Ho, Daniel E. and Kosuke Imai. (2008). ``Estimating Causal Effects of Ballot Order from a Randomized Natural Experiment: California Alphabet Lottery, 1978-2002.'' Public Opinion Quarterly, Vol. 72, No. 2 (Summer), pp. 216-240.

Randomized natural experiments provide social scientists with rare opportunities to draw credible causal inferences in real world settings. We capitalize on such a unique experiment to examine how the name order of candidates on ballots affects election outcomes. Since 1975 California has randomized the ballot order for statewide offices with a complex alphabet lottery. Adapting statistical techniques to this lottery and addressing methodological problems of conventional approaches, our analysis of statewide elections from 1978 to 2002 reveals that in general elections ballot order significantly impacts only minor party candidates, with no detectable effects on major party candidates. These results contradict previous research finding large effects in general elections for major party candidates. In primaries, however, we show that being listed first benefits everyone. Major party candidates generally gain one to three percentage points, while minor party candidates may double their vote shares. In all elections, the largest effects are for nonpartisan races, where candidates in first position gain three percentage points.

Here is an earlier version with a full appendix that includes additional tables.

You might also be interested in Ho, Daniel E. and Kosuke Imai. (2006). ``Randomization Inference With Natural Experiments: An Analysis of Ballot Effects in the 2003 California Recall Election''. Journal of the American Statistical Association, Vol 101, No. 475 (September), pp.888 - 900.