CrossRef citations to date

A second look at partisanship’s effect on receptivity to social pressure to vote

& ORCID Icon
Pages 1-13
Received 12 Mar 2018
Accepted 17 Jan 2019
Published online: 26 Jan 2019


Social pressure can exert a powerful, but sometimes counterproductive, influence on compliance with the social norm of voting. Scholars have tested several implicit social pressure techniques to reduce negative reactions to these methods. Among the most innovative is the use of ‘watching eyes’ in voter mobilization messages. Using three large randomized field experiments, this study attempts to reproduce Panagopoulos and van der Linden’s finding that political partisanship moderates the effect of watching eyes messages on voter turnout. Our findings diverge from previous findings statistically and substantively and indicate partisanship may have limited influence on the effectiveness of watching eyes in mobilizing voters.

The foundation of democracy is voting, and scholars have expended substantial resources studying how to increase voter turnout. Researchers have examined mobilization techniques such as door-to-door canvassing, phone calling, sign posting, and direct and electronic mailing using different mobilization messages such as self-directed implementation intentions, appeals to group loyalty, and social norm compliance (see Green & Gerber, 2015, for an overview). This research suggests one of the most effective means to mobilize voters is norm compliance messaging in the form of social pressure. Broadly speaking, social pressure messages are designed to gain social compliance by ‘play[ing] upon a basic human drive to win praise and avoid chastisement’ (Green & Gerber, 2010, p. 331).

In general, this takes the form of messages urging the recipient to comply with a social norm, informing the recipient compliance with the norm is easily verified, and warning the recipient that compliance will be tracked (Green & Gerber, 2015). In voter mobilization terms, this generally means telling the citizen voting is a civic duty and informing her or him that individuals’ voting histories are a public record that will be monitored. More extreme forms of social pressure threaten to share with neighbors whether the recipient did or did not vote (Gerber, Green, & Larimer, 2008). Experiments testing social pressure mobilization show these messages generate substantial effects, with maximum pressure messages generating effects rivaling those of the gold standard of mobilization techniques, door-to-door, in-person canvassing (Gerber et al., 2008).

While effective at increasing voter turnout, social pressure techniques can anger recipients concerned about privacy and create a negative and counterproductive backlash against candidates or organizations that use them (Mann, 2010; Matland & Murray, 2013). As a result, researchers have investigated and developed a number of less objectionable social pressure messages that also increase turnout, although to a lesser degree. These messages include recognizing voters who reliably turnout to vote (Panagopoulos, 2013), thanking citizens who voted in a previous election (Panagopoulos, 2011), and applying public pressure to vote directly or indirectly via social media messaging (Bond et al., 2012; Haenschen, 2016; Teresi & Michelson, 2015).

Among the most innovative of these implicit social pressure techniques is exerting social pressure using ‘watching eyes’ in mobilization messages. Researchers hypothesize eye displays create a feeling of being watched, which stimulates a subconscious desire to act prosocially in compliance with social norms (e.g., Kawamura & Kusumi, 2017), and scholars have presented substantial evidence that watching eyes do indeed promote prosociality (e.g., Bateson, Nettle, & Roberts, 2006; Ernest-Jones, Nettle, & Bateson, 2011; Pfattheicher, Strauch, Diefenbacher, & Schnuerch, 2018; Powell, Roberts, & Nettle, 2012). Panagopoulos first tested and found effects for watching eyes as a voter mobilization technique in two randomized field experiments in the United States in Key West, Florida (2014a), and Lexington, Kentucky (2014b).

