Losing Hurts: The Happiness Impact of Partisan Electoral Loss

Abstract

Partisan identity shapes social, mental, economic, and physical life. Using a novel dataset, we study the consequences of partisan identity by examining the immediate impact of electoral loss and victory on happiness and sadness. Employing a quasi-experimental regression discontinuity model we present two primary findings. First, elections strongly affect the immediate happiness/sadness of partisan losers, but minimally impact partisan winners. This effect is consistent with psychological research on the good-bad hedonic asymmetry, but appears to dissipate within a week after the election. Second, the immediate happiness consequences to partisan losers are relatively strong. To illustrate, we show that partisans are affected two times more by their party losing the 2012 U.S. Presidential Election than both respondents with children were to the Newtown shootings and respondents living in Boston were to the Boston Marathon bombings. We discuss implications regarding the centrality of partisan identity to the self and its well-being. Word count: 3,196

Full text

!
Losing Hurts:
The Happiness Impact of Partisan Electoral Loss
Lamar Pierce
1
, Todd Rogers
2
, and Jason A. Snyder
3∗
Forthcoming in Journal of Experimental Political Science
Abstract (Words: 149):
Partisan identity shapes social, mental, economic, and physical life. Using a novel dataset, we
study the consequences of partisan identity by examining the immediate impact of electoral
loss and victory on happiness and sadness. Employing a quasi-experimental regression
discontinuity model we present two primary findings. First, elections strongly affect the
immediate happiness/sadness of partisan losers, but minimally impact partisan winners. This
effect is consistent with psychological research on the good-bad hedonic asymmetry, but
appears to dissipate within a week after the election. Second, the immediate happiness
consequences to partisan losers are relatively strong. To illustrate, we show that partisans are
affected two times more by their party losing the 2012 U.S. Presidential Election than both
respondents with children were to the Newtown shootings and respondents living in Boston
were to the Boston Marathon bombings. We discuss implications regarding the centrality of
partisan identity to the self and its well-being.
Word count: 3,196
Keywords: Partisanship, Political Psychology, Happiness, Elections, Identity, Well-Being,
Obama
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
The authors thank John Dick, Ross McGowan, and Zack Sharek at CivicScience, Inc. for providing the data, as
well as Craig Fox, Francesca Gino, Eitan Hersh, Karim Kassam, Marc Meredith, Brendan Nyhan, Mike
Norton, Erik Snowberg, and Kathleen Vohs for thoughtful comments and suggestions. Carly Robinson
provided excellent research assistance.
1
Washington University in St. Louis, Olin Business School, One Brookings Drive Box 1133, St. Louis MO
63130.
2
Harvard Kennedy School, Mailbox 124, 79 JFK Street, Cambridge MA 02138.
3
UCLA, Anderson School of Management, 110 Westwood Plaza, Cornell Hall, Suite D506, Los Angeles CA
90095.
∗
Authors contributed equally and listed alphabetically. Direct correspondence to pierce@wustl.edu.

How important is partisan identity to happiness? It might be of considerable
importance to the two-thirds of Americans who identify with a political party, given its
powerful influence on other dimensions of people’s lives. Partisan identity is stable across
people’s lifetimes (Campbell et al, 1960; Green, Palmquist and Schickler, 2002), causally
shaping political preferences and the factual qualities people associate with policies (Cohen,
2003). People more frequently live near (Glaeser and Ward, 2006; Gimpel and Schuknecht,
2004) and interact with (Grentzkow and Shapiro, 2001) those who share their partisan
identity than with those who do not. Furthermore, partisan identity tends to define media
consumption (Prior, 2007) and other economic behavior (Gerber and Huber, 2009), and can
bias social perceptions and favoritism (Caruso, Mead and Belcetis, 2009; Rand et al, 2009). In
short, social, mental, economic, and physical life is shaped by partisan identity.
Given this importance, political outcomes such as the 2012 U.S Presidential Election
could profoundly impact the happiness of both partisan winners (Democrats) and partisan
losers (Republicans). This research uses a novel dataset that tracks fluctuations in happiness
and sadness to address two questions about the importance of partisan identity to well-being.
First, are the shocks to happiness from winning and losing equivalent? Diverse research
suggests they might not be. Bad events cause stronger reactions than comparable good ones
(Baumeister et al, 2001; Rozin and Royzman, 2001), similar to predictions from prospect
theory’s value function about the gain-loss asymmetry (Kahneman and Tversky, 1979;
McDermott, 2004).
Second, how strong is the shock of partisan loss to happiness? We compare the well-
being consequences of partisan loss to that of two national tragedies that dominated the
national news media for weeks. On December 14, 2012, twenty children and six adults were
fatally shot at Sandy Hook Elementary School in Newtown, Connecticut (“Newtown
1

shootings”). On April 15, 2013, three people were killed and 283 injured after terrorists
attacked the Boston Marathon (“marathon bombings”). While such tragedies are
qualitatively different than elections, comparing their well-being consequences to that of
partisan loss illustrates simply the relative importance of partisan identity to well-being.
Tragedies have both political repercussions (Gillis, 1996) and elicit emotional, financial, and
civic responses from people not directly affected by the trauma (Preston and De Waal, 2002;
Singer et al, 2004). Consequently, one might sensibly expect the hedonic (happiness-based)
impact of partisan loss to pale in comparison.
Using daily responses from CivicScience, Inc., an online polling and data intelligence
company, we employ a quasi-experimental regression discontinuity (RD) design to estimate
the happiness shock of specific events. The RD design overcomes many of the sampling bias
problems in survey-based studies of happiness by focusing on nearly identical respondents
immediately before and after an independent shock (Shadish, Cook and Campbell, 2002;
Imbens and Lemieux, 2008).
We find that the pain of losing an election is much larger than the joy of winning
one, but that this happiness loss is short-lived. Election outcomes strongly affect the short-
term happiness/sadness of partisan losers, with minimal impact on partisan winners. This
result is consistent with studies finding that “bad emotions, bad parents, and bad feedback
have more impact than good ones…bad information is processed more thoroughly than
good…[and] the self is more motivated to avoid bad self-definitions than to pursue good
ones” (Baumeister et al, 2001). Despite the strength of the loss, happiness appears to recover
within a week, consistent with research on people’s tendency to adapt to bad events more
quickly than expected (Gilbert et al, 2004). This temporariness suggests partisan loss impairs
emotional well-being rather than broader life satisfaction (Kahneman and Deaton, 2010).
2

The short-term strength of partisan loss is contrasted with responses to mass
national tragedies; partisans are affected twice as much by their candidate losing the U.S.
Presidential Election than both respondents with children were to the Newtown shootings
and respondents in Boston were to the marathon bombings. The fact that the pain
experienced by partisan losers is stronger than that of people for whom the tragedies were
self-relevant benchmarks the centrality of partisan identity to the self and well-being.
Method
Data. CivicScience polls over 300,000 unique individuals daily across the United
States on over 500 third-party websites. Unpaid volunteers answer three questions in
embedded polls. Tracking technology allows the company to identify returning respondents
across all partner websites, thereby collecting a panel of detailed demographic and attitudinal
data for many respondents. One question that is randomly and continuously distributed
across all partner websites each day asks “How happy are you today -- very happy, happy, so so,
unhappy, or very unhappy?” This question is similar to one used in the Euro-Barometer Survey
Series and the United States General Social Survey—widely used to study happiness in
economics (Easterlin, 2003; Argyle, 2003; Di Tella, MacCulloch, and Oswald, 2003; Alesina,
Di Tella and MacColloch, 2004; Easterlin, 2006). Consistent with the prior literature we
create an indicator variable “happy” equal to 1 if respondents reported being happy or very
happy.
4
CivicScience also collects extensive socio-demographic information (e.g., gender,
income, race, age, and partisan identity) in a pre-determined sequence from return
respondents over multiple visits to partner websites. We had access to data on all
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
4
This dichotomized variable is easier to interpret and more meaningful than a 1-5 scale assuming each unit
change is equal. Tables S7A and S7B in the supplementary materials present similar results with the scaled
dependent variable.
3

respondents who had answered the happiness and sadness questions, but not all of these
respondents answered all socio-demographic, party affiliation, and parental status questions.
For example, for the week before and after the election approximately 67% of respondents
with happiness responses and full socio-demographic data had answered the partisan identity
question; eighty five percent of such respondents from the week before and after Newtown
had answered the parental status question. Missing responses for each of question yields a
sparse data matrix, which makes imputing missing data very unreliable. We therefore restrict
the data for each analysis to the respondents for whom there are no missing observations.
CivicScience asks “Politically, do you consider yourself more of a: Republican, Democrat, or
Independent?” Like all surveys, the sample of these individuals is conditioned on the decision
to participate in repeated CivicScience polls. CivicScience respondents were somewhat more
Republican than the general population. We are unable to observe data on which
respondents chose not to answer specific questions. Figure 1 previews our core results.
During the two weeks surrounding Election Day, an average of 210 Republicans and 111
Democrats answered an online happiness question each day. Notice the little change in the
likelihood that Democrats report being happy, while immediately following the election
Republicans’ self-reported happiness drops from approximately 60% to 30%. These data are
collected with enough frequency that daily shocks can be clearly identified, a feature unique
to most research on happiness. We note, however, that our models’ key identifying
assumption is that the sample is similar before and after the election. Additionally, given the
self-selection and uneven geographic distribution of the sample, one must be careful in
extrapolating specific effect magnitudes to the general population. Finally, we note how days
are coded. Across all studies we code days as being 24-hour periods immediately preceding
and following focal events. For instance, the 2012 presidential election was called by the
4

