Upload
costas
View
213
Download
0
Embed Size (px)
Citation preview
Positive Social Pressure and Prosocial Motivation: Evidence froma Large-Scale Field Experiment on Voter Mobilization
Costas PanagopoulosFordham University
Political scientists are increasingly exploring the psychological underpinnings of voting behavior using fieldexperimental techniques. Research in psychology demonstrates that positive reinforcement—what I describe aspositive social pressure—motivates prosocial behavior. A distinctive feature of the current study is the focus onkey subgroups of voters, namely unmarried women and minorities. Attention to these voter subgroups allows usto build upon findings reported in previous studies that leave questions about the generalizability of the reportedeffects of positive social pressure to key demographic subgroups of voters largely unanswered. This articlereports the results of a large-scale randomized field experiment designed to investigate the impact of positivesocial pressure on voter turnout. The experiment was conducted during the November 2009 gubernatorialelection in New Jersey, and the results suggest positive social pressure mobilizes voters. Moreover, the effectsappear to be robust across subgroups of voters, including minorities and unmarried women, and both lower-and higher-propensity voters.
KEY WORDS: positive reinforcement, voting behavior, mobilization, field experiment, positive social pressure
Voting in elections, in the United States and elsewhere, persists despite the fact that the costsassociated with electoral participation typically outweigh the benefits (Downs, 1957). Researchershave offered numerous explanations for the seemingly irrational act of voting (Downs, 1957;Rosenstone & Hansen, 1993), but recent scholarship has focused on the psychological underpinningsof voting behavior. One line of research that has generated considerable interest of late is the role ofsocial pressure in promoting voting. Scholars argue that voting is perceived as a social norm and thatcitizens will engage in this activity more frequently when they are informed their behavior is publicand that it can—or will—be monitored or publicized in order to appear norm compliant (Gerber,Green, & Larimer, 2008; Panagopoulos, 2010). Several studies suggest that shame aversion or thethreat of being “called out” for norm-incompliant behavior (failure to vote, for instance) boostscompliance levels significantly (Gerber et al., 2008), but a series of field experiments suggests votersare also responsive to efforts that acknowledge compliance with social norms and invoke pride orother positive sentiments (Panagopoulos, 2010, 2011).
The attention to the impact of shame, pride, and other affective reactions on voting underscoresthe importance of triggering emotional stimuli on prosocial motivation. Theorists have long estab-lished that psychological mechanisms shape effort and behavior (Gollwitzer & Moskowitz, 1996),and there is an emerging consensus that emotions exert powerful influence over individuals’ politicalattitudes and behavior (Brader, 2005; Civettini & Redlawsk, 2009; Lodge & Taber, 2000, 2005;
Political Psychology, Vol. 34, No. 2, 2013doi: 10.1111/pops.12007
bs_bs_banner
265
0162-895X © 2013 International Society of Political PsychologyPublished by Wiley Periodicals, Inc., 350 Main Street, Malden, MA 02148, USA, 9600 Garsington Road, Oxford, OX4 2DQ,
and PO Box 378 Carlton South, 3053 Victoria, Australia
Marcus, Neuman, & MacKuen, 2000; Neuman, Crigler, & MacKuen, 2007; Redlawsk, 2002;Redlawsk, Civettini, & Lau, 2007). Theories of motivated reasoning ascribe a central role to hotcognition, the claim that all sociopolitical concepts are affect laden (Civettini & Redlawsk, 2009;Lodge & Taber, 2005, p. 456; Redlawsk, 2002; Redlawsk et al., 2007), and assert that “affectpermeates all [political] thinking and reasoning” (Lodge & Taber, 2005, p. 456). Extended todecisions about electoral participation, this view suggests voters’ choices about whether or not tovote may not necessarily be guided exclusively by rational, cognitive considerations but also byaffective or emotional influences. One mechanism by which emotions may exert influence is throughmemory; Civettini and Redlawsk (2009) show that voters are more likely to remember informationthat generates affective reactions as opposed to information for which subjects report no emotionalresponse. In fact, some researchers argue that a primary function of emotions is to initiate and guidegoal-oriented, including prosocial, behavior like voting (Barrett & Campos, 1987; Cosmides &Tooby, 2000; LeDoux, 1996). Manipulations that activate shameful feelings—what I term negativesocial pressure—in particular appear to stimulate prosocial behavior considerably (Bear, Manning,& Izard, 2003; Gerber et al., 2008), but there is also evidence that invoking positive emotions, suchas pride—which I view as positive social pressure—has the capacity to motivate prosocial behavior(Panagopoulos, 2010, 2011; Williams & DeSteno, 2008).