As Panagopoulos and van der Linden state, though, ‘the “watchful eyes” effect has become a source of active debate, as several recent studies have failed to replicate the effect’ (2016, p. 177). Since their article, more studies have failed to reproduce the effect. For instance, Manesi, Van Lange, Van Doesum, and Pollet (2018) found that watching eyes do not increase the probability of donating to victims of a natural disaster or the amount donated to victims. Pfattheicher, Schindler, and Nockur (2018) showed they do not reduce cheating on dice-rolling or coin flipping tasks. Koornneef et al. (2018) presented evidence from a field experiment they do not improve the hand hygiene of medical students treating simulated patients. Mansi and Pollet (2017) found in three field experiments they do not increase the probability that individuals will help strangers by dropping a ‘lost letter’ into a mailbox. Perhaps emblematic of the watching eyes literature, Kawamura and Kusumi (2017) failed to replicate their own lab experiment showing watching eyes promote charitable donations.

Among the failed efforts to reproduce the effect that Panagopoulos and van der Linden cite, Northover, Pedersen, Cohen, and Andrews (2017) found in a pair of meta-analyses that watching eyes do not increase generosity or the probability that individuals will show generosity. More pertinently, Matland and Murray (2016) showed in a five-site replication statistically and substantively weak effects that were inconsistent with Panagopoulos’ (2014a, b) results. In nine of 10 cases, watching eyes messages failed to generate a significant effect, raising questions about their effectiveness for voter mobilization.

In a follow-up study designed to assess moderating factors that may explain the variable effects of watching eyes, Panagopoulos and van der Linden (2016) re-analyzed the data from the Key West and Lexington field experiments. They plausibly argue political identity in the form of party identification should moderate the effect of watching eyes on voter mobilization. Research on political ideology over the past 10 to 15 years has shown liberals and conservatives not only vary systematically across a range of policy preferences but also personal preferences and evaluations of the morality of specific actions. Conservatives prioritize order and social conformity, have an aversion to ambiguity, and strongly prefer closure (Hibbing, Smith, & Alford, 2013; Jost, Glaser, Kruglanski, & Sulloway, 2003). They also score higher than liberals on conscientiousness, one of the big five personality traits (Carney, Jost, Gosling, & Potter, 2008). In developing Moral Foundations Theory, Haidt and Joseph (2004) argue liberals prioritize the care/harm and fairness/reciprocity foundations. Conservatives, on the other hand, put greater weight on the foundations of in-group/loyalty, authority/respect, and purity/sanctity (Graham et al., 2011; Haidt & Graham, 2007). Especially the third and fourth moral foundations emphasize loyalty, duty, obedience, and respect for traditions. Based on these differences, Panagopoulos and van der Linden plausibly predict conservatives will react more strongly to the pressure to conform to social norms regarding voting than liberals. They found Republican voters, who typically hold a conservative belief system (Lewis-Beck, Jacoby, Norpoth, & Weisberg, 2008, Chapter 9), consequentially responded to watching eyes messages regarding voting, while Democratic voters, who typically hold a liberal belief system, did not. reports the results from Panagopoulos and van der Linden (2016) and shows watching eyes increased Republican turnout by 3.0 percentage points relative to the control group in Key West and 4.4 percentage points in Lexington. Watching eyes did not significantly increase turnout relative to the control group in either location among Democrats, individuals who identified with other parties, or voters who were unenrolled (neither Democratic nor Republican).

Table 1. Turnout effects for watching eyes and placebo mailers in key West, FL and Lexington, KY.

It is important to understand this effect in application and in theory. Political campaigns have frequently applied voter mobilization techniques, including social pressure techniques, presented in the academic literature. Moreover, scholars have recently expressed concern about reproducibility (e.g., Open Science Collaboration, 2015), a foundation of the scientific endeavor. McDermott captures this noting, ‘If a genuine cause-and-effect relationship exists across variables, then it should emerge over time, within different contexts, using various methods of measurement, and across population groups’ (2011, p. 34). In this spirit, the objective of our study is to attempt to reproduce Panagopoulos and van der Linden’s (2016) results that Republicans respond more strongly to watching eyes voter mobilization messages than Democrats. Re-analyses of data from our (Matland & Murray, 2016) randomized field experiments show, contrary to Panagopoulos and van der Linden’s (2016) findings, partisanship fails to moderate the effects of watching eyes on voter turnout.