Associated Press at approximately 11PM EST on Election Day, so the previous day began at
11PM EST the day before Election Day.
Model. We use quasi-experimental regression discontinuity models to test how the
happiness levels of Democrats and Republicans discretely change immediately following the
presidential election. RD models assign observations to treatment and control groups based
on a discrete threshold in a continuous assignment variable, which in our case is time (days).
The discrete threshold is the focal event (e.g., Election Day or day of the tragedy). Any
response after the focal event threshold is considered “treated,” while prior days are in the
“control.” RD models are most commonly used in political science, economics, and
psychology (Dal Bo, Dal Bo and Snyder, 2009; Snyder, 2010; Gerber, Kessler, and Meredith,
2011; Pierce, Dahl, and Nielsen, 2013; Hersh, 2014), with many examples applying RD
models to time series data, as we do (Busse, Silvo-Risso, and Zettelmeyer, 2006; Pierce and
Snyder, 2012).
RD models assume that observations just above and below the threshold are
identical on all dimensions except the focal treatment. Table S1-S3 provides detailed
evidence that respondents one week before and after the three events in our data are
reasonably identical on observable dimensions. Simple t-tests of differences in pre/post
means reveal few systematic demographic differences, nor do regression discontinuity
models that replace Happiness with each demographic as the dependent variable in equation
(1) below. Our base specification is as follows at the individual-level respondent:
(1) 𝐻𝑎𝑝𝑝𝑖𝑛𝑒𝑠𝑠
!
=!∝ ! +!𝛽
!
∗ 𝑃𝑜𝑠𝑡𝐸𝑣𝑒𝑛𝑡
!
+ ! 𝛽
!
∗ 𝐿𝑖𝑛𝑒𝑎𝑟!𝑇𝑖𝑚𝑒!𝑉𝑎𝑟𝑖𝑎𝑏𝑙𝑒
!
+𝛽
!
∗ 𝑃𝑜𝑠𝑡𝐸𝑣𝑒𝑛𝑡
!
∗ ! 𝐿𝑖𝑛𝑒𝑎𝑟!𝑇𝑖𝑚𝑒!𝑉𝑎𝑟𝑖𝑎𝑏𝑙𝑒
!
+ ! 𝜀
!
Linear Time Variable runs from -7 to 6, where 0 is the day immediately following the
5

focal event. PostEvent is an indicator equal to 1 if the event has already occurred. Figure 2
illustrates this specification for Republican respondents’ happiness in relation to the election.
𝛽
!
estimates the discrete jump between the two regression estimates. 𝛽
!
is the slope prior to
the election and 𝛽
!
is the slope after the election. This specification therefore estimates the
size of the break while controlling for the different time trends before and after the event.
Other specifications include socio-demographic characteristics and higher order time
polynomials for robustness. All results are clustered at the MSA level.
To test the persistence of the happiness effect, a second model relaxes the RD
assumption to examine weekly happiness rates for Republicans and Democrats, conditioning
on socio-demographic information and location. Although this model provides some
evidence of effect persistence, we cannot observe the counterfactual time trend in weeks
distant from Election Day. Any inference about effect length must assume that happiness
would return to pre-election levels absent the election’s effect.
(2) 𝐻𝑎𝑝𝑝𝑖𝑛𝑒𝑠𝑠
!
=!∝ ! +!𝜷
𝟏
∗ 𝑊𝑒𝑒𝑘!𝐼𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟𝑠
!
+ ! 𝛽
!
∗ 𝑆𝑜𝑐𝑖𝑜𝐷𝑒𝑚𝑜𝑔𝑟𝑎𝑝ℎ𝑖𝑐!𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 + ! 𝜀
!
Study 1: 2012 U.S. Presidential Election
Tables 1a and 1b report the results for happiness around the election depicted in
Figures 1 & 2, splitting the sample by partisan winners (Democrats) and partisan losers
(Republicans). Across specifications there is little robust evidence that Democrats’ responses
changed immediately after the election. The sign across specifications is positive, but the
statistical significance is inconsistent across specifications. In contrast, partisan losers
experienced significantly larger hedonic shocks than partisan winners. Table 1B shows a
strong negative effect on the baseline level of happiness following the election. The models
are robust to including extensive demographic (race, gender, age, income), geographic
6

(metropolitan area fixed effects defined by IP address), and time control variables. This
robustness across specifications casts doubt on concerns that results are driven by
differences in the types of respondents before and after the election, as do the nearly
identical respondent characteristics before and after the election presented in Table S1. The
negative happiness impact to partisan losers, for example, actually increases from -.151 to -
.246 after all controls are added. Across each pair of columns from Tables 1A and 1B, the
coefficients are statistically different from each other at the 5% confidence level.
Figures 3A and 3B depict parameter estimates and confidence intervals associated
with equation (2). The weekly differences in happiness are all relative to the baseline 8
th
week
before the election and condition on the socio-demographic characteristics described earlier.
Over the 8 weeks before and after the election happiness is relatively constant except for
Republicans in the week immediately following the election. This evidence again shows that
Republicans’ well-being drops after the event, and also suggests that it recovers quickly.
Alternative Explanations. Three possible alternative explanations stand out. First,
the asymmetric hedonic response could stem from rational responses to the election’s policy
implications for Democrats and Republicans (Gerber and Huber, 2009). This seems unlikely,
however, since Figures 1 and 2 show happiness levels converging within one week following
the election, becoming statistically indistinguishable within four weeks.
Second, the asymmetric hedonic response could reflect different outcome
expectations. Overconfidence among partisan losers is common even in blowout elections
(Granberg and Brent, 1983), partly because simply supporting a candidate causes people to
believe that candidate will win (Krizan, Miller, and Johar, 2010). In this alternative
explanation, Republicans would be more affected because they expected a victory and were
7

disappointed, while Democrats, also expecting a victory, were unsurprised. Supplementary
analysis suggests a similar asymmetric shock for only those respondents expecting their
candidate to win, casting doubt on this expectations explanation.
Finally Republicans may simply become less happy after an election, regardless of the
outcome. This alternative seems unlikely, but we cannot directly test this hypothesis with this
data.
Study 2: Newtown Shootings and Marathon Bombings
Two major national tragedies that dominated the media for weeks occurred after the
election: the Newtown shootings and the Boston Marathon bombings. Many respondents
answered the happiness/sadness questions in the weeks surrounding the two tragedies,
averaging 445/day for the Newtown shootings and 639/day for the marathon bombings.
These data are analyzed using the same strategy as with the election data, defining the post-
event treatment dummy by whether each response was before or after the precise time of
the event’s first news coverage. Of course, learning that one’s party lost an election differs in
important ways from observing a national tragedy. For example, partisans are personally
invested in and occasionally involved in elections, while very few people are personally
involved in national tragedies. That said, of the 60% of Americans who identify with a
political party, only about 0.40% were personally involved in the 2012 election by donating
over $200 to a candidate, party, or PAC (Opensecrets.org). Nonetheless, comparing the
hedonic impact of these two national tragedies to that of losing an election can be insightful
because they were the most affectively intense events impacting the mass public during this
period. This comparison serves simply to benchmark the hedonic intensity of partisan loss,
and cannot account for other psychological impacts such as anxiety or fear.
8

Both the Newtown shootings and marathon bombings caused significant negative
hedonic shocks, but they are much smaller than those suffered by partisan losers in the
election. Table 2 presents the RD estimates for respondent happiness in relation to the
Newtown shootings. Across columns (1) – (5) the results are not consistently statistically
significant. The fully-controlled model in column 2 estimates a 7.6% happiness decrease
immediately following the Newtown shootings—only one fifth the size of the decrease
experienced by partisan losers. Likewise Table 3, column 4, shows that happiness decreases
following the marathon bombings by only 4.8%. The statistical significance varies across
multiple reasonable specifications.
Election outcomes are relevant to partisans’ identities. As such, it may not mean
much to compare the hedonic impact of partisan loss to that of national tragedies to a broad
swath of respondents. We therefore assess the hedonic impact of tragedies on those for
whom the tragedies are identity-relevant: the Newtown shootings on self-reported parents
and the marathon bombings on respondents using Boston-based IP addresses. The RD
models are reported in columns (7) & (8) of Tables 2 & 3. As one would predict, these
subsamples show larger impacts than the more general sample. However, the effects are still
only half those on partisan losers from the election (ps<.01). The differences between the
coefficient on Post-Newtown in columns (7) & (8) of Table 2 is significant at the 5%
confidence level, while the difference between the Boston and Non-Boston region is not
statistically significant.
Figure 4 presents the daily happiness and sadness results for all three events
(Presidential election, Newtown shootings, and marathon bombings) for the identity-
relevant groups (Republicans, parents, and Boston residents). The visual comparison,
9