In this article, I focus on assessing the effectiveness of positive social pressure in stimulatingvoting. I report the results of a large-scale, randomized field experiment designed to test thehypothesis that positive social pressure stimulates voting in elections. The experiment, conducted inthe context of the November 2009 gubernatorial election in New Jersey, randomly assigned voters tobe exposed to a mailing that listed neighbors who had voted in prior elections as part of a “civichonor roll” and encouraged recipients to vote and to attain the distinction themselves. The treatmentwas designed to exert positive social pressure to vote. The results I describe below suggest positivesocial pressure promotes electoral participation.
A distinctive feature of the current study is the focus on key subgroups of voters, namelyunmarried women and minorities. While the experimental sample is also comprised of other types ofindividuals, there are adequate numbers of black, Latino, and unmarried female subjects to permit usto hone in on testing the effects of social norms on these demographic segments. Studies show thatminorities tend to vote less frequently, compared to white voters and to the electorate overall (Garcia& Sanchez, 2008; Leighley, 2001); And even as men and women overall often vote at comparablerates, unmarried women generally vote at a lower rate compared to married women and to theelectorate as a whole (Mankani & Roeder, 2010; Women’s Voices Women Vote, 2010). In the 2008general election, for instance, 60% of unmarried women voted compared to 71% of married women(Women’s Voices Women Vote, 2010). Moreover, the drop-off rates in voting for minorities andunmarried women in off-year (nonpresidential) elections are typically higher than for voters overall(Mankani & Roeder, 2010). These patterns make the emphasis on mechanisms that have the capacityto stimulate voting among these voters especially compelling. Since these demographic attributes arerelated to baseline propensities to vote, it is conceivable, as Arceneaux and Nickerson (2009) argue,that these voters react differently to positive social pressure or mobilization messages more generallycompared to other types of voters. Attention to these voter subgroups also allows us to build uponfindings reported in previous studies that leave questions about the generalizability of the reportedeffects of positive social pressure to key demographic subgroups of voters largely unanswered.
This article proceeds as follows. The following section provides a more detailed theoreticalcontext for the hypothesis that is developed and that will be empirically tested using field experi-mental techniques. The next sections describe the experimental setting and procedures and presentdetails about the experimental treatment. The experimental results are presented and analyzed in thefollowing section. Reflections about the implications of the experimental findings are advanced inthe final section.
266 Panagopoulos
Recognition and Motivation
There is considerable debate amongst social psychologists about the impact of rewards onbehavior. Some researchers assert extrinsic rewards depress intrinsic motivation to perform taskssuch as voting (Deci & Ryan, 1985), but there is greater consensus that intangible rewards likeverbal praise, positive feedback, reinforcement, and recognition are positive motivators (Deci,Koestner, & Ryan, 2001). Underlying this motivational capacity may be individuals’ innate desireto avoid shame or to experience pride. Mascolo and Fischer (1995) define pride as “[an emotion]generated by appraisals that one is responsible for a socially-valued outcome or for being asocially-valued person” (p. 66). Williams and DeSteno (2008) assert, “when feeling proud abouta recognized accomplishment, an individual might feel an incentive to pursue further action in thatvalued domain” (p. 1008). Conceptualized as such, pride is integrally connected to social surveil-lance and public attention. As Williams and DeSteno (2008) argue that “public recognition ofone’s abilities appears to be a central aspect of pride (cf. Gaines et al., 2005; Webster et al., 2003)and that “pride derives from social acclaim and motivates behaviors aimed at maintaining it”(p. 1014).
Recent scholarship suggests activating prideful emotions motivates prosocial activity like par-ticipating in elections (Gerber, Green, & Larimer, 2010; Panagopoulos, 2010). A randomized fieldexperiment conducted by Panagopoulos (2010), for example, reveals that treated voters who weretold their names would be published in the local newspaper if they voted in an upcoming electionwere more likely to participate than those in the control group. In the current study, I investigatethe impact of the promise of being included on a “civic honor roll” of voters with a perfect votinghistory along with one’s neighbors. Subjects were also shown an example of such an honor rolllisting actual neighbors who had voted consistently in recent elections (treatment details providedbelow). Inclusion on the civic honor roll would signal—publicly—compliance with social votingnorms and generate prideful sentiments or, at the very least, avoid shameful sentiments that wouldpresumably arise from being excluded from the list. I expect this approach would exert positivesocial pressure to vote, and my hypothesis is that such an appeal enhances the likelihood ofelectoral participation.
