How does symbolic success affect redistribution in left-wing voters? A focus on the 2017 French presidential election.

Redistribution preferences depend on factors such as self-interest and political views. Recently, Deffains et al. (2016) reported that redistributive behavior is also sensitive to the actual experience of success or failure in a real effort task. While successful participants ('overachievers') are more likely to attribute their success to their effort rather than luck and opt for less redistribution, unsuccessful participants ('underachievers') tend to attribute their failure to external factors and opt for more redistribution. The aim of the present study was to test how the experience of success (symbolic success) and political views interact in producing redistributive behavior in an experimental setting. The study was conducted during the 2017 French presidential election. Our sample was biased towards left-wing, and most participants reported voting for Mélenchon, Hamon or Macron. Our findings reveal that 1) Macron voters redistribute less than Hamon voters who themselves redistribute less than Mélenchon voters, 2) overachievers redistribute less than underachievers only among Mélenchon voters. This suggests that redistributive behavior is governed primarily by political opinions, and that influence by exogenous manipulation of symbolic success is not homogenous across left-wing political groups.

Introduction

Support for redistribution varies greatly across individuals within a society, and is a major component of their political positioning. Political parties put forward different redistributive policies in their respective agendas. Accordingly, understanding the determinants of support for redistribution has been a topic of major interest for researchers in economics and political sciences.

One can distinguish two main factors contributing to this support, namely self-interest and fairness considerations [1]. On the one hand, the individual attitude towards a more redistributive or a less redistributive system is shaped by the economic self-interest of the individual, i.e. the effect that the redistributive system has upon the individual’s net income. Obviously, self-interest pushes wealthy individuals to support redistribution less than poor individuals. On the other hand, support for redistribution is also dependent on fairness considerations [2,3]. The redistributive policy chosen in a society reflects the beliefs about the determinants of income inequality and the main causes of poverty [1]. If wealth is primarily determined by chance or by factors that are not under the control of individuals, then support for redistribution increases [4,5], in accordance with the accountability principle [6].

Surveys have shown that such beliefs about the determinants of inequality are not homogeneous across the population [e.g. 7]. Relatedly, support for redistributive policies varies across social groups defined by race, gender, age or socioeconomic status [8]. In the United States, whites are more averse to redistribution than blacks, even after controlling for individual characteristics such as income, education, etc. [e.g., 9,10]. Past upward mobility also decreases the support for redistribution [e.g., 10,11]. Some of these observations have been confirmed by experimental data. For instance, when participants are presented with mock news articles reporting high (vs. low) rates of social mobility, their tolerance for inequality increased [12]. Providing American adults with factual information about the rise of inequalities in the United States (vs. control information) increased their beliefs that economic inequalities are due to structural rather than individual factors and increased support for redistribution [13,14].

The present work follows up on a recent study by Deffains, Espinosa, and Thöni [15] who introduced an exogenous manipulation of status and found this manipulation to affect the redistributive behavior of participants, even when self-interest was not at stake. After a real effort task, each subject was randomly given a status of either ‘overachiever’ (performance above the median) or ‘underachiever’ (performance below the median). In a subsequent disinterested dictator game, participants were asked to reallocate money between two randomly chosen individuals in their session, from the richest to the poorest individual. It turned out that on average, overachievers redistribute less than underachievers. The information provided to the subjects about the determinants of task performance (i.e. luck or effort) was very vague, and the authors found that overachievers also emphasized more the role of effort in their outcome than underachievers. Noteworthy, Deffains et al. suggested that participants exhibit a self-serving bias [16] by adopting beliefs favorable to them. More precisely, successful individuals attribute their own success to effort and others’ failure to a lack of effort, and in accordance with the accountability principle, they believe that no redistribution should take place. On the contrary, unsuccessful individuals attribute their own failure to bad luck and others’ success to favorable circumstances, so they support redistribution towards the most disadvantaged.

Since beliefs about the role of luck can be affected both by exogenous manipulations [15] and political opinions [e.g., 17], one could anticipate that these two factors may interact in their influence on redistributive behavior. The goal of the present study is to evaluate this interaction. To do so, we tested an exogenous manipulation of status much like Deffains et al., while evaluating political opinions of participants, in the context of the French 2017 presidential election.

Is the effect of status uniform across the different voters? More precisely, we hypothesized that the exogenous manipulation of Deffains et al. would have an effect on redistributive behavior for subjects who hold moderate political views but no effect for subjects who hold extreme political views. The 2017 French presidential election provided a unique opportunity to compare extreme voters to moderates. Indeed, in 2017, most electors moved away from the candidates of the two major traditional parties (Hamon for the left-wing “Parti Socialiste” vs. Fillon for the right-wing party “Les Républicains”), who together gathered only 25% of the votes in the first round. Instead, electors supported the moderate candidate Macron (who eventually won the election) and the candidates of radical parties (Mélenchon for the far-left and Le Pen for the far-right). In other words, as was seen in other western democracies in the last decade, this election moved away from the traditional left-right opposition towards a center-extreme polarization.