Data and methods

This study includes re-analyses of data from the three sites described in Matland and Murray (2016) for which party affiliation data are available (El Paso, Texas; Midland, Texas; and Toledo, Ohio). The unit of analysis is the individual registered voter, and the outcome variable is validated individual voter turnout as reported by election officials. The El Paso election took place in June 2013, while the Midland and Toledo elections were held in November 2013. Each was a local election that included a mayoral race. We used three different mailers at each site: one with female watching eyes, one with male watching eyes (both presented in Figure 1), and a placebo where a US flag replaces the eyes but with the identical civic duty text. The eye images were computer generated using FaceGen Modeller, which is 3D face-modeling software. The intent was to standardize the faces while manipulating their femininity and masculinity.

Figure 1. GOTV Mailers: El Paso, TX, Midland, TX, and Toledo, OH.

Election officials provided current lists of their jurisdiction’s registered voters approximately 30 days before their election. At each site, we removed addresses with more than four registered voters on the assumption these were apartment buildings with substantial turnover and, therefore, with a high probability of incorrect voter addresses. In the remaining households, we randomly selected one voter per household to ensure no household was included in more than one experimental group. This resulted in subject pools of 109,231 in El Paso, 18,170 in Midland, and 68,635 in Toledo. Finally, we blocked subjects based on voting propensity and randomly assigned each to one of the three treatment groups or the control group. Randomization checks indicated the groups did not differ statistically. We assigned just over 3,000 per treatment group in El Paso and 2,000 per treatment group in Midland and Toledo. Panagopoulos assigned 1,500 per treatment group in Key West (2014a) and 1,000 per group in Lexington (2014b). The mailers were sent by first-class, presorted mail six days before the elections. Election officials provided validated voter turnout information after their election; that is, each subject’s voter turnout was verified by government records and not based on self-reports.1

We determined each subject’s partisanship by her or his validated history of voting in party primaries. Subjects who had voted only in Democratic primaries were categorized as Democrats, subjects who had voted only in Republican primaries were categorized as Republicans, and subjects not voting in any party primaries were categorized as Independents. The few citizens who had voted in both Democratic and Republican primaries were dropped. Like Panagopoulos and van der Linden (2016), we report results for the combined female and male watching eyes treatments because there are no noticeable differences between them in effect.


Results using previous tests

, columns 2 to 5, present our results using tests used by Panagopoulos and van der Linden. The second column shows that when combined, Democrats’, Republicans’, and Independents’ reactions to the watching eyes treatment are close to statistically distinguishable from the control group in Midland. Ultimately, however, the eyes treatment does not have a statistically significant effect for all voters at any of our three sites. The placebo mailer had a significant negative effect among all voters and for each group of voters in Toledo but showed no effect in El Paso or Midland. The third column shows the watching eyes treatment is close to statistically significant among Democrats in Midland but is clearly insignificant in El Paso and Toledo.

Table 2. Turnout effects for watching eyes and placebo mailers in El Paso, TX, Midland, TX, and Toledo, OH.

More pertinently, the fourth column shows the Republican results. Panagopoulos and van der Linden (2016) predicted Republicans would be significantly more likely to be mobilized and sensitive to social pressure to vote than Democrats. However, none of the three separate tests yields a statistically significant effect for Republicans compared to the control group. Moreover, shows the substantive effect for Republicans is negative at two of the three sites, while the effect for Democrats is positive at all three sites. Overall, contrary to Panagopoulos and van der Linden’s findings (2016), in our field experiments there are no statistically significant effects of watching eyes compared to the control groups for Democrats, Independents, or especially Republicans.