Working together with a political consulting firm and a voter mobilization organization, wedesigned and implemented a randomized field experiment to test this hypothesis. Political scientistshave increasingly turned to field experiments to isolate the impact of various activities on voterturnout (Green & Gerber, 2008). Experimentation is a research method in which units of observationare assigned randomly to treatment and control groups. Field experiments, as distinct from labora-tory experiments, study the effects of an intervention within a naturalistic setting. In this case, theunits of observation are registered voters, the intervention is a postcard mailing, and the dependentvariable is validated voter turnout in the election of interest. The experimental protocols aredescribed in detail next.
Experimental Setting
I partnered with a nationally reputable political strategy and direct mail firm and Our Commu-nity Votes, a national 501(c)4 organization (both based in Washington, DC), to conduct a fieldexperiment in the context of the November 2009 general election in New Jersey. The election cyclefeatured a competitive statewide race for governor between incumbent Democrat, Jon Corzine, hisRepublican opponent, former United States Attorney for the District of New Jersey, Chris Christie,as well as Chris Daggett, an Independent candidate, and nine minor party contenders. Tax hikesand economic stagnation in New Jersey left Corzine vulnerable, and preelection polls throughout
267Positive Social Pressure and Prosocial Motivation
suggested a tight race. Christie eventually won the election following an intense campaign thatattracted 47% of registered voters on Election Day.1
Study Population
Against this backdrop, a sample of 41,301 registered New Jersey voters who had voted in theNovember 2008 general election, but not in the November 2005 election in New Jersey, wasrandomly assigned to receive a mailing designed to exert positive social pressure to vote or to acontrol group. The sample was stratified to be comprised of one-third African American, one-thirdHispanic, and one-third unmarried female subjects. This stratification scheme facilitates the exami-nation of differences across these select subgroups of voters, a key objective of the current study.That said, I note my experimental sample is comprised primarily, and by design, of lower-propensityvoters and excludes, for example, white males. The advantage to this approach is the ability toestimate effects for key demographic subgroups (minorities and unmarried women), but this comesat the expense of reliable comparisons to previous experimental studies conducted using a fullercross-section of voters.
Subjects assigned to the treatment group in my sample were sent a mailing within the week priorto the election, while those assigned to the control condition did not receive the mailing. A total of11,499 subjects were randomly assigned to receive the treatment, while 29,802 were assigned to thecontrol group that received no mailing.2
To ensure that random assignment generated treatment and control groups that were balanced interms of observable characteristics, I conducted a series of randomization checks.3 Table 1 displaysmean levels of available voter traits, including prior voting in the 2006 general elections, for subjectsin the treatment and control groups. Differences across experimental conditions are, as expected,
1 New Jersey Secretary of State, November 23, 2009, available at http://www.state.nj.us/state/elections/2009results/09general/2009-gen-election-ballots-cast-by-county-112309.pdf. I note that since my experimental sample was comprisedprimarily of lower-propensity voters (see below), overall turnout for my sample was undoubtedly lower, as suggested by theturnout rates for the control group I report below.
2 Resource limitations precluded the inclusion of a separate, reminder mailing in this study, but experimental evidence thatsuch reminders are ineffective is abundant (Green & Gerber, 2008).
3 Randomization was conducted at the individual level.
Table 1. Relationship between Treatment-Group Assignment andCovariates (Mean Levels) in the New Jersey Gubernatorial ElectionExperiment (November 2009)
Positive Control Prob > Fa
Voted November 2006 32.6 33.0 .44Male 27.4 26.7 .12Single 52.1 52.6 .36Democrat 47.4 47.0 .46Republican 8.9 8.6 .36Unaffiliated 43.6 44.3 .22Black 33.3 33.3 1.00Hispanic 33.3 33.3 1.00White 21.6 21.1 .25
Note. Figures in columns represent mean levels (percentages).aTest statistics generated using one-way ANOVA to evaluatewhether mean levels differ across categories of randomassignment. In all cases, I cannot reject the hypothesis of equalmeans at conventional significance (p < .05) levels, implyingbalance across groups.
268 Panagopoulos
insignificant at conventional (p < .05) levels. Successful randomization can be confirmed usingregression analysis to predict experimental assignment as a function of these covariates. The resultsof an F-test indicates no significant imbalances exist: F (7, 41293) = .82, p = .57.
