PoliticiansTheoriesOfVotingBehavior
Politicians’ Theories of Voting Behavior
Politicians’ Theories of Voting Behavior (DOI: https://doi.org/10.1017/S0003055424001060) was written by Jack Lucas, Lior Sheffer, Peter John Loewen, Stefaan Walgrave, Karolin Soontjens, Eran Amsalem, Stefanie Bailer, Nathalie Brack, Christian Breunig, Pirmin Bundi, Linda Coufal, Patrick Dumont, Sarah Lachance, Miguel M. Pereira, Mikael Persson, Jean-benoit Pilet, Anne Rasmussen, Maj-britt Sterba, and Frédéric Varone in 2024. It was published in the American Political Science Review (vol. 119, no. 3).
Politicians' views on voting behavior matter. Those views inform which policies are given priority. (i.e., a theory of behavior that looks to short-term goals leads to priority on short-term policies) If there is a mismatch between their views and actual behavior, then there will be inefficiencies in the representative democracy.
Politicians' views can also reorient the electorate by recasting expectations around candidates. (i.e., a focus on short-term policies from one official can lead to expectations that all subsequent candidates will compete similarly) In other words, there is dynamism between politicians' views and citizens' views.
The authors designed 8 questions that put two opposing world views forward, then ask "Where would you position yourself in this debate?" on an 11-point scale.
"Some say that voters make their decisions based on their policy preferences. Others say that voters’ choices have much more to do with their deeply held partisan or other group identities..."
Some say that voters are impatient and think about the short term when they vote. Others say that voters focus on the long term..."
"Some say that when citizens vote they are by and large knowledgeable about political issues, while others say they generally know very little..."
"Some say that voters make voting decisions based on one or two policy issues they care strongly about. Others say voters decide based on a wide range of policy issues..."
"Some say that voters care more about the ideas parties stand for than about the party leader’s character and competence. Others say that voters care about the leader's qualities more than the party’s platform..."
"Some say that voters make decisions based on candidates’ policy commitments and promises for the next term. Others say that voters base their decisions on rewarding or punishing their elected representatives for how well they have performed in the previous term..."
Some say that voters judge governments on whether they’ve improved everyone’s lives. Others say that voters judge governments on whether they’ve improved their own personal lives..."
sociotropic vs egocentric
- "Some say that voters often blame or reward politicians for events that are totally outside the politician’s control. Others say that voters are good at knowing which events politicians are and are not responsible for..."
blind vs clear-eyed
The battery was included in the POLPOP project and fielded from Feb 2022 to Mar 2023. This forms the elected official survey sample.
- CAPI survey
- 982 elected national officials responded
- In most countries, all members of the national legislature were invited. In federal states, members of regional legislatures were also invited.
11 countries: Australia, Flanders, Canada, Czechia, Denmark, Germany, Israel, the Netherlands, Portugal, Sweden, and Switzerland
The POLPOP teams responsible for Francophone Belgium, Norway, and Luxembourg declined to include the full battery in their interviews.
Note that Australia, Israel, and Sweden held elections so the sample included some members who lost re-election.
Response rates vary widely across countries. The biggest outlier is Belgium, where the very high response rate (85%) leads to an outsized proportion in the analytic sample (23%). Compare to Canada, with a response rate of 12%, being just 9% of the analytic sample.
A general population sample was also collected.
- Web survey using Dynata panel
- Same countries; quotas for each country were set by age, sex, and education.
- Within each country, about half of respondents were randomly assigned the battery of questions above.
- 12,082 responses total; about 2,000 from each country
Comparing the distributions between the elected officials sample and the general population sample, some differences are quickly evident. More importantly, politicians are much more homogenous (thinner distribution) than gen pop.
The authors regress a politician indicator on each of the 8 dimensions using the pooled survey responses. Country level fixed effects are used. They infer from the significance of coefficients whether there is a significant difference. 6 of the 8 are counted as being significant. These analyses are then repeated within each country to arrive at country-specific significant differences.
The most country-robust differences are that elected officials are more likely to believe in...
- blind retrospective voting
- in short-term policy goals
- and in single-issue voting
These are also the three measures with the largest absolute coefficients in the pooled analysis.
The authors then use latent class analysis to fit elected officials into some number of categories representing disparate theories of voting behavior. They test with 2-20 classes, landing on 4. The first two are of actual interest.
- Democratic optimists "tend to think of voters as fair in their retrospective assessments, policy oriented, prospective, sociotropic, multiple-issue-focused, interested in policy rather than political leaders, knowledgeable, and oriented to the long term."
- Democratic realists "tend to see voters as unfair in their blame, identity oriented, retrospective, egocentric, single-issue-focused, leader-driven, ignorant, and short-termist."
- Undecideds give moderate responses across the board, suggesting no strong opinions (or unwillingness to divulge).
- Inattentives give high responses across the board, suggesting inattention to the survey or acquiescence bias. Practically no elected officials are inattentive.
Overall, elected officials skew heavily towards democratic realism (73%; compare to 16% being optimists). The skew is most extreme in Canada (82% vs 6%), although it was not a clear outlier. Citizens are more evenly spread.