combined with the regression results, strongly suggests that Republicans’ hedonic response
to the election was larger than either response to the two tragedies.
General Discussion
People’s social, physical, economic, and mental lives are shaped by their partisan
identities – and these social identities are widely and deeply held. The current research vividly
shows that these identities also have important consequences to people’s hedonic lives.
Winning an election is fine, but losing one is painful, at least in the short run. Losing an
election appears to dominate the pain caused by national tragedies, even among those
particularly connected to them. While enhancing our understanding of the centrality of
people’s partisanship to their lived experiences, these results also speak to the growing
literature in economics, psychology, and other fields on the factors that affect well-being
(Kahneman, Diener and Schwarz, 2003).
In addition to expanding our understanding of the well-being importance of partisan
identity, this work makes several methodological contributions. First, it tackles a causal
political psychology question by employing a research design (regression discontinuity) that
is under-used in other political psychology research (Shadish, Cook, and Campbell, 2002).
Second, it leverages digital technologies that allow large-scale, yet granular, data collection
over time. One will notice in Figure 4 the rapid adaption of partisan losers to losing an
election; of parents to the Newtown shootings; and of Boston residents to the marathon
bombings. As far as we know, this is the first paper to map the contours of hedonic
adaptation to societal events at this level of granularity. This type of data source provides
new opportunities for scholars involved in the study and measurement of happiness
(Kahneman and Krueger, 2006). Finally, by using hedonic reactions to multiple unrelated
events that are each associated with distinct identities, we illustrate an approach to
10

comparing the importance of different beliefs, ideologies, or events to people’s identities
with relatively high ecological validity (Settles, 2004).
One of our main findings is that the pain of losing the 2012 Presidential Election
dominated the joy of winning it. A challenge to making a general claim is the many
idiosyncrasies to this specific election. First, the impact of losing the election may be specific
to Republicans since partisans appear to have systematic differences in how they process and
respond to information (Jost, Federico, and Napier, 2009). Second, it is difficult for us to
disentangle the role of party affiliation from simple candidate preferences. Third, since
President Obama was the incumbent, partisan winners might have perceived retaining the
presidency as maintaining the status quo, thereby muting the joy of winning. In this scenario,
however, partisan losers would have viewed the status quo as not attaining the presidency
(i.e., losing), making this status quo argument inconsistent with the results. It is also
inconsistent with the robust finding that partisans expect their preferred candidates to win,
even when the polls show that winning is unlikely (Granberg and Brent, 1983). The current
findings should be replicated in future elections to resolve these questions.
Furthermore, the results appear inconsistent with research suggesting that prospect
theory’s gain-loss asymmetry arises when people forecast their hedonic reactions, but not
when people actually experience gains and losses with monetary gambles (Kermer et al,
2006). One possible explanation for this inconsistency might be that partisans expect to win
elections (Granberg and Brent, 1983; Krizan, Miller, and Johar, 2010), whereas
overconfidence may be more muted for monetary gambles.
Finally, we note that although partisan losers appear to be only temporarily affected,
such transitive emotional shocks have important personal and social implications. Card and
11

Dahl (2011), for example, find that upset losses in football games increased local domestic
violence reports for a short period following the game.
In sum, partisan identity is even more central to the self than past research suggests.
In addition to affecting thinking, preferences, and behavior, it also has sizable hedonic
consequences, especially when people experience partisan losses.
12

References
Alesina, A., Di Tella, R., & MacCulloch, R. (2004). Inequality and happiness: are Europeans
and Americans different? Journal of Public Economics, 88(9), 2009-2042.
Argyle, M. (2003). 18 Causes and Correlates of Happiness. Well-Being: The Foundations of
Hedonic Psychology, 353.
Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger
than good. Review of General Psychology, 5(4), 323.
Busse, M., Silva-Risso, J and Zettelmeyer, F. (2006). "$1,000 Cash Back: The pass-through of
auto manufacturer promotions." American Economic Review, 96(4): 1253-1270.
Campbell, A., Converse, P. E., Miller, W. E., & Donald, E. (1966). Stokes. 1960. The
American Voter.
Card, D., & G. Dahl. (2011). Family violence and football: The effect of unexpected
emotional cues on violent behavior. Quarterly Journal of Economics, 126(1): 103-143.
Caruso, E., Mead, N., & E. Balcetis. (2009). Political partisanship influences perception of
biracial candidates’ skin tone. Proceedings of National Academy of Sciences, 106(48): 20168-
20173.
Cohen, G. L. (2003). Party over policy: The dominating impact of group influence on
political beliefs. Journal of Personality and Social Psychology, 85(5), 808-822.
Dal Bó, E., Dal Bó, P., & Snyder, J. (2009). Political dynasties. The Review of Economic
Studies, 76(1), 115-142.
Di Tella, R. D., MacCulloch, R. J., & Oswald, A. J. (2003). The macroeconomics of
happiness. Review of Economics and Statistics, 85(4), 809-827.
Easterlin R. A. (2003). Explaining happiness. PNAS 100: 11176-11183.
Easterlin, R. A. (2006). Life cycle happiness and its sources: Intersections of psychology,
economics, and demography. Journal of Economic Psychology,27(4), 463-482.
Hersh, E.!(2014). The long-term effect of September 11 on the political behavior of victims'
families and neighbors. Proceedings of the National Academy of Sciences, 110 (52): 20959-
20963.
Gentzkow, M., & Shapiro, J. M. (2011). Ideological segregation online and offline. The
Quarterly Journal of Economics, 126(4), 1799-1839.
Gerber, A. S., & Huber, G. A. (2009). Partisanship and economic behavior: Do partisan
differences in economic forecasts predict real economic behavior? American Political
Science Review, 103(3), 407-426.
Gerber, A. S., Kessler, D. P., & Meredith, M. (2011). The persuasive effects of direct mail: A
regression discontinuity based approach. Journal of Politics, 73(1), 140-155.
Gilbert, D. T., Lieberman, M. D., Morewedge, C. K., & Wilson, T. D. (2004). The peculiar
longevity of things not so bad. Psychological Science, 15, 14-19.
Gillis, J. R. (Ed.). (1996). Commemorations: The politics of national identity. Princeton
University Press.
Gimpel, J. G., & Schuknecht, J. E. (2004). Patchwork Nation: Sectionalism and Political Change in
American Politics. University of Michigan Press.
Glaeser, E. L., & Ward, B. A. (2006). Myths and realities of American political geography.
Journal of Economic Perspectives, 20(2), 119-144.
Granberg, D., & Brent, E. (1983). When prophecy bends: The preference–expectation link
in US presidential elections, 1952–1980. Journal of Personality and Social Psychology, 45(3),
477.
13

Green, D., Palmquist, B., & Schickler, E. (2002). Partisan Hearts and Minds: Political Parties and
the Social Identities of Voters. Yale University Press.
Imbens, G. W., & Lemieux, T. (2008). Regression discontinuity designs: A guide to
practice. Journal of Econometrics, 142(2), 615-635.
Jost, J. T., Federico, C. M., & Napier, J. L. (2009). Political ideology: Its structure, functions,
and elective affinities. Annual review of psychology, 60, 307-337.
Kahneman, D., & Krueger, A. (2006). Developments in the measurement of subjective well-
being. Journal of Economic Perspectives, 20(1), 3-24.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under
risk. Econometrica, 263-291.
Kahneman, D., & Deaton, A. (2010). High income improves evaluation of life but not
emotional well-being. Proceedings of the National Academy of Sciences, 107(38), 16489-
16493.
Kahneman, D., Diener, E., and Schwarz, N. (2003). Well-Being: The Foundations of Hedonic
Psychology. Russell Sage Foundation.
Kermer, D. A., Driver-Linn, E., Wilson, T. D., & Gilbert, D. T. (2006). Loss aversion is an
affective forecasting error. Psychological Science, 17(8), 649-653.
Krizan, Z., Miller, J. C., & Johar, O. (2010). Wishful thinking in the 2008 US presidential
election. Psychological Science, 21(1), 140-146.
Langford, D. J., Crager, S. E., Shehzad, Z., Smith, S. B., Sotocinal, S. G., Levenstadt, J. S.,
Chanda, M. L., Levitin, D. J., & Mogil, J. S. (2006). Social modulation of pain as
evidence for empathy in mice. Science 312(5782), 1967-1970.
McDermott, R. (2004). Prospect theory in political science: Gains and losses from the first
decade. Political Psychology, 25(2), 289-312.
Pierce, L, & J.A. Snyder. (2012). Discretion and manipulation by experts: Evidence from a
vehicle emissions policy change. B.E. Journal of Economic Analysis & Policy, 13(3).
Pierce, L., Dahl, M. S., & Nielsen, J. (2013). In sickness and in wealth: Psychological and
sexual costs of income comparison in marriage. Personality and Social Psychology
Bulletin, 39(3), 359-374.
Preston, S. D. & De Waal, F. B. (2002). Empathy: Its ultimate and proximate bases.
Behavioral and Brain Sciences, 25(1), 1-71.
Prior, M. (2007). Post-Broadcast Democracy: How Media Choice Increases Inequality in Political
Involvement and Polarizes Elections. Cambridge University Press.
Rand, D., Pfeiffer, T., Dreber, A., Sheketoff, R., Wernerfelt, N., & Benkler, Y. (2009).
Dynamic remodeling of in-group bias during the 2008 presidential election.
Proceedings of the National Academy of Sciences, 106(15): 6187-6191.
Rozin, P., & Royzman, E. B. (2001). Negativity bias, negativity dominance, and
contagion. Personality and Social Psychology Review, 5(4), 296-320.
Settles, I. H. (2004). When multiple identities interfere: The role of identity
centrality. Personality and Social Psychology Bulletin, 30(4), 487-500.
Shadish, W.R, Cook, T.D., & Campbell, D.T. (2002). Experimental and Quasi-Experimental
Designs for Generalized Causal Inference. Boston: Houghton Mifflin.
Singer, T., Seymour, B., O’Doherty, J., Kaube, H., Dolan, R. J., & Frith, C. D. (2004).
Empathy for pain involves the affective but not sensory components of pain. Science,
303(5661), 1157-1162.
Snyder, J. (2010). Gaming the liver transplant market. Journal of Law, Economics, and
Organization, 26(3), 546-568.
14