Combining the results across our three sites confirms these findings. A meta-analysis of the Democrats results in an average treatment effect (ATE) of 1.24 percentage points (SE = 0.92, p = 0.21 two-tailed) versus the controls, while meta-analyses for Republicans and Independent voters result in insignificant ATE effects of 0.01 percentage points (SE = 1.18, p = 0.99 two-tailed) and −0.18 percentage points (SE = 0.31, p = 0.86 two-tailed).

A meta-analysis combining data from all five sites finds some support for the initial hypothesis. There is virtually no impact among Independents (ATE = −0.11, SE = 0.28) and a modest effect that is statistically insignificant among Democratic voters (ATE = .97, SE = .61). Republicans, though, show a residual effect (ATE = 1.82, SE = 0.79, p < 0.03 two-tailed). This suggests adding our three new tests weakens the conclusion that Republicans are more sensitive to watching eyes, but the effect is not so powerful that it conclusively undermines the previous result.

Results using alternative tests

While the failure in our tests to find meaningful support for the hypothesis raises questions, we also are concerned whether the right test was conducted. Panagopoulos and van der Linden hypothesize there will be ‘a stronger effect of eyespot images on voter mobilization for Republican/conservative voters compared to Democratic/liberal voters’ (2016, p. 178). While they find the treatment is statistically significant compared to the control group in models of Republicans but not for models of Democrats, they do not directly test, as the hypothesis suggests, whether the difference between the two treatment effects is statistically significant. We believe the correct test for comparing the impact of watching eyes on Republicans and Democrats is testing whether the difference between the Republican and Democratic coefficients is statistically significant, not whether one is significantly different from the control group and the other is not.

The last column of presents Chow tests of the differences between the regression coefficients for Democrats and Republicans. These are insignificant (p-values are 0.16, 0.44, and 0.92). The last column of evaluates the difference between the Democratic and Republican coefficients in Key West and Lexington. These are also insignificant indicating the difference in effect between Democrats and Republicans fails statistical significance in the original study. In Key West the difference is 0.02 with a standard error of 0.02, which fails a p = 0.10 one-tailed test. The results are weaker in Lexington where the impact is 0.03 with a standard error of 0.03 and a t-statistic barely above 1.00. We believe this more direct test clearly fails to support the hypothesis.

Another test, different data

Panagopoulos and van der Linden posited an emphasis on conscientiousness and loyalty would make Republicans more sensitive than Democrats to the implicit social pressure generated by watching eyes. Our results fail to bear this out. The broader theoretical assertion, though, is that party identification moderates the effect of social pressure. To evaluate this assertion further, it is reasonable to look at other field experiments that use different types of social pressure to increase voter turnout. If the argument is accurate, then other social pressure treatments should more effectively mobilize Republicans to vote than Democrats.

reports the effect of a ‘self-mailer’ used in a field experiment designed to increase turnout for the general election held in Lubbock, Texas, in November 2010 (Murray & Matland, 2014). We tested this social pressure mailer and a civic duty mailer. The self-mailer includes for each recipient a listing of her or his voting record over the past five elections along with a message noting voting turnout is a matter of public record. This form of social pressure has consistently been found to raise turnout. Our results show both the self-mailer and the civic duty mailer raised turnout. Most important are the results for differences across the three party groups. Panagopoulos and van der Linden would predict the social pressure mailer would have a greater effect on Republican than Democratic voters. The results show this does not happen. Neither mailer has a statistically significant effect among Republicans, and the weakest substantive effects are among Republicans. Both mailers have significant positive effects among Independents, while Democrats are not significantly moved by the social pressure mailer but are by the civic duty mailer. A Chow test of the Democratic and Republican regression coefficients shows the difference between the parties is not statistically significant. The results from this independent test taking advantage of the downstream results of an earlier experiment again fails to support the broader proposition that Republicans are more sensitive to social pressure messaging than Democrats.

Table 3. Turnout effects for self-mailer and placebo mailers in Lubbock, TX.