Method

Participants

A total of 649 unpaid participants completed the experiment (see “Description of our sample” below). Participants were essentially French people who responded to an announcement we posted on the Parisian Experimental Economics Laboratory (LEEP) portal, the Paris School of Economics portal, and the main social networks (Facebook and Twitter) inviting them to participate in an online survey on the presidential election. The website that hosted the experiment provided participants with all information about the research (the purpose and nature of the study, the voluntary nature of participation, and the possibility of withdrawing from the experiment at any time without any penalty or consequences). This research was reviewed and approved by Institutional Review Board–Ecole d’économie de Paris (approval number: IRB00010601). Written informed consent was obtained from all participants.

Procedure and measures

The experiment took place during the two weeks separating the two rounds of voting in the 2017 French presidential election (April 23-May 7). Participants first performed a computerized effort task without monetary reward linked to performance. This task was an Implicit Association Test (IAT) aimed at measuring their implicit attitude towards France (in its preliminary version, this study was intended to examine to what extent implicit and explicit attitudes predict participants’ voting intention. Because of our skewed sample, however, we could not really evaluate properly the voting intention towards Marine Le Pen. Thus, the variable “voting intention for the second round” was not taken into account in the analysis. We then focused our analysis essentially on the determinants of redistributive behavior). Participants were asked to respond as fast and accurately as possible, and they were informed that their performance would be their mean reaction time over the task. After completion of the task, participants were given a (fake) feedback on their performance and were randomly assigned to the overachiever or underachiever groups (status). Then, they completed a disinterested dictator game in which they were asked to reallocate money between two fictive individuals, a rich and a poor individual. The game was scripted as follows: “Imagine that 100 euros were allocated to two participants A and B based on their performance on the previous speeded-response task. A received 80 euros based on her good performance, B received 20 euros based on her weak performance. If you could reallocate the 100 euros to A and B, how would you reallocate them?” Participants chose the amount of money (between 50 and 100) they would allocate to A, B receiving the rest. Next, participants responded to five self-report items on a 7 points Likert scale measuring fatalism (“to what extent do you relate you performance to 1: chance or 7: effort); their views on income inequality (1: egalitarian, 7: liberal); their attitudes towards economic patriotism (“Do you think that the French government should take more patriotic measures in the economic and the social domain?”, 1: unfavorable, 7: favorable); their attitudes towards France (“Do you like France?” 1: positive, 7: negative); and their political position on the left-right continuum (1: extreme left, 7: extreme right). Then, participants reported their vote in the first round. Here, the response modalities included the 11 candidates involved plus the two options “I did not vote in the first round” and “I voted blank or null in the first round”. Finally, participants reported their voting intention for the second round. At this stage, four response modalities were presented: “I will vote for Macron”, “I will vote for Le Pen”, “I will vote blank or null”, and “I will not vote”.

Results

Description of our sample

Participants were 357 females and 292 males (mean age 33.62 years, SD = 15.44 years) (due to a technical error in the data collection, redistribution choices could only be analyzed for 626 participants). Regarding the socio-professional category, it turned out that managers and white-collar professions (34.36%) and students (41.60%) were overrepresented in our sample (both categories representing 75% of the sample). Regarding reported votes for the first round, our sample was clearly left-wing oriented, and voters for the two main right-wing candidates (Fillon and Le Pen) were underrepresented, whereas voters for Mélenchon, Hamon, and Macron were overrepresented (Fig 1). Therefore, in subsequent analyses we focus on participants who reported having voted for Mélenchon, Hamon, or Macron in the first round of the election (N = 506, 78% of the initial sample), given the lack of data for the other cases. Accordingly, in what follows the variable “First-round vote” is a categorical variable with 3 possible values, namely Mélenchon, Hamon, and Macron. Table 1 reports the age, gender, socio-economic category, and average status for the different group of voters in our sample.

Fig 1

Distribution of the First-round vote and comparison with actual results at the national level.

Table 1

Socio-demographic characteristics of participants reporting voting for Mélenchon, Hamon or Macron in the first round of the election in our dataset.

Participants who reported another vote are pooled together in this table, and were not analyzed further in the present study.