West, S. G. (2009). Alternatives to randomized experiments. Current Directions in Psychological
Science, 18(5), 299-304.
15

Independent variables (1) (2) (3) (4) (5)
Post-election
.039
(.049)
.033*
(.019)
.037
(.047)
.049
(.052)
.084
(.137)
Post-election * One de-
gree polynomial of days
No No Yes Yes No
Post-election * ree de-
gree polynomial of days
No No No No Yes
Socio - Demographic &
MSA Controls
No No No Yes Yes
Time restriction +/- one day +/- one week +/- one week +/- one week +/- one week
Observations 265 1,553 1,553 1,553 1,553
Note: *, **, and *** indicate signicance at the 10%, 5%, and 1% condence levels, respectively. Standard errors are clus-
tered at the Metropolitan level. Socio-demographic controls include gender, age indicators, race indicators, and income
indicators. MSA controls included indicators for the metropolitan statistical area.
Table 1A: Democrat happiness one week surrounding 2012 election
Dependent variable: Are you happy today?
Independent variables (1) (2) (3) (4) (6)
Post-election
-.310***
(.040)
-.151***
(.016)
-.243***
(.035)
-.246***
(.039)
-.316***
(.096)
Post-election * One de-
gree polynomial of days
No No Yes Yes No
Post-election * ree de-
gree polynomial of days
No No No No Yes
Socio - Demographic &
MSA Controls
No No No Yes Yes
Time restriction +/- one day +/- one week +/- one week +/- one week +/- one week
Observations 465 2,934 2,934 2,934 2,934
Table 1B: Republican happiness one week surrounding 2012 election
Dependent variable: Are you happy today?
16

Independent variables (1) (2) (3) (4) (5) (6) (7)
Post-Newtown
-.063
(.039)
-.014
(.014)
-.076***
(.029)
-.062**
(.029)
-.035
(.077)
.060
(.061)
-.100***
(.034)
Post-Newtown * One degree
polynomial of days
No No Yes Yes No Yes Yes
Post-Newtown * ree degree
polynomial of days
No No No No Yes No No
Socio - Demographic & MSA
Controls
No No No Yes Yes Yes Yes
Time restriction +/- one day +/- one week +/- one week +/- one week +/- one week +/- one week +/- one week
Parents & Non-Parents Both Both Both Both Both Non-Parents Parents
Observations 695 5,304 5,304 5,304 5,304 1,216 4,088
Note: *, **, and *** indicate signicance at the 10%, 5%, and 1% condence levels, respectively. Standard errors are clustered at the Metropolitan level. Socio-demograph-
ic controls include gender, age indicators, race indicators, and income indicators. MSA controls included indicators for the metropolitan statistical area.
Table 2: Self-reported happiness before and after Newtown shooting
Dependent variable: Are you happy today?
17

Independent variables (1) (2) (3) (4) (5) (7) (8)
Post-Boston
-.064**
(.039)
-.022***
(.008)
-.048**
(.022)
-.050**
(.023)
-.069
(.051)
-.048**
(.024)
-.204
(.144)
Post-Boston * One degree poly-
nomial of days
No No Yes Yes No Yes Yes
Post-Boston * ree degree poly-
nomial of days
No No No No Yes No No
Socio - Demographic & MSA
Controls
No No No Yes Yes Yes Yes
Time restriction +/- one day +/- one week +/- one week +/- one week +/- one week +/- one week +/- one week
Boston Region & Non-Boston
Region
Both Both Both Both Both Non-Boston Boston
Observations 1,360 8,939 8,939 8,939 8,939 8,763 176
Note: *, **, and *** indicate signicance at the 10%, 5%, and 1% condence levels, respectively. Standard errors are clustered at the Metropolitan level. Socio-demograph-
ic controls include gender, age indicators, race indicators, and income indicators. MSA controls included indicators for the metropolitan statistical area.
Table 3: Self-reported happiness before and after Boston bombing
Dependent variable: Are you happy today?
18

.3 .4 .5 .6 .7
Percentage indicating that they are happy today
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Days before and after election
.2 .4 .6 .8 1
Percentage indicating that they are happy today
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Days before and after election
Figure 1: Self-reported happiness one week before and after the election
First day after
2012 election
Figure 2: Self-reported happiness one week before and after the
election for Republicans with discontinuity model
First day after
2012 election
Note: Dashed lines are +/- 1.96 standard errors. e dierence between the lines at the rst day after the election cor-
responds to the b
1
in equation (1). e dierences in slopes of the estimated line are the consequence of the interaction
between the linear term and the post-election indicator.
Intervals represent
± 1.96 standard errors
Democrats
Republicans
19

-.15 -.1 -.05 0 .05 .1
Week Coefficient (Baseline is 8th Week Before Election)
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
Weeks before and after Election
-.1 -.05 0 .05 .1 .15
Week Coefficient (Baseline is 8th Week Before Election)
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
Weeks before and after Election
Figure 3A: Self-reported happiness for Democrats eight weeks before
and after the election
Figure 3B: Self-reported happiness for Republicans eight weeks before
and after the election
Regression coecents on week indicator from
equation (1) for the set of Republicans in the
dataset. Week 1 is the baseline. Regressions
included controls for gender, age indicators,
race indicators, and income indicators. MSA
controls included indicators for the metropoli-
tan statistical area. Clustered at the MSA level.
95% condence intervals show as dashed lines.
Regression coecents on week indicator
from equation (1) for the set of Demo-
crats in the dataset. Week 1 is the base-
line. Regressions included controls for
gender, age indicators, race indicators,
and income indicators. MSA controls
included indicators for the metropolitan
statistical area clustered at the MSA level.
95% condence intervals show as dashed
lines.
Drop in happiness for Re-
publicans in the rst week
post-election is disipates in
the subsequent weeks.
20

0 .2
.4 .6 .8 1
Percentage indicating that they are happy today
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Days before and after event
Figure 4: Comparison of salient groups before and after event
Event Date
Intervals represent
± 1.96 standard errors
Parents before and after Newtown shooting
Republicans before and after 2012 election
Boston region before and after Boston bombing
21

Supplementary Appendix
Losing Hurts: The Happiness Impact of Partisan Loss
Addressing Alternative Expectations Explanation
To address the concern that Democrats and Republicans suffered different
happiness shocks because of different expectations about winning, another question that
CivicScience asked before the election was analyzed: Who do you think will win the US
Presidential election: Barack Obama, not sure but probably Obama, not sure but probably Romney, Mitt
Romney, not sure? 286 respondents answered this question between October 31 and
November 6. 94% of partisan losers and 87% of partisan winners expected to win the
election in the week before the election. Of these respondents, 210 were Republicans.
Uncontrolled RD models run separately on the 197 (94%) expecting a Republican victory
and the 13 (6%) expecting a Democrat victory. Partisan losers expecting to win experienced
a similar negative hedonic shock to earlier results, -0.47 (ps<.001); partisan losers anticipating
losing also experienced a negative impact (-0.88, ps<.05) that is larger but not statistically
different from the shock to other Republicans. Similar analysis for partisan winners shows
no meaningful difference in happiness gains. We must be cautious in drawing strong
conclusions from such a sample of Republicans expecting to lose because it is both very
small and may be fundamentally different than other Republicans. Still, these results cast
some doubt on the alternative explanation of unmet expectations.
!
1

Before After Dierence T-Stat Coe T-Stat
Self-Report: Happy today? 0.613 0.529 -0.084 7.010 -0.143 -6.065
Self-Report: Sad today? 0.076 0.169 0.094 -11.792 0.177 10.308
Democrat 0.223 0.236 0.013 -1.250 0.003 0.144
Independent 0.328 0.347 0.019 -1.674 0.018 0.775
Republican 0.449 0.417 -0.032 2.661 -0.021 -0.813
Female 0.425 0.419 -0.006 0.507 0.004 0.156
Age: Under 18 0.006 0.008 0.002 -0.783 0.001 0.506
Age: 18-24 0.020 0.026 0.006 -1.662 -0.006 -0.749
Age: 25-29 0.022 0.029 0.007 -1.917 0.003 0.329
Age: 30-34 0.037 0.043 0.006 -1.284 0.005 0.480
Age: 35-44 0.119 0.108 -0.011 1.423 -0.018 -1.277
Age: 45-54 0.251 0.233 -0.018 1.738 -0.033 -1.502
Age: 55-64 0.295 0.298 0.004 -0.324 0.006 0.288
Age: 65 & Older 0.251 0.256 0.004 -0.416 0.042 2.095
Caucasian 0.864 0.855 -0.009 1.051 -0.014 -0.867
Hispanic 0.029 0.029 -0.001 0.175 0.002 0.229
African American 0.044 0.046 0.002 -0.391 -0.001 -0.052
American Indian 0.012 0.017 0.004 -1.499 -0.001 -0.193
Aleut Eskimo 0.006 0.005 0.000 0.149 0.002 0.796
Asian or Pacic Islander 0.009 0.010 0.001 -0.229 -0.006 -1.349
Other 0.035 0.038 0.003 -0.651 0.017 1.677
Income: Under $25k 0.127 0.126 -0.001 0.165 -0.003 -0.205
Income: $25k - $35k 0.094 0.094 -0.001 0.078 -0.017 -1.406
Income: $35k -$50k 0.156 0.166 0.010 -1.159 -0.015 -0.797
Income: $50k - $75k 0.205 0.199 -0.006 0.574 -0.012 -0.570
Income: $75k - $100k 0.169 0.161 -0.007 0.825 0.011 0.697
Income: $100k - $125k 0.096 0.102 0.006 -0.815 0.031 2.062
Income: $125k - $150k 0.053 0.059 0.006 -0.993 0.017 1.837
Income: $150k or more 0.100 0.093 -0.007 0.962 -0.012 -0.739
Observations 3,483 3,290
Note: Standard deviations omitted because dummies means and observations describe all moments of the distribution.
Columns (3) & (4) calculate the raw dierences. Columns (5) & (6) apply a discontinuity specication from to equation
(1) using the set of controls as the dependent variables. Coe corresponds to the coecient, similar to specication (3) in
table (1). See supplemental gure S1A for the distribution of the t-statistics on the dierences and S1B for the distribution
of the t-statistics on the coecients.
Table S1A: Summary statistics one week before and after the the elec-
tion
2