Panagopoulos and van der Linden (2016) find their two randomized field experiments show watching eyes significantly mobilized Republicans but not Democrats or other voters. Contrary to those findings, the three randomized field experiment results we report indicate watching eyes never mobilized Republicans while getting close to mobilizing Democrats at one of three sites. Most importantly, we find no case in which the difference between the effects on Democrats and Republicans is statistically significant. It is reasonable to ask how to explain these contradictory outcomes.

Context may matter. First, in our experiments, control group turnout was 16% in El Paso, 26% in Midland, and 37% in Toledo. Panagopoulos and van der Linden’s reported control group turnout was 4% in Key West and 23% in Lexington. There is a wide range of turnout across the five sites, and they follow no consistent pattern. A theoretical argument based on turnout that would fit these findings must be highly complex and, therefore, be highly unlikely. Next, four of the five elections are mayoral elections and perhaps mayoral elections are unique. Holding office constant, however, does not lead to constant outcomes. In Key West watching eyes mobilized Republicans but not Democrats in the mayoral race, while in Midland’s mayoral election Democrats were closer to being mobilized. In addition, Panagopoulos and van der Linden find similar results at their two sites even though the office varies. Similarly, the partisan composition of the electorate may matter, but the results suggest otherwise. Key West and Lexington are heavily Democratic, so Republicans there may receive little campaign contact resulting in Republicans being especially sensitive to mobilization messages. But El Paso and Toledo are also heavily Democratic, and Republicans there did not respond to the treatments. Finally, political culture, as defined by Elazar (1972), is generally constant but with differences in outcomes. Midland, Toledo, El Paso, and Key West are individualistic or partially individualistic cultures, but the results differ. Overall, we have difficulty coming up with a plausible theoretical reason to believe context should lead to differentiated treatment effects across partisans.

We believe we are on stronger ground when we suggest methodological decisions underlie the differences. Perhaps the eyes used stimulate different responses. As shown in Figure 2, Panagopoulos and van der Linden used stylized female eyes in Key West and somber male eyes in Lexington. Our experiments used computer-generated eyes, both female and male, with similar neutral expressions. Although there are clear differences, there is no theoretical or empirical evidence suggesting the type of eyes should matter or that Democrats and Republicans should be sensitive to different sets of eyes. Furthermore, the watching eyes literature has generated significant positive effects with a variety of eyes, using human eyes in many cases but often using much more abstract images such as eye stickers (Powell et al., 2012).

Figure 2. GOTV Mailers: Key West, FL, and Lexington, KY.

Another methodological distinction between the two studies is how the samples are drawn. Panagopoulos drew his samples from single-voter households, while we drew our samples from households of one to four voters. Perhaps individuals without voting cohabitators are more susceptible to watching eyes, but it seems unlikely that should be truer of Republicans than Democrats. To test this possibility, we reran our regressions and limited the samples to single-voter households. There were some changes in effect sizes but no clear pattern to those changes. Watching eyes were insignificant for all three single-voter household samples of Republicans and the three samples of Independents. They were significant in one of the three Democratic samples, strongly suggesting sampling protocol cannot explain the difference.

The method used to establish partisanship also differs. Panagopoulos and van der Linden drew samples in states with closed primary systems, which indicates that a voter’s partisanship is known because a voter must have an official affiliation with a party to vote in its primary elections. We drew samples in states with open primary systems, which means that a voter’s partisanship is uncertain because a voter does not need an official affiliation with a party to vote in its primary elections. This makes strategic voting – e.g., voting for a weaker opposition party candidate in the primary election to get a more desirable (weaker) opposing candidate for the preferred party candidate in the general election – more probable. Strategic voting, in turn, makes partisan misidentification more probable as voters can easily switch which party primary they vote in. While research has found that strategic voting occurs in the United States (Abramson, Aldrich, Paolino, & Rohde, 1992), we believe it was unlikely to affect significantly our results. To be coded as a partisan in our study a voter must have voted only in one party’s primaries and done it at least twice.

