The psychology of voting, digested

Psychology tells us that not only are political candidates competing with each other, they also have to contend with the foibles of human nature. Many of us like to think that we vote according to sound reason, perhaps for the good of the country, our own family's best interests, or by selecting the most fair and competent candidate. In fact, there's evidence that our votes are frequently influenced by more superficial factors, from a candidate's looks to the weather on election day. Here we digest the psychology of psephology, running down the evidence for 10 factors that affect people's behaviour in the ballot box:

Candidate Appearance

It would be reassuring to think that the electorate choose who to vote for based on the candidates' track records and future policy promises. In truth, many of us are swayed simply by the way that politicians look. Consider a 2009 study that asked Swiss students to look at multiple pairs of unfamiliar French political candidates and in each case to select the one who looked most competent. Most of the time, the candidate selected by students as looking the most competent was also the one who'd had real life electoral success, the implication being that voters too had been swayed by the candidates' appearance (there's little evidence that appearance and competence actually correlate). Unsurprisingly, being attractive also helps win votes, especially in war time (in peace time, looking trustworthy is more of an advantage). Other research has shown that we're more likely to vote for male and female candidates with deeper voices. Meanwhile obesity is a disadvantage for female candidates, but may help male candidates. People ignorant about politics are more swayed by politicians' appearance, especially if the politician has had plenty of TV exposure.

Candidate Personality

Journalists are often criticised for focusing overly on politicians' personalities rather than the "real issues" – in the UK election campaign at the time of writing, just look at the media commentary on opposition leader Ed Miliband. Psychology research suggests candidates' perceived traits are relevant, at least in the sense that they are related to the way we vote. A study from 2007 found that we tend to vote for politicians who we think have similar personalities to ourselves – for instance, prior to the 2004 US Presidential election, people who thought John Kerry shared their traits were more likely to vote for him in the election, whereas people who thought they were like George W Bush tended to vote in his favour. A similar effect has been found in the context of Italian and Spanish politics. Meanwhile, in a study published last year, students said they would be more willing to vote for politicians whom they considered to be more open-minded, friendly, and emotionally stable (the politicians' extraversion and conscientiousness were not related to the students' voting intentions).

The Polling Station

A growing body of evidence suggests that the places we go to vote, influence the way we vote. For example, in 2008, US researchers reported that people who voted at a polling station housed in a school were more likely to back a bill proposing more funding for education; and a 2010 study found voting in a church (rather than school or other location) boosted support for a conservative candidate. Sometimes these priming effects are less predictable: a study published last year (pdf) found that voters at a polling station in a church were more likely to support the introduction of same-sex marriages: possibly the religious symbolism reminded them that the arguments against such marriages are faith-based, which only served to increase their support for the marriages. There's even evidence that an uneven flooring could affect us: in this study from 2010, people leaning to the left (because of missing wheels on a chair) were found to be more sympathetic towards left-wing political attitudes (and vice versa if a wheel was missing on the right). A similar finding was obtained more recently using a wonky Wii balance board. A somewhat related and intriguing line of research finds that many people suspect their ballot choice is not truly secret and this influences them to vote according to social pressures, such as to conform with their declared affiliations.

Rain and Sunshine

Evidence from the USA (pdf), Spain and the Netherlands suggests that for each extra inch of rain fall on voting day, turnout reduces by around one per cent. Conversely, sunny weather and higher temperatures increase turn out (but not in Sweden where poor weather made no difference to turn out). There are also some more intriguing meteorological effects on voting. For instance, based on evidence that people's attitudes towards climate change are influenced by the local weather (higher temperatures increase belief in man-made global warming), the UK's Green Party might wish for a heat wave to strike at election time. Yet local sunshine was also found to increase approval ratings for US President George W Bush when he was in office, so as I write today, perhaps current Prime Minister Cameron would also benefit from a sunny spell. But consider too how poor weather affects people's risk aversion. A study presented in 2013 showed that people are less likely to vote for risky candidates when the weather is poor. A key feature of the General Election here in the UK is said to be the rise of minor parties and untested candidates. Perhaps the major parties should start their rain dances?

Shark Attacks, Sports Results and Storms

Following a dramatic series of shark attacks in New Jersey in 1916, voters punished the incumbent President Woodrow Wilson (according to an analysis published in 2012). This is just one example of how the electorate tends to blame governing parties for unwelcome events, even if those events are beyond the politicians' control. The converse is that incumbent politicians gain from positive circumstances. For instance, a 2010 US study found that the incumbent President benefited from extra votes in districts that had enjoyed football and basketball wins in the days leading up to an election. The effects of uncontrollable events are not always predictable and may depend on how politicians are seen to respond. When Hurricane Sandy struck in the days before the 2012 Presidential Election, this apparently (pdf) increased local votes for the incumbent, President Obama.