Before After Dierence T-Stat Coe T-Stat
Self-Report: Happy today? 0.650 0.683 0.033 -1.381 0.037 0.794
Self-Report: Sad today? 0.057 0.055 -0.001 0.104 0.014 0.602
Democrat 1.000 1.000 0.000 0.000
Independent 0.000 0.000 0.000 0.000
Republican 0.000 0.000 0.000 0.000
Female 0.557 0.534 -0.024 0.940 0.034 0.792
Age: Under 18 0.010 0.008 -0.003 0.534 0.011 1.599
Age: 18-24 0.032 0.041 0.009 -0.949 -0.026 -1.627
Age: 25-29 0.031 0.034 0.003 -0.292 0.015 0.999
Age: 30-34 0.041 0.064 0.023 -2.050 0.008 0.292
Age: 35-44 0.129 0.134 0.005 -0.310 -0.022 -0.723
Age: 45-54 0.264 0.241 -0.023 1.036 0.016 0.427
Age: 55-64 0.318 0.302 -0.016 0.696 -0.015 -0.424
Age: 65 & Older 0.175 0.177 0.002 -0.078 0.012 0.352
Caucasian 0.767 0.767 0.000 0.014 -0.001 -0.013
Hispanic 0.049 0.036 -0.013 1.253 -0.030 -1.170
African American 0.134 0.135 0.001 -0.084 0.007 0.220
American Indian 0.006 0.008 0.001 -0.305 0.003 0.362
Aleut Eskimo 0.005 0.006 0.001 -0.336 0.000 0.056
Asian or Pacic Islander 0.009 0.013 0.004 -0.734 0.004 0.513
Other 0.030 0.035 0.005 -0.579 0.016 0.990
Income: Under $25k 0.147 0.142 -0.005 0.278 0.042 1.146
Income: $25k - $35k 0.112 0.111 -0.001 0.072 -0.039 -1.494
Income: $35k -$50k 0.167 0.144 -0.023 1.248 -0.049 -1.354
Income: $50k - $75k 0.229 0.227 -0.002 0.107 -0.022 -0.480
Income: $75k - $100k 0.148 0.165 0.017 -0.919 0.056 1.696
Income: $100k - $125k 0.084 0.086 0.003 -0.190 0.028 1.066
Income: $125k - $150k 0.035 0.046 0.012 -1.162 0.003 0.174
Income: $150k or more 0.079 0.079 0.000 -0.007 -0.019 -0.797
Observations 777 776
Note: Standard deviations omitted because dummies means and observations describe all moments of the distribution.
Columns (3) & (4) calculate the raw dierences. Columns (5) & (6) apply a discontinuity specication from to equation
(1) using the set of controls as the dependent variables. Coe corresponds to the coecient, similar to specication (3) in
table (1). See supplemental gure S1C for the distribution of the t-statistics on the dierences and S1D for the distribution
of the t-statistics on the coecients.
Table S1B: Summary statistics one week before and after the the elec-
tion (Democrats only)
3

Before After Dierence T-Stat Coe T-Stat
Self-Report: Happy today? 0.615 0.464 -0.151 8.270 -0.243 -6.846
Self-Report: Sad today? 0.075 0.225 0.149 -11.379 0.273 9.634
Democrat 0.000 0.000 0.000 0.000
Independent 0.000 0.000 0.000 0.000
Republican 1.000 1.000 0.000 0.000
Female 0.393 0.377 -0.016 0.874 -0.006 -0.156
Age: Under 18 0.004 0.010 0.006 -2.033 0.006 1.624
Age: 18-24 0.013 0.018 0.005 -1.037 0.010 0.967
Age: 25-29 0.019 0.027 0.008 -1.395 -0.007 -0.557
Age: 30-34 0.032 0.034 0.002 -0.237 0.000 -0.037
Age: 35-44 0.120 0.093 -0.028 2.433 -0.044 -1.969
Age: 45-54 0.245 0.225 -0.020 1.301 -0.059 -1.853
Age: 55-64 0.298 0.287 -0.011 0.683 0.009 0.242
Age: 65 & Older 0.269 0.308 0.039 -2.331 0.085 2.644
Caucasian 0.936 0.921 -0.015 1.548 -0.026 -1.333
Hispanic 0.019 0.023 0.003 -0.644 0.008 0.717
African American 0.004 0.013 0.009 -2.465 0.009 1.416
American Indian 0.006 0.012 0.005 -1.492 0.003 0.386
Aleut Eskimo 0.004 0.005 0.001 -0.511 0.009 2.218
Asian or Pacic Islander 0.010 0.007 -0.002 0.683 -0.003 -0.479
Other 0.020 0.019 -0.002 0.294 0.001 0.062
Income: Under $25k 0.094 0.114 0.020 -1.743 0.022 1.050
Income: $25k - $35k 0.081 0.085 0.004 -0.392 -0.007 -0.430
Income: $35k -$50k 0.146 0.177 0.031 -2.247 -0.010 -0.424
Income: $50k - $75k 0.205 0.201 -0.004 0.279 0.003 0.107
Income: $75k - $100k 0.198 0.150 -0.047 3.400 -0.032 -1.128
Income: $100k - $125k 0.103 0.105 0.002 -0.179 0.006 0.249
Income: $125k - $150k 0.065 0.064 -0.001 0.118 0.020 1.409
Income: $150k or more 0.109 0.105 -0.004 0.327 -0.002 -0.115
Observations 1,563 1,371
Note: Standard deviations omitted because dummies means and observations describe all moments of the distribution.
Columns (3) & (4) calculate the raw dierences. Columns (5) & (6) apply a discontinuity specication from to equation
(1) using the set of controls as the dependent variables. Coe corresponds to the coecient, similar to specication (3) in
table (1). See supplemental gure S1E for the distribution of the t-statistics on the dierences and S1F for the distribution
of the t-statistics on the coecients.
Table S1C: Summary statistics one week before and after the the elec-
tion (Republicans only)
4

Before After Dierence T-Stat Coe T-Stat
Self-Report: Happy today? 0.584 0.571 -0.014 1.012 -0.076 -2.627
Self-Report: Sad today? 0.082 0.091 0.008 -1.086 0.030 2.045
Parent 0.774 0.768 -0.006 0.520 -0.016 -0.784
Female 0.388 0.376 -0.012 0.871 0.007 0.281
Age: Under 18 0.006 0.006 0.001 -0.308 0.004 0.804
Age: 18-24 0.023 0.021 -0.001 0.371 -0.001 -0.152
Age: 25-29 0.028 0.025 -0.003 0.700 -0.008 -0.776
Age: 30-34 0.031 0.038 0.007 -1.467 0.001 0.110
Age: 35-44 0.092 0.098 0.006 -0.787 -0.003 -0.190
Age: 45-54 0.200 0.223 0.023 -2.007 0.028 1.019
Age: 55-64 0.300 0.306 0.006 -0.469 0.019 0.714
Age: 65 & Older 0.321 0.283 -0.038 3.031 -0.040 -1.394
Caucasian 0.856 0.859 0.003 -0.325 0.003 0.115
Hispanic 0.019 0.022 0.003 -0.784 0.004 0.516
African American 0.056 0.047 -0.008 1.321 -0.019 -1.529
American Indian 0.013 0.018 0.005 -1.382 -0.005 -0.723
Aleut Eskimo 0.006 0.003 -0.003 1.841 0.002 0.325
Asian or Pacic Islander 0.007 0.009 0.001 -0.566 0.000 0.026
Other 0.043 0.043 -0.001 0.128 0.016 1.291
Income: Under $25k 0.146 0.145 -0.002 0.171 0.040 2.392
Income: $25k - $35k 0.117 0.113 -0.004 0.433 0.031 1.720
Income: $35k -$50k 0.167 0.166 -0.001 0.086 0.022 1.032
Income: $50k - $75k 0.203 0.202 -0.001 0.065 -0.009 -0.398
Income: $75k - $100k 0.148 0.153 0.005 -0.476 -0.039 -1.577
Income: $100k - $125k 0.082 0.084 0.002 -0.316 -0.024 -1.612
Income: $125k - $150k 0.054 0.049 -0.005 0.763 -0.042 -3.233
Income: $150k or more 0.084 0.088 0.005 -0.598 0.021 1.203
Observations 2,630 2,674
Note: Standard deviations omitted because dummies means and observations describe all moments of the distribution.
Columns (3) & (4) calculate the raw dierences. Columns (5) & (6) apply a discontinuity specication from to equation
(1) using the set of controls as the dependent variables. Coe corresponds to the coecient, similar to specication (3) in
table (1). See supplemental gure S2A for the distribution of the t-statistics on the dierences and S2B for the distribution
of the t-statistics on the coecients.
Table S2A: Summary statistics before and after the Newtown shootings
5