Finally, such results may be due to a lack of statistical power or sampling variability. It is reasonable for a study reporting null results to raise questions about underpowered statistical tests. This is not a concern here. shows the treatment effect is more positive on Democrats than Republicans at two of the three sites, making statistical power irrelevant, and at the one site where the effect is more positive for Republicans (El Paso), the coefficient is one-tenth of one percentage point greater, making it substantively and statistically insignificant. Regarding sampling variability, although the samples were deemed equivalent from the randomization checks, the groups can still be unbalanced in terms of factors we do not measure that impact turnout.


Panagopoulos and van der Linden (2016) developed a plausible hypothesis based on partisan differences in the effectiveness of social pressure. Their study found this hypothesis was a promising avenue for closer analysis. We attempted to reproduce their results using new tests and data. The results, however, do not support the hypothesis that conservatives will be more sensitive to watching eyes than their liberal counterparts. At no point do Republican voters react more positively and strongly to the watching eyes across our three sites. Reanalysis of the results from the initial two sites, directly testing whether the Republican and Democratic effects are significantly different, fails to find the hypothesized effect at either site. Finally, our previous social pressure experiment in Lubbock County in 2010 failed to find the broader effect.

More generally, in response to increasing concerns about type I error (Open Science Collaboration, 2015), there has been an ongoing methodological debate as to whether researchers should tighten standards of statistical significance. While we are undecided as to which standard is preferable and practicable, it is striking that Benjamin et al. note (Benjamin et al., 2018, p. 1): ‘For fields where the threshold for defining statistical significance for new discoveries is p < 0.05, we propose a change to p < 0.005. This simple step would immediately improve the reproducibility of scientific research in many fields.’ The research on watching eyes and voter mobilization supports this proposition. For the studies presented here, most of the significant results disappear if we use the more rigorous standard as the number of significant results in and , columns 2 to 5, drops from eight to two. A stricter statistical standard leaves us with a significant effect at only one of the five sites for Republicans and at none of the sites when assessing the difference between Democrats and Republicans. Under this aspirational standard, there is only very modest support for the hypothesis.

Overall, the evidence here indicates the effects of watching eyes on voter turnout are limited and do not dramatically differ due to partisanship. This suggests to us that political campaigns should use watching eyes with great care, but given the broader research on watching eyes, scholars should pursue them with great interest.


We would like to thank the electoral officials who provided us with data and Don Green for his support. This experiment was preregistered with Experiments in Government and Politics at http://e-gap.org/design-registration.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information


This work was supported by Donald P. Green.