First-round vote	N	Age (SD)	Gender (% women)	Occupation (% White Collar)	Occupation (% Student)	Status (% overachiever)
Mélenchon	219	31.21 (12.42)	0.62	0.29	0.45	0.46
Hamon	121	32.57 (14.59)	0.67	0.33	0.45	0.57
Macron	166	37.22 (16.75)	0.45	0.47	0.32	0.47
Other	120	32.40 (14.65)	0.46	0.23	0.46	0.43

As this selection resulted in a restriction of variance of the Political position variable (Fig 2), we considered the vote reported for the first round (hereafter First-round vote) as the only measure of political opinions in the analysis. Note that at the time of this experiment, Mélenchon, Hamon, and Macron were all considered left-wing candidates. Specifically, Mélenchon was considered as the main candidate of the radical left, Hamon was the official candidate of the major French left-wing party (“Parti Socialiste”), and Macron was associated both with a left-wing government under former president Hollande and with a social-liberal position with a pronounced liberal component.

Fig 2

Relation between First-round vote and political position (1: extreme left, 7: extreme right).

Status manipulation

Table 2 reports age, gender, socio-economic category and First-round vote of overachievers and underachievers, showing that our random manipulation of status did not create an unwanted bias between overachievers and underachievers.

Table 2

Comparison of First-round vote and the socio-demographic characteristics of participants in the overachiever and underachiever conditions in our final sample (i.e. including only participants who reported voting for Mélenchon, Hamon or Macron).

Condition	N	Age (SD)	Gender (% women)	First-round vote Mélenchon / Hamon / Macron
Overachiever	248	33.73 (15.34)	0.54	0.41 / 0.28 / 0.31
Underachiever	258	33.28 (14.06)	0.60	0.46 / 0.20 / 0.34
test		t = 0.344	χ2 = 1.64	χ2 = 1.10 / 3.67 / 0.29
p		0.731	0.2	0.29 / 0.06 / 0.59

Redistributive behavior and self-report measures

Table 3 indicates the descriptive statistics (means and standard deviations) and the correlations between the different behavioral and personality measures. We note that almost all pairwise correlations measures were significant, except for the correlation between Fatalism and Political position. In particular, the share given to the richer player in the disinterested dictator game, which quantifies participants’ attitude towards income inequality in a simple behavioral test, was correlated positively with the explicit attitude towards inequality (r = 0.24, p<0.001), which was most correlated with the stated political position (r = 0.60, p<0.001).

Table 3

Descriptive statistics and zero-order correlations between individual measures: The share left to the richer agent in a disinterested dictator game, fatalism (relating performance to 1: chance 7: effort), income equality (from 1: egalitarian to 7: liberal), economic patriotism (from 1: unfavorable to 7: favorable), attitude towards France (from 1: negative to 7: positive), and political position (from 1: extreme left to 7: extreme right) (N = 649).

		Mean	SD	1	2	3	4	5	6
1	Disinterested dictator	60.08	15.19	—
2	Fatalism	4.78	1.51	.17***	—
3	Income inequality	2.98	1.78	.24***	.10**	—
4	Economic patriotism	3.40	1.91	.16***	.09*	.20***	—
5	Attitude France	5.72	1.34	.14***	.19***	.20***	.16***	—
6	Political position	2.68	1.53	.22***	.04	.60***	.27***	.19***	—

*p < .05,

**p < .01,

***p < .001, two-tailed.

Effect of status on redistribution

The main point of interest of the analysis was how the amount of money (between 50 and 100) reallocated to the “richer agent” A in the disinterested dictator game was affected by the manipulation of Status and by the vote reported by participants. To assess this, we conducted a 2 (First-round vote) × 2 (Status) ANOVA for independent samples on the share given to A as a dependent variable. This analysis yielded a significant main effect of First-round vote (F(1, 500) = 8.65, p = .0002, ηp 2 = 0.0334), with no main effect of Status (F(1, 500) = 1.72, p = .19, ηp 2 = .0034), but an interaction between First-round vote and Status (F(2, 500) = 3.16, p = .043, ηp 2 = .0124). The main effect of First-round vote confirmed our expectations that participants who reported voting for more left-wing candidates would also exhibit greater redistribution. Indeed, Mélenchon voters allocated a mean amount of 58.45 to A (SD = 11.50), Hamon voters 60.54 (SD = 12.04), and Macron voters 63.65 (SD = 13.18). The manipulation of Status did not produce a statistically significant effect, although the observed pattern was in the expected direction. Indeed, overachievers allocated slightly more to A (M = 61.39, SD = 12.81) than underachievers (M = 59.95, SD = 11.94). This is comparable to what was reported by Deffains, et al. [15] (converted to our measure, they respectively obtained M = 61.46, SD = 14.8 for overachievers and M = 59.95, SD = 11.94 for underachievers). With regards to the interaction, we expected that Status would have an effect on redistributive behavior for participants with moderate political views (Hamon and Macron voters), but not necessarily for those with extreme political views (Mélenchon voters), whom we expected to strongly redistribute irrespectively of our Status manipulation. Inspection of the different groups (Table 4) however revealed a different pattern, and separate analyses for each group of voters indicated that the effect of Status was significant for Mélenchon voters (F(1, 217) = 6.30, p = .0128, ηp 2 = .0282), but not for Hamon voters (F(1, 119) = 1.99, p = .16, ηp 2 = .0164) or Macron voters (F(1, 164) = 0.50, p = .47, ηp 2 = .0030) . In other words, we found the opposite pattern to that expected.