Before After Dierence T-Stat Coe T-Stat
Self-Report: Happy today? 0.481 0.499 0.019 -0.645 0.056 1.067
Self-Report: Sad today? 0.113 0.130 0.018 -0.951 0.023 0.564
Parent 0.000 0.000 0.000 0.000
Female 0.370 0.372 0.002 -0.081 -0.033 -0.566
Age: Under 18 0.024 0.024 0.001 -0.071 0.017 0.802
Age: 18-24 0.086 0.079 -0.007 0.431 -0.011 -0.295
Age: 25-29 0.086 0.079 -0.007 0.431 -0.050 -1.524
Age: 30-34 0.055 0.072 0.017 -1.213 0.022 0.958
Age: 35-44 0.116 0.106 -0.010 0.537 -0.025 -0.544
Age: 45-54 0.190 0.237 0.047 -1.995 0.047 1.036
Age: 55-64 0.259 0.254 -0.004 0.175 0.024 0.384
Age: 65 & Older 0.185 0.148 -0.037 1.718 -0.024 -0.571
Caucasian 0.844 0.820 -0.024 1.121 -0.090 -1.525
Hispanic 0.037 0.026 -0.011 1.119 0.008 0.357
African American 0.040 0.050 0.010 -0.805 0.007 0.259
American Indian 0.013 0.018 0.004 -0.601 -0.008 -0.447
Aleut Eskimo 0.013 0.006 -0.007 1.226 -0.004 -0.255
Asian or Pacic Islander 0.008 0.014 0.006 -1.000 0.002 0.124
Other 0.044 0.066 0.022 -1.713 0.085 3.135
Income: Under $25k 0.218 0.203 -0.016 0.666 0.059 1.392
Income: $25k - $35k 0.118 0.124 0.006 -0.339 -0.003 -0.105
Income: $35k -$50k 0.175 0.184 0.009 -0.399 0.057 1.438
Income: $50k - $75k 0.175 0.172 -0.002 0.114 -0.049 -1.145
Income: $75k - $100k 0.118 0.111 -0.007 0.358 -0.005 -0.158
Income: $100k - $125k 0.071 0.093 0.023 -1.451 0.021 0.615
Income: $125k - $150k 0.040 0.050 0.010 -0.805 -0.033 -1.289
Income: $150k or more 0.086 0.063 -0.023 1.521 -0.047 -1.432
Observations 595 621
Note: Standard deviations omitted because dummies means and observations describe all moments of the distribution.
Columns (3) & (4) calculate the raw dierences. Columns (5) & (6) apply a discontinuity specication from to equation
(1) using the set of controls as the dependent variables. Coe corresponds to the coecient, similar to specication (3) in
table (1). See supplemental gure S2C for the distribution of the t-statistics on the dierences and S2D for the distribution
of the t-statistics on the coecients.
Table S2B: Summary statistics before and after the Newtown shootings
(Non-Parents only)
6

Before After Dierence T-Stat Coe T-Stat
Self-Report: Happy today? 0.615 0.592 -0.022 1.466 -0.115 -3.506
Self-Report: Sad today? 0.073 0.078 0.005 -0.628 0.031 1.881
Parent 1.000 1.000 0.000 0.000
Female 0.393 0.377 -0.016 1.026 0.019 0.708
Age: Under 18 0.000 0.001 0.000 -0.571 0.000 -0.137
Age: 18-24 0.004 0.004 -0.001 0.261 0.000 -0.043
Age: 25-29 0.011 0.008 -0.003 0.832 0.003 0.530
Age: 30-34 0.024 0.028 0.004 -0.844 -0.006 -0.607
Age: 35-44 0.085 0.096 0.011 -1.220 0.004 0.202
Age: 45-54 0.203 0.219 0.015 -1.196 0.022 0.702
Age: 55-64 0.312 0.321 0.009 -0.649 0.019 0.643
Age: 65 & Older 0.361 0.323 -0.037 2.512 -0.041 -1.262
Caucasian 0.859 0.870 0.011 -1.071 0.032 1.411
Hispanic 0.014 0.021 0.007 -1.760 0.003 0.374
African American 0.060 0.047 -0.013 1.876 -0.027 -1.632
American Indian 0.013 0.018 0.005 -1.246 -0.004 -0.586
Aleut Eskimo 0.004 0.002 -0.002 1.402 0.003 0.797
Asian or Pacic Islander 0.007 0.007 0.000 0.023 0.000 -0.093
Other 0.043 0.036 -0.008 1.263 -0.006 -0.486
Income: Under $25k 0.125 0.127 0.002 -0.176 0.032 1.679
Income: $25k - $35k 0.116 0.110 -0.007 0.693 0.041 1.877
Income: $35k -$50k 0.165 0.161 -0.004 0.336 0.011 0.485
Income: $50k - $75k 0.211 0.211 0.000 -0.007 0.004 0.165
Income: $75k - $100k 0.157 0.165 0.008 -0.728 -0.048 -1.727
Income: $100k - $125k 0.085 0.081 -0.004 0.424 -0.037 -2.471
Income: $125k - $150k 0.057 0.049 -0.009 1.252 -0.045 -3.275
Income: $150k or more 0.083 0.096 0.013 -1.446 0.042 2.037
Observations 2,035 2,053
Note: Standard deviations omitted because dummies means and observations describe all moments of the distribution.
Columns (3) & (4) calculate the raw dierences. Columns (5) & (6) apply a discontinuity specication from to equation
(1) using the set of controls as the dependent variables. Coe corresponds to the coecient, similar to specication (3) in
table (1). See supplemental gure S2E for the distribution of the t-statistics on the dierences and S2F for the distribution
of the t-statistics on the coecients.
Table S2C: Summary statistics before and after the Newtown shootings
(Parents only)
7

Before After Dierence T-Stat Coe T-Stat
Self-Report: Happy today? 0.619 0.596 -0.022 2.155 -0.048 -2.162
Self-Report: Sad today? 0.070 0.077 0.007 -1.284 0.006 0.470
Boston Region 0.019 0.021 0.002 -0.607 0.001 0.123
Female 0.445 0.439 -0.005 0.504 -0.014 -0.650
Age: Under 18 0.019 0.018 -0.001 0.290 -0.016 -2.758
Age: 18-24 0.049 0.053 0.004 -0.832 0.010 0.999
Age: 25-29 0.051 0.051 0.000 -0.023 0.005 0.516
Age: 30-34 0.055 0.049 -0.006 1.378 0.015 1.598
Age: 35-44 0.134 0.143 0.010 -1.301 0.038 2.501
Age: 45-54 0.222 0.232 0.010 -1.170 0.030 1.914
Age: 55-64 0.261 0.245 -0.016 1.693 -0.026 -1.331
Age: 65 & Older 0.210 0.209 -0.001 0.112 -0.057 -3.359
Caucasian 0.839 0.838 -0.001 0.125 -0.001 -0.058
Hispanic 0.034 0.031 -0.003 0.712 0.006 0.809
African American 0.055 0.056 0.001 -0.182 -0.003 -0.258
American Indian 0.014 0.012 -0.002 0.837 -0.003 -0.705
Aleut Eskimo 0.007 0.006 -0.001 0.392 -0.003 -0.989
Asian or Pacic Islander 0.014 0.009 -0.005 2.023 -0.004 -0.919
Other 0.038 0.048 0.010 -2.348 0.008 0.940
Income: Under $25k 0.140 0.151 0.011 -1.420 -0.035 -2.589
Income: $25k - $35k 0.100 0.104 0.004 -0.623 0.003 0.275
Income: $35k -$50k 0.157 0.157 0.000 -0.054 0.011 0.698
Income: $50k - $75k 0.204 0.194 -0.011 1.273 0.002 0.097
Income: $75k - $100k 0.161 0.160 -0.001 0.120 0.022 1.470
Income: $100k - $125k 0.095 0.091 -0.004 0.635 -0.017 -1.088
Income: $125k - $150k 0.058 0.058 -0.001 0.172 0.007 0.675
Income: $150k or more 0.086 0.087 0.001 -0.243 0.007 0.582
Observations 4,316 4,623
Note: Standard deviations omitted because dummies means and observations describe all moments of the distribution.
Columns (3) & (4) calculate the raw dierences. Columns (5) & (6) apply a discontinuity specication from to equation
(1) using the set of controls as the dependent variables. Coe corresponds to the coecient, similar to specication (3) in
table (1). See supplemental gure S3A for the distribution of the t-statistics on the dierences and S3B for the distribution
of the t-statistics on the coecients.
Table S3A: Summary statistics before and after the Boston bombings
8

