information partially filled a void. Purely speculation but plausible.
• The first ballot position gave an advantage of roughly 6.7 percentage points relative to a non-first-position candidate. In a 2018 published article I explained how this is a bit of an oversimplification and that any higher position is better than any lower position. But as a single metric, the 6.7 figure is consistent with recent past primaries.
• Campaign spending figures aren’t yet available and so will be examined in a future post. However, spending in countywide judicial contests historically has almost no detectable effect on vote totals, and in reviewing preliminary spending reports there is no reason to expect anything different this time.

The subcircuits

Subcircuit contests are more difficult to analyze because every subcircuit has unique characteristics and each has only a small number of contests to study. But overall the subcircuit results to add further support to the crowding-out idea offered above, with local endorsements and candidate evaluations appearing to be less influential than usual relative to ballot cues. For instance, eleven of the twelve contests with at least one female and one male were won by a female, and fifteen of the sixteen Democratic contests were won either by a female or by a male in the first ballot position. And several candidates who diligently rounded up local party endorsements appeared to gain fewer votes than expected from them.

*   *   *   *   *   *   *

Two final bits of fine print. To reiterate, the statistical estimates cited here do have margins of error, but all are considered highly statistically significant. Also, the estimates might change slightly when the remaining mail votes are counted, but unless those ballots heavily over- or under-represent different categories of voters, any such changes will be trivial.

----------------------------------------------

Albert J. Klumpp has been a generous and frequent contributor to FWIW over the years. A research analyst with a public policy PhD, Klumpp is the author of several scholarly works analyzing judicial elections including, most recently, "Campaign Spending in Cook County Judicial Elections," CBA Record, Nov.-Dec. 2019 (p. 30)