Treatment
Subjects randomly selected to be treated were assigned to receive a mailing within the weekprior to the election. The positive social-pressure message was nonpartisan; it reinforced the generalimportance of voting, stressed the significance of the upcoming gubernatorial election, and provideda blandishment to vote. Crucially, the mailing also informed recipients that voting is a public act anddisplayed a list of 10 of each subject’s neighbors (along with their addresses) who had “perfectvoting records in general elections over the last four years.” Recipients’ names did not appear on thelist because their voting records were imperfect. The mailing presented the list as a “Civic HonorRoll” of “perfect voters,” and it encouraged subjects to join their neighbors as perfect voters byparticipating in the upcoming (November 2009) election. Messages were printed in black on plain,white paper, (81⁄2 × 11 inches) folded in half. An example of the full contents of the treatment appearsin Appendix A.
Experimental Results
Following the November 2009 election, I obtained validated voter-turnout data from the officialNew Jersey voter file. Table 2 reports the basic turnout rates for the group of subjects assigned toeach treatment condition. The control group in the experiment voted at a rate of 36.7%.4 Turnoutamongst voters assigned to receive the positive social pressure mailing voted at a rate of 38.7%,implying an intent-to-treat (ITT) effect of 2.0 percentage points (SE = 0.5) on average. I acknowl-edge that some subjects assigned to be treated may not have been successfully contacted, but reliableestimates of contact rates for direct mailings are unavailable. Thus, I report intent-to-treat effectsthroughout, noting these are likely conservative estimates of the treatment effects. Taking contactrates into account would only magnify the treatment effects I report. That said, any differences wouldlikely be trivial; previous research suggests failure-to-treat is not very problematic in studies of this
4 Readers are reminded that the experimental sample consisted of voters who had not voted in the 2005 election.
Table 2. Experimental Results (New Jersey, November 2009)
Experimental Group N Turnout (%) Intent-to-Treat
ALLTreatment 11,499 38.7 +2.0 (0.5)Control 29,802 36.7
BLACKTreatment 3,833 37.1 +2.3 (0.9)Control 9,934 34.8
HISPANICTreatment 3,833 36.3 +2.3 (0.9)Control 9,934 34.0
UNMARRIED WOMENTreatment 3,833 42.5 +1.3 (0.9)Control 9,934 41.2
Note. Standard errors in parentheses.
269Positive Social Pressure and Prosocial Motivation
sort as only a relatively small percentage of subjects typically fail to receive direct mailings (Gerber& Green, 2000; Panagopoulos, 2010).
As a robustness check, I employ multiple regression (OLS) to obtain estimates of the treatmenteffects. This approach permits the inclusion of control variables to correct for imbalances betweenexperimental groups due to chance. I estimate two models: Model 1 expresses individual voterturnout as a linear function of the experimental treatment, while Model 2 is embellished to includethe full battery of available covariates. The inclusion of covariates is optional, but it may reduce thedisturbance variance and improve the statistical precision of the estimated treatment effects.5
The results are presented in Table 3. Model 1 (column 1) confirms the treatment elevated turnoutby 2.0 percentage points on average, and the inclusion of covariates (Model 2, column 2) leaves theestimates essentially unchanged. Both sets of estimated treatment effects are statistically significantat the p < .01 level. I underscore the fact that the magnitude of these effects is about four timesstronger than what is typical for a nonpartisan mail intervention (Green & Gerber, 2008). Calculatinga 95% confidence interval around the estimated treatment effect of 1.9 percentage points generatedby Model 2 reveals the boost in turnout attributable to a single mailing that exerts positive socialpressure can range from 0.92 percentage points, which is already almost twice the impact of anaverage piece of mail, to as much as 2.88 percentage points, an effect that is roughly on par with theimpact of mailings designed to exert social pressure by showing voters their own voting records(Gerber et al., 2008), expressing gratitude for prior voting (Panagopoulos, 2011) and even otherpolicy interventions fashioned to stimulate turnout like Election Day registration and vote-by-mail(Knack, 2001). Even though the magnitude of the treatment effect I detect may not be as strong asstandard predictors of turnout developed over many years, such as partisanship (Rosenstone &Hansen, 1993), it does represent a noteworthy marginal improvement on the likelihood of voting,especially for an intervention that plays out in a day or two; effects of this magnitude have rarelybeen encountered in randomized field experiments designed to mobilize electoral participation usingdirect mail treatments, even when subjects have received as many as nine pieces of mail (Gerberet al., 2008).
A Closer Look at Heterogeneity in Treatment Effects
Recent studies suggest there may be heterogeneity in the impact of voter mobilization treatmentsattributable to voter attributes. Baseline voting propensity, for example, may condition the relation-ship between efforts designed to mobilize voters and the decision to vote (Arceneaux & Nickerson,2009). In a recent study, Panagopoulos (2010) shows that negative social pressure (shaming) affectsboth low- and high-propensity voters, while the effects of positive social pressure (like inducingpride) are restricted to high-propensity voters. It is also conceivable that demographic characteristics,like race or gender, may moderate the impact of the positive social pressure treatment. To evaluatethese possibilities, I conduct a series of analyses in which treatment effects are estimated separatelyfor key subgroups of voters in the experiment.