Before After Dierence T-Stat Coe T-Stat
Self-Report: Happy today? 0.618 0.597 -0.022 2.077 -0.047 -2.065
Self-Report: Sad today? 0.069 0.076 0.007 -1.185 0.004 0.330
Boston Region 0.000 0.000 0.000 0.000
Female 0.442 0.437 -0.005 0.469 -0.010 -0.486
Age: Under 18 0.019 0.017 -0.002 0.745 -0.016 -2.830
Age: 18-24 0.048 0.053 0.005 -1.043 0.013 1.314
Age: 25-29 0.049 0.050 0.002 -0.369 0.005 0.459
Age: 30-34 0.055 0.049 -0.006 1.309 0.014 1.459
Age: 35-44 0.136 0.144 0.009 -1.200 0.038 2.468
Age: 45-54 0.223 0.231 0.008 -0.876 0.031 1.910
Age: 55-64 0.261 0.246 -0.015 1.628 -0.027 -1.391
Age: 65 & Older 0.209 0.209 0.000 -0.021 -0.056 -3.328
Caucasian 0.839 0.839 -0.001 0.111 -0.001 -0.040
Hispanic 0.033 0.031 -0.002 0.566 0.006 0.736
African American 0.055 0.057 0.002 -0.354 -0.003 -0.264
American Indian 0.014 0.012 -0.002 1.015 -0.003 -0.629
Aleut Eskimo 0.007 0.006 -0.001 0.518 -0.004 -1.313
Asian or Pacic Islander 0.014 0.009 -0.005 2.148 -0.003 -0.591
Other 0.038 0.047 0.009 -2.193 0.008 0.860
Income: Under $25k 0.140 0.151 0.011 -1.407 -0.033 -2.416
Income: $25k - $35k 0.100 0.105 0.005 -0.741 0.004 0.338
Income: $35k -$50k 0.155 0.157 0.001 -0.184 0.010 0.657
Income: $50k - $75k 0.205 0.193 -0.012 1.398 0.001 0.079
Income: $75k - $100k 0.162 0.161 -0.001 0.153 0.022 1.446
Income: $100k - $125k 0.095 0.091 -0.004 0.669 -0.019 -1.168
Income: $125k - $150k 0.058 0.057 0.000 0.086 0.007 0.723
Income: $150k or more 0.085 0.086 0.001 -0.151 0.006 0.502
Observations 4,235 4,528
Note: Standard deviations omitted because dummies means and observations describe all moments of the distribution.
Columns (3) & (4) calculate the raw dierences. Columns (5) & (6) apply a discontinuity specication from to equation
(1) using the set of controls as the dependent variables. Coe corresponds to the coecient, similar to specication (3) in
table (1). See supplemental gure S3C for the distribution of the t-statistics on the dierences and S3D for the distribution
of the t-statistics on the coecients.
Table S3B: Summary statistics before and after the Boston bombings
(Non-Boston region only)
9

Before After Dierence T-Stat Coe T-Stat
Self-Report: Happy today? 0.642 0.589 -0.053 0.711 -0.165 -1.881
Self-Report: Sad today? 0.086 0.116 0.029 -0.645 0.119 3.567
Boston Region 1.000 1.000 0.000 0.000
Female 0.580 0.547 -0.033 0.436 -0.245 -3.996
Age: Under 18 0.000 0.063 0.063 -2.517 0.014 0.235
Age: 18-24 0.123 0.074 -0.050 1.092 -0.206 -1.867
Age: 25-29 0.148 0.063 -0.085 1.809 0.044 1.965
Age: 30-34 0.062 0.042 -0.020 0.578 0.124 1.030
Age: 35-44 0.037 0.084 0.047 -1.326 0.060 1.013
Age: 45-54 0.136 0.274 0.138 -2.304 0.023 0.165
Age: 55-64 0.222 0.189 -0.033 0.532 0.067 0.738
Age: 65 & Older 0.272 0.211 -0.061 0.938 -0.126 -0.509
Caucasian 0.802 0.800 -0.002 0.041 -0.041 -0.520
Hispanic 0.074 0.042 -0.032 0.891 0.049 0.839
African American 0.062 0.021 -0.041 1.324 0.008 0.155
American Indian 0.000 0.021 0.021 -1.422 -0.018 -1.631
Aleut Eskimo 0.012 0.021 0.009 -0.452 0.068 1.098
Asian or Pacic Islander 0.025 0.032 0.007 -0.275 -0.115 -1.213
Other 0.025 0.063 0.038 -1.261 0.049 1.704
Income: Under $25k 0.148 0.158 0.010 -0.178 -0.170 -2.665
Income: $25k - $35k 0.086 0.053 -0.034 0.867 -0.053 -1.061
Income: $35k -$50k 0.222 0.168 -0.054 0.890 0.046 0.416
Income: $50k - $75k 0.173 0.221 0.048 -0.801 0.021 0.119
Income: $75k - $100k 0.086 0.105 0.019 -0.423 0.028 0.845
Income: $100k - $125k 0.074 0.084 0.010 -0.247 0.100 5.703
Income: $125k - $150k 0.086 0.063 -0.023 0.579 -0.037 -0.170
Income: $150k or more 0.123 0.147 0.024 -0.461 0.066 0.494
Observations 81 95
Note: Standard deviations omitted because dummies means and observations describe all moments of the distribution.
Columns (3) & (4) calculate the raw dierences. Columns (5) & (6) apply a discontinuity specication from to equation
(1) using the set of controls as the dependent variables. Coe corresponds to the coecient, similar to specication (3) in
table (1). See supplemental gure S3E for the distribution of the t-statistics on the dierences and S3F for the distribution
of the t-statistics on the coecients.
Table S3C: Summary statistics before and after the Boston bombings
(Boston region only)
10

Independent variables (1) (2) (3) (4) (5)
Post-election
.036
(.030)
-.001
(.011)
.014
(.023)
.013
(.026)
-.051
(.056)
Post-election * One de-
gree polynomial of days
No No Yes Yes Yes
Post-election * ree de-
gree polynomial of days
No No No No Yes
Socio - Demographic &
MSA Controls
No No No Yes Yes
Time restriction +/- one day +/- one week +/- one week +/- one week +/- one week
Observations 265 1,553 1,553 1,553 1,553
Note: *, **, and *** indicate signicance at the 10%, 5%, and 1% condence levels, respectively. Standard errors are clus-
tered at the Metropolitan level. Socio-demographic controls include gender, age indicators, race indicators, and income
indicators. MSA controls included indicators for the metropolitan statistical area.
Table S4A: Democrat sadness one week surrounding 2012 election
Dependent variable: Are you sad today?
Independent variables (1) (2) (3) (4) (6)
Post-election
.423***
(.039)
.149***
(.012)
.273***
(.028)
.268***
(.031)
.450***
(.064)
Post-election * One de-
gree polynomial of days
No No Yes Yes No
Post-election * ree de-
gree polynomial of days
No No No No Yes
Socio - Demographic &
MSA Controls
No No No Yes Yes
Time restriction +/- one day +/- one week +/- one week +/- one week +/- one week
Observations 465 2,934 2,934 2,934 2,934
Table S4B: Republican sadness one week surrounding 2012 election
Dependent variable: Are you sad today?
11

Independent variables (1) (2) (3) (4) (5) (6) (7)
Post-Newtown
.032
(.021)
.008
(.006)
.030**
(.015)
.024
(.016)
.002
(.043)
-.003
(.042)
.029
(.018)
Post-Newtown * One degree
polynomial of days
No No Yes Yes No Yes Yes
Post-Newtown * ree degree
polynomial of days
No No No No Yes No No
Socio - Demographic & MSA
Controls
No No No Yes Yes Yes Yes
Time restriction +/- one day +/- one week +/- one week +/- one week +/- one week +/- one week +/- one week
Parents & Non-Parents Both Both Both Both Both Non-Parents Parents
Observations 695 5,304 5,304 5,304 5,304 1,216 4,088
Note: *, **, and *** indicate signicance at the 10%, 5%, and 1% condence levels, respectively. Standard errors are clustered at the Metropolitan level. Socio-demograph-
ic controls include gender, age indicators, race indicators, and income indicators. MSA controls included indicators for the metropolitan statistical area.
Table S5: Self-reported sadness before and after Newtown shooting
Dependent variable: Are you sad today?
12

Independent variables (1) (2) (3) (4) (5) (6) (7)
Post-Boston
.005
(.014)
.007
(.005)
.006
(.013)
.010
(.013)
-.003
(.033)
.008
(.013)
.090
(.062)
Post-Boston * One degree poly-
nomial of days
No No Yes Yes No Yes Yes
Post-Boston * ree degree poly-
nomial of days
No No No No Yes No No
Socio - Demographic & MSA
Controls
No No No Yes Yes Yes Yes
Time restriction +/- one day +/- one week +/- one week +/- one week +/- one week +/- one week +/- one week
Boston Region & Non-Boston
Region
Both Both Both Both Both Non-Boston Boston
Observations 1,360 8,939 8,939 8,939 8,939 8,763 176
Note: *, **, and *** indicate signicance at the 10%, 5%, and 1% condence levels, respectively. Standard errors are clustered at the Metropolitan level. Socio-demograph-
ic controls include gender, age indicators, race indicators, and income indicators. MSA controls included indicators for the metropolitan statistical area.
Table S6: Self-reported sadness before and after Boston bombing
Dependent variable: Are you sad today?
13

