by Lamar Pierce, PhD, Associate professor of organization and strategy at Olin Business School, and CivicScience academic advisor


Election Day is a difficult one for many political candidates.  But a recent study I did with Todd Rogers of Harvard and Jason Snyder of UCLA shows it’s no picnic for their supporters either.

We used thousands of daily survey responses from CivicScience to show that although winning an election barely improves the happiness of those from the winning political party, losing reduces self-reported happiness and increases sadness substantially.

We compared the happiness and sadness reported by those who identify with political parties in the days surrounding the 2012 Presidential Election.  The data from CivicScience, which provided real-time responses just before and after the election, allowed us to estimate the distinct shock to aggregate happiness. Detailed demographic and party affiliation data showed the relationship of this shock with partisan identity, as well as controlled for any potential sample selection issues.

The sadness effect lasted for about a week, but eventually partisan losers got over it.  Thus, although the shock of losing was intense, it didn’t last.

This asymmetry between winning and losing is actually consistent with past research on happiness that has shown how bad things tend to hurt more and last longer than comparable good things.  But this evidence is unique because the real-time data that is collected by CivicScience each day can show the immediate change following an immediate and uncertain outcome. Although President Obama was favored to win a second term, CivicScience data showed most Republicans expected a Romney victory.

Since this was the first time we had studied happiness using the CivicScience data, we wanted to benchmark exactly how intense the pain of election losses were and validate that the magnitude of the effect was substantial. To do so, we compared the effect to that of two national tragedies which we know broadly impacted the happiness and sadness of many people. Using the same methodology with CivicScience data:

  • We found that respondents with children were also distinctly less happy and more sad after the Newtown school shooting, but that this sadness increase and happiness decrease were smaller than the effect of an election loss on partisans.
  • Similarly, smaller effects were found for those living in Boston during the Boston Marathon bombing.

Although it is impossible to compare such events, this study at least provides validation that the happiness numbers we observe are relevant and accurate.

Now, let’s hope people are still planning to hit the polls anyway.

Image by DonkeyHotey