The results of these analyses, presented in columns 3–9 in Table 3, suggest scant evidence ofheterogeneous treatment effects. Lower- and higher-propensity voters6 in my sample were similarlymobilized by positive social pressure. At first blush, higher-propensity voters may appear to be
5 Since randomization was conducted at the individual level, it is unnecessary to cluster standard errors.6 The experiment was designed to target mainly groups of low-propensity voters overall, but even within these categories
individual subjects vote at varying rates. I take this into account to classify subjects as “lower-propensity” and “higher-propensity” voters using available past voting history. Additional past voting history was only available for the November2006 election. I classify subjects who had voted in that election as “higher-propensity” voters, while those who had failedto vote in 2006 were categorized as “lower-propensity” voters.
270 Panagopoulos
Tabl
e3.
Est
imat
esof
the
Eff
ects
ofth
ePo
sitiv
eSo
cial
Pres
sure
Mai
lT
reat
men
ton
Vot
erT
urno
utin
the
New
Jers
eyG
uber
nato
rial
Ele
ctio
n(N
ovem
ber
2009
)
Mod
elSp
ecifi
catio
nsV
ote
Prop
ensi
tyb
Dem
ogra
phic
Attr
ibut
es
(Mod
el1)
(Mod
el2)
Low
erH
ighe
rB
lack
His
pani
cU
nmar
ried
Fem
ale
Fem
ale
Mal
e
Bas
eVo
ting
Rat
e(T
urno
ut%
)28
.056
.035
.434
.641
.638
.136
.3
Tre
atm
ent
.020
**(.
005)
.019
**(.
005)
.018
**(.
006)
.026
**(.
009)
.023
**(.
009)
.023
**(.
009)
.013
*(.
009)
.018
**(.
006)
.024
**(.
010)
Nof
indi
vidu
als
41,3
0141
,301
27,7
0913
,592
13,7
6713
,767
13,7
6729
,238
10,7
53C
ovar
iate
saN
oY
esN
oN
oN
oN
oN
oN
oN
oR
MSE
.483
.457
.449
.496
.478
.476
.493
.485
.481
Not
e.E
stim
ates
repr
esen
tin
tent
-to-
trea
tef
fect
sde
rive
dfr
omO
LS
regr
essi
on.D
epen
dent
vari
able
isvo
ter
turn
out
inth
eN
ovem
ber
3,20
09,N
Jge
nera
lel
ectio
n.N
umbe
rsin
pare
nthe
ses
repr
esen
tst
anda
rder
rors
.**s
igni
fies
stat
istic
alsi
gnifi
canc
eat
the
p<
.01
leve
l,*a
tth
ep
<.1
0le
vel,
usin
gon
e-ta
iled
test
s.a C
ovar
iate
sin
clud
e:Pr
ior
turn
out
inth
e20
06ge
nera
lel
ectio
ns(N
ovem
ber)
asw
ell
asra
ce,g
ende
r,m
arita
lst
atus
,and
part
isan
affil
iatio
n.bO
nly
subj
ects
’vo
tehi
stor
yin
the
Nov
embe
r20
06el
ectio
nw
asav
aila
ble.
Acc
ordi
ngly
,Ico
nsid
ersu
bjec
tsw
hovo
ted
inth
isel
ectio
nas
“hig
her
prop
ensi
ty”
vote
rs,w
hile
thos
ew
hofa
iled
todo
soar
ede
sign
ated
“low
erpr
open
sity
”vo
ters
.
271Positive Social Pressure and Prosocial Motivation
slightly more responsive to the treatment compared to lower-propensity voters, at least in linearprobability terms, but the difference is statistically insignificant; moreover, the effects are basicallyidentical in probit or log-odds terms.
The findings tell a similar story across the range of demographic subgroups I investigate.Black and Hispanic voters’ responsiveness to the treatment was identical, boosting turnout by 2.3percentage points on average (p < .01, one-tailed) for both types of voters. The treatment elevatedturnout for unmarried women in my experimental sample by 1.3 percentage points on average(p < .10, one-tailed). The substantive effect is lower than for other voters subgroups I examine, butnot statistically different. Finally, non-White men appear to be somewhat more responsive topositive social pressure, compared to women overall; the estimates suggest the treatment raisedturnout by 2.4 percentage points on average for men (p < .01, one-tailed), and 1.8 percentagepoints for women (p < .01, one-tailed), but, once more, the difference is not statistically signifi-cant. I conclude from these results that voters across the range of demographic attributes Iexamine reacted similarly to the treatment, implying positive social pressure has the capacity tomobilize minorities and women effectively.