1. See Matland and Murray (2016) for full details of the original study.


  • Abramson, P. R., Aldrich, J. H., Paolino, P., & Rohde, D. W. (1992). “Sophisticated” voting in the 1988 presidential primaries. American Political Science Review, 86(1), 5569. [Crossref], [Web of Science ®][Google Scholar]
  • Bateson, M., Nettle, D., & Robertson, G. (2006). Cues of being watched enhance cooperation in a real-world setting. Biological Letters, 2(3), 412414. [Crossref], [PubMed], [Web of Science ®][Google Scholar]
  • Benjamin, D., Berger, J., Johannesson, M., Nosek, B., Wagenmakers, E. J., Berk, R., … Johnson, V. E. (2018). Redefine statistical significance. Nature Human Behavior Comment, 2, 610. [Crossref], [PubMed], [Web of Science ®][Google Scholar]
  • Bond, R. M., Fariss, C. J., Jones, J. J., Kramer, A. D., Marlow, C., Settle, J. E., & Fowler, J. H. (2012). A 61-million-person experiment in social influence and political mobilization. Nature, 489(7415), 295. [Crossref], [PubMed], [Web of Science ®][Google Scholar]
  • Carney, D. R., Jost, J. T., Gosling, S. D., & Potter, J. (2008). The secret lives of liberals and conservatives: Personality profiles, interaction styles, and the things they leave behind. Political Psychology, 29, 807840. [Crossref], [Web of Science ®][Google Scholar]
  • Elazar, D. J. (1972). American Federalism: A view from the states. New York, NY: Harper & Row. [Google Scholar]
  • Ernest-Jones, M., Nettle, D., & Roberts, M. (2011). Effects of eye images on everyday cooperative behavior: A field experiment. Evolution and Human Behavior, 32(3), 172178. [Crossref], [Web of Science ®][Google Scholar]
  • Gerber, A. S., Green, D. P., & Larimer, C. W. (2008). Social pressure and voter turnout: Evidence from a large-scale field experiment. American Political Science Review, 102(1), 3348. [Crossref], [Web of Science ®][Google Scholar]
  • Graham, J., Nosek, B., Haidt, J., Iyer, R., Koleva, S., & Ditto, P. H. (2011). Mapping the moral domain. Journal of Personality and Social Psychology, 101(2), 366385. [Crossref], [PubMed], [Web of Science ®][Google Scholar]
  • Green, D. P., & Gerber, A. S. (2010). Introduction to social pressure and voting: New experimental evidence. Political Behavior, 32(3), 331336. [Crossref], [Web of Science ®][Google Scholar]
  • Green, D. P., & Gerber, A. S. (2015). Get out the vote: How to increase voter turnout (3rd ed.). Washington, DC: Brookings Institution Press. [Google Scholar]
  • Haenschen, K. (2016). Social pressure on social media: Using facebook status updates to increase voter turnout. Journal of Communication, 66, 542563. [Crossref], [Web of Science ®][Google Scholar]
  • Haidt, J., & Graham, J. (2007). When morality opposes justice: Conservatives have moral intuitions that liberals may not recognize. Social Justice Research, 20(1), 98116. [Crossref][Google Scholar]
  • Haidt, J., & Joseph, C. (2004). Intuitive ethics: How innately prepared intuitions generate culturally variable virtues. Daedelus, 133(4), 5566. [Crossref], [Web of Science ®][Google Scholar]
  • Hibbing, J. R., Smith, K. B., & Alford, J. R. (2013). Predisposed: Liberals, conservatives, and the biology of political differences. New York, NY: Routledge. [Crossref][Google Scholar]
  • Jost, J. T., Glaser, J., Kruglanski, A. W., & Sulloway, F. J. (2003). Political conservatism as motivated social cognition. Psychological Bulletin, 129(3), 339375. [Crossref], [PubMed], [Web of Science ®][Google Scholar]
  • Kawamura, Y., & Kusumi, T. (2017). The norm-dependent effect of watching eyes on donation. Evolution and Human Behavior, 38(5), 659666. [Crossref], [Web of Science ®][Google Scholar]
  • Koornneef, E. J., Dariel, A., Elbarazi, I., Alsuwaidi, A. R., Robben, P. B., & Nikiforakis, N. (2018). Surveillance cues do not enhance altruistic behavior among anonymous strangers in the field. PloS One, 13(8), e0197959. [Crossref], [PubMed], [Web of Science ®][Google Scholar]
  • Lewis-Beck, M. S., Norpoth, H., Jacoby, W. G., & Weisberg, H. F. (2008). The American voter revisited. Ann Arbor, MI: University of Michigan Press. [Crossref][Google Scholar]
  • Manesi, Z., & Pollet, T. (2017). No support for the watching eyes effect across three “lost letter” field experiments. Letters on Evolutionary Behavioral Science, 8(1), 1215. [Crossref][Google Scholar]