Table 4

Means and standard errors of the amount of money allocated to A in the disinterested dictator game as a function of First-round vote and status.

First-round vote Status	Mélenchon	Hamon	Macron
Overachiever	60.5 (1.28)	59.2 (1.30)	64.4 (1.58)
Underachiever	56.7 (0.91)	62.3 (1.86)	63.0 (1.33)

Noteworthy, to evaluate whether our results were robust to changes in model specification, we conducted a new regression analysis, in which we added gender and age as covariates (Table 5). This regression revealed that gender affected redistribution, with women redistributing more than men, replicating previous findings [e.g., 2,18]. This analysis also indicated a main effect of the First-round vote, and confirmed the interaction between Status and First-round vote. When examining the effect of Status separately for the 3 groups of voters, again adding gender and age as covariates, we found that redistributive behavior was affected by Status only for Mélenchon voters (F(1, 215) = 5.54, p = .020, ηp 2 = .0251), but not for Hamon voters (F(1, 117) = 2.09, p = .151, ηp 2 = .0176) or Macron voters (F(1, 162) = 0.454, p = .502, ηp 2 = .0028) replicating our main finding. For completeness, we also report in S1 Appendix the results of a regression over all participants, including those who reported a different First-round vote.

Table 5

ANOVA table for redistributive behavior in the disinterested dictator game.

The different factors included in the model are the effects of gender, age, status in the experiment (overachiever vs. underachiever), First-round vote and the interaction between status and First-round vote.

	ηp 2	S.S.	d.f.	F	p
Gender	0.0080	584	1	4.001	.046
Age	0.0051	373	1	2.553	.111
Status	0.0026	188	1	1.291	.256
Vote1	0.0239	1781	2	6.101	.002
Status:Vote1	0.0119	874	2	2.992	.051
Residuals		72701	498

Finally, we evaluated the effect of Status on the fatalism measure, that is, the extent to which participants related their performance to chance or effort. We found no evidence that fatalism was affected by Status, F(1, 504) = 0.40, NS. In other words, we found no evidence for a self-serving bias in our participants, unlike Deffains et al. [15]. Note that the task used in the present study was an IAT, which was objectively less effortful than the counting task used by Deffains et al. Therefore, unlike their participants, participants in our study might have not believed that their performance could be impacted by the amount of effort deployed.

Discussion

The present study capitalized on a major political election (the 2017 French presidential election) in order to investigate how redistributive behavior is affected by political views and–experimentally induced–symbolic success [15]. We found an overall effect of First-round vote on redistribution such that the mean amount redistributed by the three main groups of voters in our sample was coherent with their respective positions on the left-right continuum. While participants who reported voting for Mélenchon (presumably the most leftists) were the most redistributive, Macron voters (the most liberal) were less redistributive, with Hamon voters falling in between. This finding confirms previous research reporting that preferences for redistribution and progressive taxation are coherent with vote choice: during the French 2012 presidential elections, strong supporters of redistribution voted for the left-wing candidate Hollande, while supporters of a flat rate tax voted for the right-wing candidate Sarkozy [11].

Our main result is that redistributive behavior is influenced by the exogenous manipulation of Status only in a subgroup of participants, specifically those who reported voting for Mélenchon. Therefore, our study partially replicated the findings of Deffains and colleagues [15]. This partial discrepancy between our study and that of Deffains might be due to incentives. In Deffains’ study, participants’ redistribution choices in the dictator game had real consequences on the payoffs of other players, whereas in our paradigm redistribution choices were only hypothetical. It is possible that incentives might have influenced our results independently of the desirability bias. Participants who reported voting for Hamon or Macron might be more sensitive to the presence of real life incentives than Mélenchon voters. Thus, incentivizing redistribution choices might be a necessary feature to obtain the effect of Status in Hamon or Macron voters, whereas Mélenchon voters would exhibit the effect of Status even in the absence of incentives. To evaluate these possibilities, further research would need to compare redistribution choices with and without incentives, for the different groups of voters.