Discussion
Field experimental studies of voter mobilization routinely prime social norm considerations byincorporating appeals to civic duty to motivate participation, but the introduction of social surveil-lance is relatively new (Gerber et al., 2008). The field experiments I conducted extend this line ofresearch and reinforce the notion that social surveillance is an active ingredient in norm compliance.More generally, the current study bolsters the emerging consensus that activating powerful, under-lying psychological mechanisms can prompt citizens to political action and that the inclusion ofpsychological triggers likely exerts more potent effects than generic appeals. Accordingly, thefindings support claims advanced by numerous scholars about the powerful capacity of emotions toguide political behavior (Brader, 2005; Civettini & Redlawsk, 2009; Lodge & Taber, 2000, 2005;Marcus et al., 2000; Neuman et al., 2007; Panagopoulos, 2010, 2011; Redlawsk, 2002; Redlawsket al., 2007).
The experimental results I report provide empirical support for the hypothesis that interven-tions designed to publicize individuals’ voting behavior (or lack thereof) will effectively promoteparticipation. Social pressure, triggered by emotional reactions to stimuli, appears to motivateprosocial behavior. The effects detected in this experiment are not as potent as heavy-handed,negative social pressure treatments designed to shame voters into voting (Gerber et al., 2008), butthe findings suggest positive feedback, reinforcement, or recognition will promote socially respon-sible activity overall, presumably by activating feelings of pride and avoiding shame. The esti-mated impact of the positive social pressure/pride treatment I report in the current study ismodestly lower than for the “pride” treatment reported in Gerber et al. (2010), in which theauthors report a turnout boost of about 4 percentage points on average, but the authors’ treatmentinvolved a strong shaming component. The effects I observe above are strikingly consistent withsimilar treatments used in Panagopoulos (2010) and Panagopoulos (2011). Panagopoulos (2010)reports that when pride was induced by telling voters their names would be published in thenewspaper if they voted, turnout was 2 percentage points higher on average compared to thecontrol group that did not receive a blandishment to vote.7 A different set of field experiments
7 I note the alternative, “negative” social pressure message in Panagopoulos (2010) that induced shame by telling subjectstheir names would be publicized if they failed to vote was far more effective on average, raising turnout by about 7percentage points, compared to the control group, an effect that is on par with the 8 percentage point effect Gerber et al.(2008) report for their shame-inducing “neighbors” social pressure treatment that exposed subjects to their own as well astheir neighbors’ vote histories.
272 Panagopoulos
conducted using a more nuanced message that arguably incorporated positive social pressure byexpressing gratitude to voters for having voted in a prior election raised turnout by about 2 to 3percentage points on average compared to control conditions that were untreated (Panagopoulos,2011). Even as I note these intriguing similarities, I acknowledge that direct comparisons tostudies conducted on full cross-sections of voters may be inappropriate given that the experimentalsample in the current study was comprised primarily of lower-propensity voters in the first place.Still, I present this discussion by way of speculation.
The subgroup analyses I conduct, a key contribution of the current study, reveal some intriguingpatterns that both support extant findings and raise new questions that beg further examination. Incontrast with previous research that found positive social pressure treatments designed to invokepride motivated only high-propensity voters (Panagopoulos, 2010), for example, the current studyfinds no such heterogeneity in treatment effects; positive social pressure seems to motivate bothlower- and higher-propensity voters in my experimental sample similarly. I also show that positivesocial pressure effectively motivates voting among minorities and unmarried women in my sample,a finding that may speak to the generalizability of the positive social pressure effects established inprevious studies (Gerber et al., 2010; Panagopoulos, 2010, 2011) to these key subgroups of voters,at least in certain electoral contexts.
From a practical perspective, the findings I report above have significant implications forthe electoral process generally in which mobilizing voters, especially disadvantaged, low voting-propensity voters, is persistently challenging. Even a modest bump in turnout among these groupsof voters attributable to a single intervention along the lines I describe in this study is substan-tively meaningful.
I acknowledge that a potential limitation of the current study is the use of one election (the2009 gubernatorial race) in one state (New Jersey). Despite evidence that positive social pressuretreatments likely operate similarly on average across a wide range of electoral contexts (Pana-gopoulos, 2011), whether positive social pressure affects key subgroups of voters consistently (asthe current study suggests) in other states or in midterm or presidential election cycles remainopen questions. These findings are theoretically compelling and open up new possibilities forsubsequent inquiry and experimental exploration. Replication is necessary to sharpen and confirmthese initial conclusions.
