Assessing Voter Competence: A Study of Media Attention and Preferential Votes

The question of people’s competence to participate in politics has been discussed since Plato’s Republic. While many scholars have found the people to be wanting, others have been more optimistic. Lately, a specific question has been whether voters are competent enough to separate relevant from irrelevant information when making political decisions. We aim to study this question using a completely new base of evidence. In Norwegian local elections, voters can influence list order through preferential votes to individual candidates. We combine data on the number of preferential votes received by candidates in local elections from 2003 in the capital of Norway, Oslo (which is the largest municipality), with data on media attention they receive. Apart from attributes on candidates, our data consists of more than one million articles from Norwegian media outlets over the last decades. We identify i) how many times each candidate has been mentioned in the media; ii) in what kinds of magazines or newspapers they have been mentioned; and finally, using topic models, we identify iii) what topics candidates were mentioned in. In line with former findings, we find that media attention is a strong predictor of the number of preferential votes a candidate receives, even when controlling for the candidate’s sex, age, party, list position, and former office holding - in fact, even when applying candidate fixed effects across elections. Is this a product of incompetent voters susceptible to mere name recognition bias and focus on celebrity status, or a product of competent voters making meaningful choices? We hypothesize that competent voters distinguish between relevant and irrelevant sources of information, while incompetent voters do not. If voters are incompetent, the media outlet in which a candidate is mentioned should not matter. However, it does. While getting attention in serious newspapers is an important predictor of preferential votes, attention in tabloid magazines is not; if anything, it reduces the predicted number of preferential votes. We further hypothesize that if voters are incompetent, the topic of the article in which the candidate is mentioned should not matter. But it does. Being mentioned in an article about politics or Oslo’s urban planning increases the predicted preferential votes, while being mentioned in articles about sports or international affairs does not increase the number of preferential votes in local elections. The results are important for democratic policy. In the controversial Norwegian debate about extending preferential votes to the national elections, one of the major counterarguments is that voters do not know enough about the candidates to separate them from each other, and that voters will vote for celebrities, at the expense of competent candidates. This does not seem to be the case. We find that Oslo voters distinguish between relevant and irrelevant information. In fact, they display traits of being highly competent. While only 20-30% of the Oslo voters choose to place preferential votes, the local election studies confirm that these voters do not stand out in terms of educational or class background. The results thus have potential for generalization.