Daughters and Sisters

How we vote could depend on the gender of our children. That's according to a longitudinal analysis of British citizens published in 2010 – after having a daughter, people's political attitudes were more likely to swing to the left, and vice versa after having a son. The researchers think this happens because having a daughter increases awareness of issues facing women, such as pay discrimination, and increases sympathy for the typically greater desire among women for investment in public services. Note this is a contentious area: a Europe-wide study published last year failed to replicate this finding, while a US study found daughters increased parents' support for the (right-wing) Republican party. Of course, parents also influence their children's political persuasions: there's evidence that sons are affected by both parents, but daughters only by their mothers. The effects of parents on children's voting is both socio-cultural and genetic. We're also influenced by our siblings, especially our elder siblings. Just as the British analysis showed daughters increase parents' left-wing sympathies, a 2011 US study found that so too does having an older sister.

Scandals

An analysis of the last UK General Election in 2010 found that voters punished candidates who'd been found out by the expenses scandal – but the effect was modest and less than expected. A study of local Spanish politics also found that voters punished politicians caught up in corruption scandals, but the extent depended on media coverage and whether charges were brought. Voters' responses to scandals tends to be highly partisan – that is, we're lenient when the transgressing politician is from the party we support (and vice versa). Timing is important: a study from last year found that scandals that break later in an election campaign may be less harmful because voters have acquired policy information by then. A drip, drip of new scandal information sustains its damaging effects. The grammar used in reports also makes a difference: the imperfect tense "was fiddling his expenses" is more damaging than "fiddled". Some commentators warn that political scandals distract us from real issues, but a 2010 study found that when a politician is caught up in a scandal, this actually improves our memory for their policies – this is consistent with an associative memory account, in which the salience of the scandal boosts our memory for other information related to the politician.

Voter Emotions

When we're feeling happy with life, we're more likely to vote for the ruling party, so says an analysis from 2014 which controlled for the influence people's economic circumstances. A lab study (and this one) found that when we're angry we pay less attention to details about candidates; when we're fearful, by contrast, we scrutinise information more carefully, arguably making us more informed voters (but see here for a critique). Israeli research (pdf) finds that living in fear of rocket attacks increases people's support for right wing parties (although note, there's evidence that terrorist attacks in Madrid increased support for the country's opposition left-wing party in the election that came days later). Meanwhile, research shows that people who are more prone to disgust (for example, they dislike sitting on a bus seat left warm by a stranger) are more likely to hold right-wing conservative views [although see this 2019 update].

Political Adverts and Negative Campaigns

Political parties spend enormous amounts on advertising: this 2011 study on television ads found the effects on voting preferences to be strong, but short-lived. Ads with moody music and lighting are more effective. What about negative campaigns? In the current UK General Election Campaign, the incumbent Tory defence secretary recently made an attack on the character of the leader of the opposition and was widely criticised for doing so. This largely fits the findings from a 2010 lab study on negative campaigns – politicians who made negative statements about their opponents suffered a backlash, while the target of the attack was unaffected. It's worth noting though that based on voters' subconscious attitudes, the target of the attack did suffer a loss in standing (as did the politician making the attack). This meta-analysis from 2007 found that negative campaigns don't adversely affect voter turnout, but they do reduce trust in politics and lower public mood. A new study recently found that watching adverts that are congruent with our political beliefs makes us more likely to vote; watching an ad that clashes with our views has little effect.

"It's The Economy, Stupid" …

… this apparently was Bill Clinton's campaign mantra back in the 90s. With the British economy showing signs of recovery, today's incumbent Tory party will be hoping that Bill Clinton was correct – that ultimately, if the economy is doing well, people will reward the ruling party. However, British research suggests that this is not the case: for example, non-Tory voters who were financially comfortable at the time of the 1997 General Election did not reward the incumbent Tory party at that time by switching allegiance (and ditto in 2001 when previously non-Labour supporters in a good financial position failed to switch to voting Labour). These results might be explained in part by most people's partisanship (and "motivated reasoning") – when things go well under our preferred party, we credit the party, but if things go well under a party we oppose, then we don't. That said, there is evidence that sudden increases in people's personal wealth does influence their voting tendencies – winning the lottery makes it more likely that people will vote Republican, says this US study (and this one), and more likely that they'll support this incumbent party, says this Spanish research. This UK paper (pdf) found that as housing prices increase (to the benefit of homeowners, in terms of the wealth they have invested in their property), so too do intentions to vote for the Tory party.

- Find much more on voting across our website.