Independent variables (1) (2) (3) (4) (5)
Post-election
.073
(.106)
.073*
(.037)
.088
(.096)
.097
(.110)
.226
(.247)
Post-election * One de-
gree polynomial of days
No No Yes Yes No
Post-election * ree de-
gree polynomial of days
No No No No Yes
Socio - Demographic &
MSA Controls
No No No Yes Yes
Time restriction +/- one day +/- one week +/- one week +/- one week +/- one week
Observations 265 1,553 1,553 1,553 1,553
Note: *, **, and *** indicate signicance at the 10%, 5%, and 1% condence levels, respectively. Standard errors are clus-
tered at the Metropolitan level. Socio-demographic controls include gender, age indicators, race indicators, and income
indicators. MSA controls included indicators for the metropolitan statistical area. Happiness on scale of 1-5 where 1 is very
unhappy, 2 unhappy, 3 neither happy nor sad, 4 happy, and 5 very happy.
Table S7A: Democrat happiness one week surrounding 2012 election
Dependent variable: Happiness on scale of 1-5
Independent variables (1) (2) (3) (4) (6)
Post-election
-1.139***
(.104)
-.448***
(.034)
-.796***
(.074)
-.789***
(.081)
-1.231***
(.217)
Post-election * One de-
gree polynomial of days
No No Yes Yes No
Post-election * ree de-
gree polynomial of days
No No No No Yes
Socio - Demographic &
MSA Controls
No No No Yes Yes
Time restriction +/- one day +/- one week +/- one week +/- one week +/- one week
Observations 465 2,934 2,934 2,934 2,934
Table S7B: Republican happiness one week surrounding 2012 election
Dependent variable: Happiness on scale of 1-5
14

0 .1 .2 .3 .4
-2 -1 0 1 2 3
T-statistics of coefficients from Supplemental Table 1A
0 .1 .2 .3 .4
-2 -1 0 1 2 3
T-statistics of differences from Supplemental Table 1A
Figure S1A: Kernel density distribution for mean differences t-statistics
from Table S1A
Figure S1B: Kernel density distribution for coefcient t-statistics from
Table S1A
Note: Normal approximation shown with kernel density estimate. Variables Happy & Sad omitted from density estimation
because they are dependent variables.
15

0 .1 .2 .3 .4
-2 -1 0 1 2
T-statistics of coefficients from Supplemental Table 1B
0 .2 .4 .6
-2 -1 0 1 2
T-statistics of differences from Supplemental Table 1B
Figure S1C: Kernel density distribution for mean differences t-statistics
from Table S1B
Figure S1D: Kernel density distribution for coefcient t-statistics from
Table S1B
Note: Normal approximation shown with kernel density estimate. Variables Happy & Sad omitted from density estimation
because they are dependent variables.
16

0 .1 .2 .3 .4
-2 0 2 4
T-statistics of coefficients from Supplemental Table 1C
0 .05 .1 .15 .2 .25
-4 -2 0 2 4
T-statistics of differences from Supplemental Table 1C
Figure S1E: Kernel density distribution for mean differences t-statistics
from Table S1C
Figure S1F: Kernel density distribution for coefcient t-statistics from
Table S1C
Note: Normal approximation shown with kernel density estimate. Variables Happy & Sad omitted from density estimation
because they are dependent variables.
17

0 .1 .2 .3
-4 -2 0 2 4
T-statistics of coefficients from Supplemental Table 2A
0 .1 .2 .3 .4
-2 0 2 4
T-statistics of differences from Supplemental Table 2A
Figure S2A: Kernel density distribution for mean differences t-statistics
from Table S2A
Figure S2B: Kernel density distribution for coefcient t-statistics from
Table S2A
Note: Normal approximation shown with kernel density estimate. Variables Happy & Sad omitted from density estimation
because they are dependent variables.
18

0 .1 .2 .3 .4
-2 0 2 4
T-statistics of coefficients from Supplemental Table 2B
0 .1 .2 .3 .4
-3 -2 -1 0 1 2
T-statistics of differences from Supplemental Table 2B
Figure S2C: Kernel density distribution for mean differences t-statistics
from Table S2B
Figure S2D: Kernel density distribution for coefcient t-statistics from
Table S2B
Note: Normal approximation shown with kernel density estimate. Variables Happy & Sad omitted from density estimation
because they are dependent variables.
19

0 .1 .2 .3 .4
-4 -2 0 2
T-statistics of coefficients from Supplemental Table 2C
0 .1 .2 .3 .4
-2 -1 0 1 2 3
T-statistics of differences from Supplemental Table 2C
Figure S2E: Kernel density distribution for mean differences t-statistics
from Table S2C
Figure S2F: Kernel density distribution for coefcient t-statistics from
Table S2C
Note: Normal approximation shown with kernel density estimate. Variables Happy & Sad omitted from density estimation
because they are dependent variables.
20

0 .1 .2 .3
-4 -2 0 2 4
T-statistics of coefficients from Supplemental Table 3A
0 .1 .2 .3 .4
-4 -2 0 2
T-statistics of differences from Supplemental Table 3A
Figure S3A: Kernel density distribution for mean differences t-statistics
from Table S3A
Figure S3B: Kernel density distribution for coefcient t-statistics from
Table S3A
Note: Normal approximation shown with kernel density estimate. Variables Happy & Sad omitted from density estimation
because they are dependent variables.
21

0 .1 .2 .3
-4 -2 0 2 4
T-statistics of coefficients from Supplemental Table 3B
0 .1 .2 .3 .4
-4 -2 0 2 4
T-statistics of differences from Supplemental Table 3B
Figure S3C: Kernel density distribution for mean differences t-statistics
from Table S3B
Figure S3D: Kernel density distribution for coefcient t-statistics from
Table S3B
Note: Normal approximation shown with kernel density estimate. Variables Happy & Sad omitted from density estimation
because they are dependent variables.
22

0 .1 .2 .3
-4 -2 0 2 4 6
T-statistics of coefficients from Supplemental Table 3C
0 .1 .2 .3 .4
-3 -2 -1 0 1 2
T-statistics of differences from Supplemental Table 3C
Figure S3E: Kernel density distribution for mean differences t-statistics
from Table S3C
Figure S3F: Kernel density distribution for coefcient t-statistics from
Table S3C
Note: Normal approximation shown with kernel density estimate. Variables Happy & Sad omitted from density estimation
because they are dependent variables.
23

-.5 -.4 -.3 -.2 -.1 0
Coefficient estimate on Post-Election for Republicans
0 5 10 15 20 25 30
Days before and after Election
-.2 -.1 0 .1 .2 .3
Coefficient estimate on Post-Election for Democrats
0 5 10 15 20 25 30
Days before and after Election
Supplemental Figure S4A: Table (1A) column (4) regression where time
window changes
Supplemental Figure S4B: Table (1B) column (4) regression where time
window changes
24

-.5 -.4 -.3 -.2 -.1 0
Coefficient estimate on Post-Boston
0 5 10 15 20 25 30
Days before and after Boston bombings
-.5 -.4 -.3 -.2 -.1 0
Coefficient estimate on Post-Newtown
0 5 10 15 20 25 30
Days before and after Newtown shootings
Supplemental Figure S4C: Table (2) column (4) regression where time
window changes
Supplemental Figure S4D: Table (3) column (4) regression where time
window changes
25

-.15 -.1 -.05 0 .05 .1
Week Coefficient (Baseline is 8th Week Before Election)
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Weeks before and after Election
-.1 -.05 0 .05 .1
Week Coefficient (Baseline is 8th Week Before Election)
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Weeks before and after Election
Supplemental Figure S5A: Democrat self-reported emotional state two
months before and after the election: “Are you happy today?”
Supplemental Figure S5B: Republican self-reported emotional state two
months before and after the election: “Are you happy today?”
Democrats before and
after 2012 election
Republicans before and
after 2012 election
Regression coecents on week indicator from
equation (1) for the set of Democrats in the data-
set. Week 1 is the baseline. Regressions include no
controls and are clustered at the MSA level.
Drop in happiness for Re-
publicans in the rst week
post-election is disipates in
the subsequent weeks.
Regression coecents on week indicator from
equation (1) for the set of Democrats in the data-
set. Week 1 is the baseline. Regressions include no
controls and are clustered at the MSA level.
26

250 300 350 400 450
Total Responses
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Days before and after Newtown shooting
300 400 500 600 700
Total Responses
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Days before and after election
Supplemental Figure S6A: Count of responses before and after the Elec-
tion
Supplemental Figure S6B: Count of the responses before and after New-
town
27

400 600 800 1000
Total Responses
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Days before and after Boston Marathon bombing
Supplemental Figure S6C: Count of responses before and after the Bos-
ton
28

.45 .5 .55 .6 .65
(mean) happy
-10 -5 0 5 10
weekelect
Treated: predicted Treated: actual
Regression with Newey-West standard errors
Supplemental Figure S8A: Republican happiness two months before and
after the election: Weekly interrupted time series
Break has point estimate of
-.1247 and t-stat of -3.02 on 16
data points using interrupted
time series approach
29

.3 .4 .5 .6 .7 .8
(mean) happy
-50 0 50
dayelect1
Treated: predicted Treated: actual
Regression with Newey-West standard errors
Supplemental Figure S8B: Republican happiness two months before and
after the election: Daily interrupted time series
Break has point estimate of
-.1217 and t-stat of -5.15 on 120
data points using interrupted
time series approach
30