Apr 212021

How important are social incentives for turnout?

Since the earliest election studies, we have assumed that social networks are important for turnout, and that voting more generally is a form of social behaviour (and also a habit). While there are heaps of data to support these ideas, this study from Denmark is quite something. Thanks to Scandinavia’s love for efficiency and science, as opposed to her glorious disregard for data protection and stuff, the authors are able to link panel data to fine-grained information on when people living in the same household entered their local polling station in various elections. Yes, that’s right.

Bhatti, Y., Fieldhouse, E., & Hansen, K. M. (2020). It’s a Group Thing: How Voters Go to the Polls Together. Political Behavior, 42(1), 1–34. http://dx.doi.org/10.1007/s11109-018-9484-2

What we liked

The data: we observe actual behaviour, over years. We know where people live, and with whom. Plus, we also have socio-demographic information on them. And there are so many of them. Wow. Just wow.

What we were reading: Voting as group behaviour 1

What we did not like so much

Students did not at all worry about data protection. In this regard at least, my work on this plane is done. Otherwise, they were quite critical. They said that the paper was a bit short on theory (wow again), and that they wanted a better explanation for the “potential voting companion” effect. They also argued that local and EU elections were not necessarily comparable and worried about some other (rather specific) points. But in the end we agreed that this is seriously cool stuff.
Apr 172021
Family parking

Is there a political gender gap amongst young Germans?

Gender gaps are everywhere, but there are some places where they are less likely. According to the authors, young Germans represent a least likely case for gender differences in political activity: female levels of educational attainment are actually higher than male ones, and adolescence is a time (for many) before those aspects of family life kick in that tend hold women back.

But as it turns out, “least likely” is actually “not determinstically likely, but very likely”. The authors find substantial differences in the socialisation of young men and women, which are linked to rather dramatic differences in internal efficacy and self esteem. They also uncover evidence that men benefit more from factors that contribute to political activity, particularly in institutional settings. Intriguing, depressing, hardly surprising.

Pfanzelt, H., & Spies, D. C. (2019). The Gender Gap in Youth Political Participation: Evidence from Germany. Political Research Quarterly, 72(1), 34–48. http://dx.doi.org/10.1177/1065912918775249

What we liked

Students were over the moon. They praised the clear structure, the strength of the theoretical argument, the clear exposition and the empirical strategy. Also (for once), this was about something that they actually deemed relevant. According to them, this was easily the best paper that we read all winter. I agree.

What we did not like so much

Nothing to declare.

Apr 142021
What we were reading: Campaign Effects and Issue Voting 2

How much of a difference do campaigns make?

Conventional wisdom says that electoral campaigns remind voters who they are and what they want. So they do not change much, but still matter, because the outcome could be very different without all the reminders. This little gem here is a bit unusual, because it tracks a referendum campaign in Denmark, not a normal electoral race. Also, there is detailed information on the timing of information.

Beach, D., & Finke, D. (2020). The long shadow of attitudes: differential campaign effects and issue voting in eu referendums. West European Politics, (online first), . http://dx.doi.org/10.1080/01402382.2020.1780829

What we liked

Students loved the fact, how the authors made use of panel data. They were also convinced by the rather intricate hypotheses that link to the literature on motivated reasoning. These complications not withstanding, they found the text quite accessible.

Politikens Hus, the central offices for Denmark’s biggest newspaper ‘Politiken’.

Photo by REVOLT on Unsplash

What we did not like so much

Ordered logit regression comes with problems of its own (hello parallel regressions assumption). We were not convinced that OLR is really necessary or even justified in this application. We also found some of the graphs a bit confusing. And like last week, students said that the summary (and even the presentation of the findings) were a bit too concise.

Apr 142021
What we were reading: Cross-cutting exposure and political participation 3

Does cross-cutting exposure help or hinder participation?

Somewhat predictably, I fell off the seminar-blogging wagon shortly after the break. But hey, this is a digital semester, so we have actual archived notes of our virtual meetings.

One of the first texts we read in the new year was this one: a specimen of the meta-analytic approach that is becoming more and more popular in Political Science. In this one, the authors want to find out if cross-cutting exposure (i.e. escaping from the filter bubble) has any positive or negative effects on political participation. Turns out that there is insufficient evidence for either.

Matthes, J., Knoll, J., Valenzuela, S., Hopmann, D. N., & Sikorski, C. v. (2019). A meta-analysis of the effects of cross-cutting exposure on political participation. Political Communication, 36(4), 523–542. http://dx.doi.org/10.1080/10584609.2019.1619638

What we liked

Most students had never heard of meta-analyses and were quite intrigued. They found the idea of pooling evidence from several published studies “stimulating” (their words) and were particularly happy that the authors had carefully documented each of their steps. They also liked the presentation of research questions and hypotheses.

Finally, some students had previously heard of publication bias. They were impressed that the authors applied a test for that, and were even more impressed that a journal had the guts to publish null results.

What we did not like so much

Some students were disappointed that the authors did not conduct any primary analysis, but (wait for it!), that’s the whole point of a meta analysis, right? A more serious criticism (levelled more at the journal and its page limit than at the authors) was that at least two interesting graphs were relegated to the appendices. And finally, students said that the summary & outlook section was too short. While I agree, I can also relate to the authors: you do all the work, you write it up, why should you have to summarise it again?