It has been proposed [e.g., 19] that in the absence of incentives, participants might try to please the experimenter or conform to some social norms, e.g. by being generous in dictator games. Could this desirability bias explain our results or the difference between our study and Deffains’ study? We believe that such an explanation is unlikely for several reasons. First, if a desirability bias was more present our study than in Deffains’ study, then we should have observed more redistribution in our participants. However, in our experiment, participants redistributed less than in Deffains’ study: our mean allocation to A was 60.08 while the corresponding value in Deffains’ study would be 57.56. Second, and more generally, it is not clear to us why this desirability bias would lead to the specific interaction between Status and First-round vote. Third, the instructions given to participants (see S2 Appendix) did not refer to the aim of our experiment, so participants were naïve about our hypothesis. Had they tried to guess our expectations, we would have found an effect of status on fatalism, which we did not observe either in the full sample (p = .52) nor in Mélenchon voters (p = .35), whose redistributive behavior was affected by status however. Finally, our experiment was conducted online and responses were anonymous, so participants have no pressure to please the experimenter or conform to social norms.

Our study provided a nuanced picture of how redistributive behavior is jointly influenced by political views and the actual experience of individuals (here, the experience of success or failure in a simple decision task). In fact, we hypothesized that the exogenous manipulation of Status would have an effect on redistributive behavior for subjects who hold moderate political views (Hamon or Macron voters), but no effect for subjects who hold extreme political views (Mélenchon voters) who would be more likely to resist any experimental manipulation. Our findings revealed a significant interaction between Status and First-round vote but the pattern we found is the opposite of our expectation, as the only group of voters who were significantly affected by status were Mélenchon voters. Being the most left-wing voters in our sample, endorsing pronounced egalitarian views of society, these voters were supposed to be the most redistributive overall (which was actually observed) but also the least sensitive to the information regarding Status (which was the opposite of what we observed). That result is even more surprising since they reported the most egalitarian views on income (M = 2.07) compared to Hamon voters (M = 2.60) and Macron voters (M = 3.86), F(2, 503) = 69, p < 0.001. Explanations of this finding in terms of age, sex, or socio-economic status are unlikely in our dataset as Mélenchon voters and Hamon voters did not differ significantly on these variables (Table 1). In addition, we verified that our Status manipulation was truly random with respect to age, sex, or socio-economic status, which did not differ between overachievers and underachievers (Table 2).

Here, we suggest one explanation for our finding that Mélenchon voters were the most affected by Status manipulation. It is worth noting that these voters were also the most versatile at the end of the electoral campaign. Indeed, the dynamics of voting intentions as measured by the polls during the month preceding the first round revealed that voting intentions for Macron remained stable around 23%, those for Hamon collapsed from 12% to 6%, while those for Mélenchon jumped from 11% to 18%. As a candidate, Mélenchon also used a communication strategy based on social influence, with a strong presence on social media, and a populist attitude that emphasized the proximity to his base (“the people”). Individuals that are highly susceptible to social influence were then more likely to become Mélenchon voters, and in our study they were also more likely to be influenced by the Status manipulation. Thus, our result could be explained by susceptibility to social influence as a common cause of voting behavior and of the effect of the Status manipulation.

Before concluding, we must acknowledge several important limitations of our study. First, our sample was limited and was not representative of the French population. In particular, our data could not allow us to investigate the sensitivity of redistribution behavior to an experimental manipulation of success for right-wing voters. One main reason for this limitation was probably the recruitment procedure employed, which was based on social media, local networks, and word of mouth. As a result, according to their reports, our participants were mostly young, left-wing supporters, and closely related to the academic sector. In particular, we did not have enough right-wing or far-right supporters to perform meaningful analyses on this part of the political spectrum. By contrast, analyses of Twitter activity at that period revealed the emergence of three main communities in the French political environment, namely supporters of Macron and Hamon, supporters of Le Pen, and supporters of Mélenchon [20]. It is possible that redistributive behavior and its sensitivity to our manipulation would have been different for right-wing and far-right voters.

We note that although our results may not be representative of right-wing voters, one could envision that they would generalize for left-wing voters in some other countries. Indeed, in the last decade, western democracies have seen a polarization of opinions, with a crisis of the traditional parties and a rise of support for extreme populist parties. Examples of populist far-left parties are die Linke in Germany, Podemos in Spain, Siriza in Greece, La France Insoumise (Jean-Luc Mélenchon’s party) in France. According to Rooduijn and Akkerman [21] these radical left parties have in common that “They do not focus on the ‘proletariat’, but glorify a more general category: the ‘good people’” , contrary to former communist parties and that”they do not reject the system of liberal democracy as such, but only criticize the political and/or economic elites within that system”. Our results regarding the susceptibility of Mélenchon voters to status manipulation could thus be evaluated and replicated in other countries.