ACKNOWLEDGMENTS
The author is grateful to Donald Green, the editor, and anonymous reviewers for invaluablecomments and suggestions. I also thank Hal Malchow, Joel Rivlin, and Bill Russell. Correspondenceconcerning this article should be sent to Costas Panagopoulos, Department of Political Science,Fordham University, 441 E. Fordham Rd., Faber 667, Bronx, NY 10458. E-mail: [email protected]
REFERENCES
Arceneaux, K., & Nickerson, D. W. (2009). Who is mobilized to vote? A re-analysis of 11 field experiments. American Journalof Political Science, 53(1), 1–16.
Barrett, K. C., & Campos, J. J. (1987). Perspectives on emotional development II: A functionalist approach to emotions. InJ. D. Osofsky (Ed.), Handbook of infant development (2nd ed., pp. 555–578). Oxford, UK: Wiley.
Bear, G., Manning, M., & Izard, C. (2003). Responsible behavior: The importance of social cognition and emotion. SchoolPsychology Quarterly, 18(2), 140–157.
Brader, T. (2005). Campaigning for hearts and minds: How emotional appeals in political ads work. Chicago: University ofChicago Press.
273Positive Social Pressure and Prosocial Motivation
Civettini, A. J. W., & Redlawsk, D. P. (2009). Voters, emotions, and memory. Political Psychology, 30(1), 125–151.
Cosmides, L., & Tooby, J. (2000). Evolutionary psychology and the emotions. In M. Lewis & J. M. Haviland-Jones (Eds.),Handbook of emotions, 2nd ed. (pp. 91–115). New York: Guilford Press.
Deci, E., & Ryan, R. (1985). Intrinsic motivation and self-determination in human behavior. New York:Plenum.
Deci, E., Koestner, R., & Ryan, R. (2001). Extrinsic rewards and intrinsic motivation in education: Reconsidered once again.Review of Educational Research, 71, 1–27.
Downs, A. (1957). An economic theory of democracy. New York: Harper and Row.
Gaines, L. M., Duvall, J., Webster, J. M., & Smith, R. H. (2005). Feeling good after praise for a successful performance: Theimportance of social comparison information. Self and Identity, 4, 373–389.
Garcia, F. C., & Sanchez, G. R. (2008). Hispanics and the U.S. political system: Moving into the mainstream. Upper SaddleRiver, NJ: Pearson Prentice Hall.
Gerber, A. S., & Green, D. P. (2000). The effects of canvassing, direct mail, and telephone contact on voter turnout: A fieldexperiment. American Political Science Review, 94, 653–663.
Gerber, A. S., Green, D. P., & Larimer, C. W. (2008). Social pressure and voter turnout: Evidence from a large scale fieldexperiment. American Political Science Review, 102(February), 33–48.
Gerber, A. S., Green, D. P., & Larimer, C. W. (2010). An experiment testing the relative effectiveness of encouraging voterparticipation by inducing feelings of pride or shame. Political Behavior, 32(3): 409–422.
Gollwitzer, P. M., & Moskowitz, G. B. (1996). Goal effects on action and cognition. In E. T. Higgins & A. W. Kruglanski(Eds.), Social psychology: Handbook of basic principles (pp. 361–399). New York: Guilford Press.
Green, D. P., & Gerber, A. S. (2008). Get out the vote! How to increase voter turnout (2nd ed.). Washington, DC: BrookingsInstitution Press.
Knack, S. (2001). Election-day registration: The second wave. American Politics Research, 29(January), 65–78.
LeDoux, J. E. (1996). The emotional brain: The mysterious underpinnings of emotional life. New York: Simon & Schuster.
Leighley, J. E. (2001). Strength in numbers? The political mobilization of racial and ethnic minorities. Princeton, NJ:Princeton University Press.
Lodge, M., & Taber, C. S. (2005). The automaticity of affect for political leaders, groups, and issues: An experimental test ofthe hot cognition hypothesis. Political Psychology, 26(3), 455–482.
Lodge, M., & Taber, C. S. (2000). Three steps toward a theory of motivated political reasoning. In A. Lupia, M. McCubbins,& S. Popkin (Eds.), Elements of reason: Cognition, choice, and the bounds of rationality (pp. 183–213). Cambridge:Cambridge University Press.
Mankani, N., & Roeder, E. (2010). 2010 turnout: Quantifying the quandry. Report issued by the New OrganizingInstitute. Accessed online on July 22, 2011 at http://neworganizing.com/wp-content/uploads/2010/09/NOI-2010-Turnout.pdf.
