CBA Record November 2018

from punch cards to optical-scan and touch-screen voting systems. More research is needed. As for the last ballot position, not only did I find no evidence of a last-position advantage, the calculations produced a negative estimate of roughly four percent- age points. However, this four-point figure comes with an asterisk, because the analysis did not include other ballot positions besides the last, and so it is unclear which other positions are more or less disadvan- taged relative to the first position. But the finding of a negative value certainly throws cold water on the idea of any sort of last- position advantage A Second Approach One limitation of the type of statistical analysis that I employed is that it struggles with variables that are closely related to one another, such as different ballot positions. A vote cast for one position is by defini- tion not cast for the others, and this sort of interrelationship can distort statistical estimates. So to take a bigger-picture look at all of the ballot positions simultaneously, I tried something very different. The tables nearby show the results of my efforts. While the tables may look impos- ing, their numbers are nothing more than simple tallies. For every primary contest since 1980 I tallied the winning candidate’s ballot position, and grouped the tallies by contest size. So, for instance, the second table in the group shows that there have been 114 contests with three candidates, with the first position winning 51, the second position winning 42, and the last position winning 21. These tables lack the statistical rigor of a more formal analysis, but simply by virtue of their including more than 500 contests, they display clear patterns. And the patterns are nothing short of remarkable. The obvious conclusion that the tables all but shout out is that higher is better. They show that the first position has produced the most victories overall, but also that the other positions are far from equal. Almost without exception, moving down the ballot produces fewer win- ning candidates. This is less clear in the larger tables, where there are fewer contests to tally, but even there the pattern is unmistakable.

Cook County Judicial Primaries 1980-2018: Numbers of Winning Candidates, by Ballot Position

2 candidates

7 candidates position 1 position 2 position 3 position 4 position 5 position 6 position 7 position 1 position 2 position 3 position 4 position 5 position 6 position 7 position 8 position 1 position 2 position 3 position 4 position 5 position 6 position 7 position 8 position 9 position 10 8 candidates 9+ candidates

position 1 position 2

99 71

7 5 2 3 2 2 2 3 4 1 1 3 2 2 0 4 1 1 3 2 2 0 2 0 1 0

170

3 candidates

position 1 position 2 position 3

51 42 21

23

114

4 candidates

position 1 position 2 position 3 position 4

30 24 17 15 86

5 candidates

16

position 1 position 2 position 3 position 4 position 5

18 17 13 12

2

62

6 candidates

position 1 position 2 position 3 position 4 position 5 position 6

10

6

11

6 5 1

lower

39

last position

0 16

The updated analysis confirmed my ear- lier results on the first-position advantage. Across the entire time period the advantage was roughly 4.3 percentage points for countywide contests. I also found more substantial evidence of an increase over time: for the 1980-1998 period the advan-

tage was 3.2 percentage points, compared to 5.9 points for the 2000-2018 period. There is no obvious explanation for this increase. It may reflect an electorate less informed about judicial candidates and more easily influenced by ballot cues; or it may somehow be connected to the switch

28 NOVEMBER 2018