In addition, one could argue that a second limitation of the present work is related to the specific timing of the study, which took place during the French presidential election. This specific timing was chosen on purpose for two reasons. One reason was to benefit from the increased interest towards political topics at this time. The other reason was to probe voters’ redistributive behavior at a time that constitutes an important step in the democratic process. However, we acknowledge that it is possible that voters’ behavior in our study is unusual, because of this unusual timing. Voters may receive more information in the context of an election, and they may react more strongly to information delivered in this context. Whether our results would generalize to another context unrelated to a particular election thus remains an open empirical issue.

The third limitation relates to the possible discrepancy between actual votes and reported votes in our participants. Poll estimates (based on self-reported votes) and actual votes can indeed differ, as famously illustrated in the 2016 US presidential election, the 2016 “Brexit” referendum, or the 2002 French presidential election, amongst others. However, we note that in the case of the election under study here (2017 French presidential election) the last polls were very accurate. One reason for the discrepancy between self-reported votes and actual votes might be a social desirability bias by which right-wing or far-right votes are expressed less easily and therefore under-estimated in opinion polls [see e.g. 22]. Critically, polling institutes use adjustment procedures to take into this bias when producing their estimates, but we did not. Therefore, right-wing opinions/votes in our sample might have been under-estimated. In sum, although we followed the common practice in studies of voting behavior, and used the terms “Mélenchon voters”, “Hamon voters” or “Macron voters”, one should bear in mind that our data is about self-reported votes, which might have differed from actual votes.

To conclude, our findings revealed that self-reported far-left voters turned out to be the more sensitive to the exogenous manipulation of symbolic success. This leads to three remarks. Firstly, we need further research to better understand to what extent, and in which groups, redistributive behavior can be manipulated through exogenous manipulations of experience of success. In particular, further studies are needed that shall use a proper manipulation of symbolic success and representative samples in terms of political and socio-economic features. Secondly, our findings suggest that the various political groups process information differently, that is, they are not cognitively homogeneous [e.g., 23–25]. Finally, and more broadly, the fact that Mélenchon voters displayed a different behavior than Hamon and Macron voters extends recent findings showing that supporters of extreme political groups have different characteristics from those with more moderate views, although they are not necessarily different on socio-demographic variables such as age or level of education [e.g. 26]. For instance, Hanel, Zarzeczna, and Haddock [27] reported that extreme (left-wing or right-wing) supporters are usually more heterogeneous than moderate ones in terms of human values and politics-related variables such as attitudes toward immigrants and trust in institutions. In the current social and political context, we believe that understanding further these differences, especially whether some groups are more susceptible to influence than others, appears a worthwhile subject for future research. Using controlled experiments during political elections can be a useful tool in such research.

ANOVA table for redistributive behavior in the disinterested dictator game including all groups of voters.

(DOCX)

Click here for additional data file.

Instructions (screen shot and translation).

(DOCX)

Click here for additional data file.

19 Sep 2019

Submitted filename: Rebuttal letter.docx

Click here for additional data file.

16 Oct 2019

PONE-D-19-26429

How symbolic success affects redistributive behavior? An experiment based on the 2017 French presidential election

PLOS ONE

Dear Dr. Berthet,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please find below the reviewer's comments, as well as those from my own.

We would appreciate receiving your revised manuscript by Nov 30 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Valerio Capraro

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

Additional Editor Comments (if provided):

I have now collected one review from one expert in the field. Unfortunately, I was unable to find a second reviewer. However, since this review is very detailed and thorough, I have opted for making a decision on your manuscript with only one review. As you will see, the reviewer is quite positive, but has several major comments. Therefore, I would like to invite you to revise your manuscript according to the reviewer's comment. Additionally, after reading your manuscript, I would like to add one more comment. I have myself done a lot of work on redistributive behavior, which I think is quite relevant to your discussion regarding the determinants of redistributive behavior, so you might want to have a look at it (of course, citing this work is not a requirement). In Capraro & Rand (2018) and Tappin & Capraro (2018) we found that redistributive behavior in the trade-off game is driven by moral preferences for doing the right thing and, consequently, it very much depends on how the decision problem is framed; in Capraro (2019), I found that women are more likely than men to choose an equitable distribution over an efficient one.

Looking forward for the revision.

References

Capraro V (2019) Gender differences in the equity-efficiency trade-off. Available at https://ssrn.com/abstract=3386124.