Marcus, G., Neuman, W. R., & MacKuen, M. (2000). Affective intelligence and political judgment. Chicago: The Universityof Chicago Press.
Mascolo, M. F., & Fischer, K. W. (1995). Developmental transformations in appraisals for pride, shame and guilt. In J. P.Tangney & K. W. Fischer (Eds.), Self-conscious emotions: The psychology of shame, guilt, embarrassment and pride(pp. 64–113). New York: Guilford Press.
Neuman, W. R., Marcus, G., Crigler, A., & MacKuen, M. (Eds.) (2007). The affect effect: Dynamics of emotion in politicalthinking and behavior. Chicago: University of Chicago Press.
Panagopoulos, C. (2010). Affect, social pressure and prosocial motivation: Field experimental evidence of the mobilizingeffects of pride, shame and publicizing voting behavior. Political Behavior, 32(3), 369–386.
Panagopoulos, C. (2011). Thank you for voting: Gratitude expression and voter mobilization. Journal of Politics,73(3), 707–717.
Redlawsk, D. P. (2002). Hot cognition or cool consideration? Testing the effects of motivated reasoning on political decisionmaking. Journal of Politics, 64(4), 1021–1044.
Redlawsk, D. P., Civettini, A. J. W., & Lau, R. (2007). Affective intelligence and voting: Information processing and learningin a campaign. In A. Crigler, M. MacKuen, G. E. Marcus, & W. R. Neuman (Eds.), The affect effect: Dynamics ofemotion in political thinking and behavior. Chicago: University of Chicago Press.
Rosenstone, S., & Hansen, J. (1993). Mobilization, participation and democracy in America. New York:Macmillan.
Webster, J. M., Duvall, J., Gaines, L. M., & Smith, R. H. (2003). The roles of praise and social comparison information in theexperience of pride. Journal of Social Psychology, 143, 209–232.
274 Panagopoulos
Williams, L., & DeSteno, D. (2008). Pride and perseverance: The motivational role of pride. Journal of Personality and SocialPsychology, 94(6), 1007–1017.
Women’s Voices Women Vote. (2010). Unmarried women in the electorate: Behind the numbers. Report issued by Women’sVoices Women Vote. Accessed online on July 22, 2011 at: http://wvwv.org/assets/2010/1/19/Unmarried_Women_in_the_Electorate_-_Behind_the_Numbers.pdf.
Appendix A: Treatment
Dear [FIRSTNAME, LASTNAME]:
There is no action more important to our democracy than going to the polls to vote. That’s why OurCommunity Votes, a non-profit organization that encourages voting, is recognizing citizens in yourneighborhood who have perfect voting records in the general elections over the past four years.
These neighbors deserve our recognition and congratulations for doing their civic duty and makingtheir voices heard. And with New Jersey’s election for governor taking place on November 3rd, wehope you will go to the polls and join your neighborhood’s Civic Honor Roll of perfect voters.
Voting records show that you voted in the presidential election of 2008 but not in the 2005 electionfor governor. Voting records are public information, so people know when you voted, but never howyou voted.
By voting on November 3rd, you will join the following voters as perfect voters.
DO YOUR CIVIC DUTY—VOTE!
Nov 05 Nov 08
[ADDRESS] [FIRSTNAME, LAST NAME] VOTED VOTED[ADDRESS] [FIRSTNAME, LAST NAME] VOTED VOTED[ADDRESS] [FIRSTNAME, LAST NAME] VOTED VOTED[ADDRESS] [FIRSTNAME, LAST NAME] VOTED VOTED[ADDRESS] [FIRSTNAME, LAST NAME] VOTED VOTED[ADDRESS] [FIRSTNAME, LAST NAME] VOTED VOTED[ADDRESS] [FIRSTNAME, LAST NAME] VOTED VOTED[ADDRESS] [FIRSTNAME, LAST NAME] VOTED VOTED[ADDRESS] [FIRSTNAME, LAST NAME] VOTED VOTED[ADDRESS] [FIRSTNAME, LAST NAME] VOTED VOTED
The November 3rd governor’s election is very important to New Jersey and Red Bank. So we hopeyou will go to the polls on November 3rd and join your neighborhood’s Civic Honor Roll of perfectvoters.
Daniel MillerOur Community Votes
P.S. Our Community Votes is a not-for-profit organization that encourages citizens to go to the pollsand vote. If you have questions about our work, you can contact us at [email protected].
275Positive Social Pressure and Prosocial Motivation