  • Manesi, Z., Van Lange, P. A., Van Doesum, N. J., & Pollet, T. V. (2018). What are the most powerful predictors of charitable giving to victims of typhoon Haiyan: Prosocial traits, socio-demographic variables, or eye cues? personality and individual differences. Available Online March, 17, 2018. [Google Scholar]
  • Mann, C. (2010). Is there backlash to social pressure? A large-scale field experiment on voter mobilization. Political Behavior, 32(3), 387407. [Crossref], [Web of Science ®][Google Scholar]
  • Matland, R. E., & Murray, G. R. (2013). An experimental test for “backlash” against social pressure techniques used to mobilize voters. American Politics Research, 41(3), 359386. [Crossref], [Web of Science ®][Google Scholar]
  • Matland, R. E., & Murray, G. R. (2016). I only have eyes for you: Does implicit social pressure increase voter turnout? Political Psychology, 37(4), 533550. [Crossref], [Web of Science ®][Google Scholar]
  • McDermott, R. (2011). Internal and external validity. In J. N. Druckman, D. P. Green, J. H. Kuklinski, & A. Lupia (Eds.), Cambridge handbook of experimental political science (pp. 2740). New York, NY: Cambridge University Press. [Crossref][Google Scholar]
  • Murray, G. R., & Matland, R. E. (2014). Mobilization effects using mail: Social pressure, descriptive norms, and timing. Political Research Quarterly, 67(2), 304319. [Crossref], [Web of Science ®][Google Scholar]
  • Northover, S. B., Pedersen, W. C., Cohen, A. B., & Andrews, P. W. (2017). Artificial surveillance cues do not increase generosity: Two meta-analyses. Evolution and Human Behavior, 38(1), 144153. [Crossref], [Web of Science ®][Google Scholar]
  • Open Science Collaboration. (2015). Estimating the reproducibility of psychological science. Science, 349(6251), aac4716. [Crossref], [PubMed], [Web of Science ®][Google Scholar]
  • Panagopoulos, C. (2011). Thank you for voting: Gratitude expression and voter mobilization. Journal of Politics, 73(3), 707717. [Crossref], [Web of Science ®][Google Scholar]
  • Panagopoulos, C. (2013). Positive social pressure and prosocial motivation: Evidence from a large-scale field experiment on voter mobilization. Political Psychology, 34(2), 265275. [Crossref], [Web of Science ®][Google Scholar]
  • Panagopoulos, C. (2014a). I’ve got my eyes on you: Implicit social pressure cues and prosocial behavior. Political Psychology, 35(1), 2333. [Crossref], [Web of Science ®][Google Scholar]
  • Panagopoulos, C. (2014b). Watchful eyes: Implicit observability cues and voting. Evolution and Human Behavior, 35(4), 279284. [Crossref], [Web of Science ®][Google Scholar]
  • Panagopoulos, C., & van der Linden, S. (2016). Conformity to implicit social pressure: The role of political identity. Social Influence, 11(3), 177184. [Taylor & Francis Online], [Web of Science ®][Google Scholar]
  • Pfattheicher, S., Schindler, S., & Nockur, L. (2018). On the impact of honesty-humility and a cue of being watched on cheating behavior. Journal of Economic Psychology. Available Online June, 23, 2018. [Google Scholar]
  • Pfattheicher, S., Strauch, C., Diefenbacher, S., & Schnuerch, R. (2018). A field study on watching eyes and hand hygiene compliance in a public restroom. Journal of Applied Social Psychology, 48(4), 188194. [Crossref], [Web of Science ®][Google Scholar]
  • Powell, K. L., Roberts, G., & Nettle, D. (2012). Eye images increase charitable donations: Evidence from an opportunistic field experiment in a supermarket. Ethology, 118(11), 16. [Crossref], [Web of Science ®][Google Scholar]
  • Teresi, H., & Michelson, M. R. (2015). Wired to mobilize: The effect of social networking messages on voter turnout. The Social Science Journal, 52(2), 195204. [Crossref], [Web of Science ®][Google Scholar]

Reprints and Permissions

Please note: We are unable to provide a copy of the article, please see our help page How do I view content?

To request a reprint or commercial or derivative permissions for this article, please click on the relevant link below.

Permission can also be obtained via Rightslink. For more information please visit our Permissions help page.