Capraro V, Rand DG (2018) Do the right thing: Experimental evidence that preferences for moral behavior, rather than equity and efficiency per se, drive human prosociality. Judgment and Decision Making 13, 99-111.

Tappin BM, Capraro V (2018) Doing good vs. avoiding bad in prosocial choice: A refined test and extension of the morality preference hypothesis. Journal of Experimental Social Psychology 79, 64-70.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Referee’s report - Manuscript ID PONE-D-19-26429

Title: " How symbolic success affects redistributive behavior? An experiment based on the 2017 French presidential election" for PLOS ONE

This paper uses experimental data to examine how redistributive preferences are shaped by the experience of success and political views. This experiment was implemented during the recent presidential election in France. It finds that redistributive behavior is affected primarily by political opinions e.g. Mélenchon voters redistribute more than other groups. In addition, the paper shows that redistributive behavior is influenced by the exogenous manipulation of experiences of success. Interestingly, this effect is not homogenous across political groups and only the Mélenchon voters were significantly affected by the status.

The paper is well-written and competently conducted. The literature review is fine although I would suggest referring also to the literature investigating experimentally if and how information about actual inequality (and political position) affects policy preferences. Overall, this paper investigates a relevant question and I enjoyed reading it.

Despite that, I am not sure about the results. I have a number of concerns and I turn to discuss each of these issues in the remainder of this report.

First, the sample is heavily biased to left position and is far to be representative of the overall population. Although, I appreciate that authors are well aware about this limitation, I still think that this is a big issue. The worry is that one would expect the left-wing participants to possibly have quite different preferences from the general population. How much we can take from this work and extend to other groups or even other contexts? Moreover, we have also to take in consideration the timing of the survey. Indeed, voters usually receive more information in the period of elections and may react to them more strongly - at least in the short term.

Second, there are not enough information about groups characteristics. Are participants in the overachiever or underachiever groups different in their characteristics? Did the exclusions of some observations affect the balancing properties of the sample? How much people voting Hamon are different from those voting Macron, for example? These are very important information to understand the validity of the final results.

Third, the findings may be biased by the fact that tasks are not incentivized as they would be in an ideal scenario. In particular, I worry that the lack of impact on most of the voters might be the consequence of the fact that the questions are asked hypothetically. A final question is raised as to whether one can believe that participants might have a desire to please who developed the study. However, there are not enough information on the structure of the survey, so it is very difficult to evaluate the validity of the research design.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

4 Nov 2019

Editor’s letter :

I have now collected one review from one expert in the field. Unfortunately, I was unable to find a second reviewer. However, since this review is very detailed and thorough, I have opted for making a decision on your manuscript with only one review. As you will see, the reviewer is quite positive, but has several major comments. Therefore, I would like to invite you to revise your manuscript according to the reviewer's comment. Additionally, after reading your manuscript, I would like to add one more comment. I have myself done a lot of work on redistributive behavior, which I think is quite relevant to your discussion regarding the determinants of redistributive behavior, so you might want to have a look at it (of course, citing this work is not a requirement). In Capraro & Rand (2018) and Tappin & Capraro (2018) we found that redistributive behavior in the trade-off game is driven by moral preferences for doing the right thing and, consequently, it very much depends on how the decision problem is framed; in Capraro (2019), I found that women are more likely than men to choose an equitable distribution over an efficient one.

Looking forward for the revision.

References

Capraro V (2019) Gender differences in the equity-efficiency trade-off. Available at https://ssrn.com/abstract=3386124.

Capraro V, Rand DG (2018) Do the right thing: Experimental evidence that preferences for moral behavior, rather than equity and efficiency per se, drive human prosociality. Judgment and Decision Making 13, 99-111.

Tappin BM, Capraro V (2018) Doing good vs. avoiding bad in prosocial choice: A refined test and extension of the morality preference hypothesis. Journal of Experimental Social Psychology 79, 64-70.

We thank you for bringing this work to our attention. We now refer to Capraro & Rand (2018) in our introduction. In addition, we now include gender as a covariate in our regression analysis. We have found indeed that women redistributed more than men, but this did not change our main results. Regarding this gender effect, we refer to Capraro (2019).

We now report our regression analysis as follows:

“Noteworthy, to evaluate whether our results were robust to changes in model specification, we conducted a new regression analysis, in which we added gender, age and the 5 self-report measures as covariates. This regression revealed that fatalism, views on income equality, and economic patriotism were significant predictors of redistributive behaviors. In addition, we found that gender affected redistribution, with women redistributing more than men, replicating previous findings (e.g. Capraro, 2019; Corneo & Grüner, 2002). We note that the main effect of the First-round vote did not reach significance in this new analysis, but critically the interaction between Status and First-round vote remained significant (Table 2). When examining the effect of Status separately for the 3 groups of voters, again adding the 5 self-report measures as covariates, we found that redistributive behavior was affected by Status only for Mélenchon voters (F(1, 210) = 6.03, p = .0149, ηp 2 = .0279), but not for Hamon voters (F(1, 112) = 2.78, p = .098, ηp 2 = .0242) or Macron voters (F(1, 157) = .790, p = .376, ηp 2 = .0050). replicating our main finding.”

We have also modified Table 2 accordingly. The new Table 2 is as follows:

ANOVA table for redistributive behavior in the disinterested dictator game. The different factors included in the model are the effects of Gender, Age, self-report measures (attitude towards Income inequality, Economic patriotism, Fatalism, Attitude towards France, Political position), Status in the experiment (Overachiever vs. Underachiever), First-round vote and the interaction between Status and First-round vote.

ηp 2 S.S. d.f. F p

Gender 0.0097 628 1 4.81 .029 *

Age 0.0006 36 1 0.28 .598

Income inequality 0.0135 879 1 6.73 .010 **

Economic patriotism 0.0139 906 1 6.94 .009 **

Fatalism 0.0482 3260 1 24.97 <.001 ***

Attitude France 0.0030 193 1 1.48 .225

Political position 0.0018 119 1 0.91 .340

Status 0.0029 187 1 1.43 .232

Vote1 0.0066 425 2 1.63 .197

Status:Vote1 0.0148 967 2 3.70 .025 *

Residuals 64360 493

Reviewer #1: Referee’s report - Manuscript ID PONE-D-19-26429

This paper uses experimental data to examine how redistributive preferences are shaped by the experience of success and political views. This experiment was implemented during the recent presidential election in France. It finds that redistributive behavior is affected primarily by political opinions e.g. Mélenchon voters redistribute more than other groups. In addition, the paper shows that redistributive behavior is influenced by the exogenous manipulation of experiences of success. Interestingly, this effect is not homogenous across political groups and only the Mélenchon voters were significantly affected by the status.

The paper is well-written and competently conducted. The literature review is fine although I would suggest referring also to the literature investigating experimentally if and how information about actual inequality (and political position) affects policy preferences. Overall, this paper investigates a relevant question and I enjoyed reading it.

We thank the reviewer for these positive comments.

Despite that, I am not sure about the results. I have a number of concerns and I turn to discuss each of these issues in the remainder of this report.

First, the sample is heavily biased to left position and is far to be representative of the overall population. Although, I appreciate that authors are well aware about this limitation, I still think that this is a big issue. The worry is that one would expect the left-wing participants to possibly have quite different preferences from the general population. How much we can take from this work and extend to other groups or even other contexts? Moreover, we have also to take in consideration the timing of the survey. Indeed, voters usually receive more information in the period of elections and may react to them more strongly - at least in the short term.

We are indeed aware of this limitation, which was already acknowledged in the discussion. In the revised manuscript, we have made explicit the possibility that right-wing or far-right voters could have exhibited a different behavior.

We now also highlight that the timing of the survey is an important part of the specific context of our study. Generalization to other populations and other contexts is always an issue in experimental studies. However, we believe that since it is an empirical issue, it can be tackled in future work. We have added the following paragraph in the discussion section to address this point:

“In addition, one could argue that a second limitation of the present work is related to the specific timing of the study, which took place during the French presidential election. This specific timing was chosen on purpose for two reasons. One reason was to benefit from the increased interest towards political topics at this time. The other reason was to probe voters’ redistributive behavior at a time that constitutes an important step in the democratic process. However, we acknowledge that it is possible that voters’ behavior in our study is unusual, because of this unusual timing. As voters usually receive more information in the context of an election, it is possible that their reaction to the information we delivered during the study (i.e. the random allocation to the overachiever vs. underachiever group) was affected in these contexts. Whether our results would generalize to another context unrelated to a particular election thus remains an open empirical issue.”

Second, there are not enough information about groups characteristics. Are participants in the overachiever or underachiever groups different in their characteristics? Did the exclusions of some observations affect the balancing properties of the sample? How much people voting Hamon are different from those voting Macron, for example? These are very important information to understand the validity of the final results.

We thank the reviewer for this comment. We have clarified in the discussion that group characteristics did not differ between Mélenchon voters and Hamon voters (Macron voters were older and more male on average), and between overachievers and underachievers.

“Explanations of this finding in terms of age, sex, or socio-economic status are unlikely in our dataset as Mélenchon voters and Hamon voters did not differ significantly on these variables. In addition, we verified that our Status manipulation was truly random with respect to age, sex, or socio-economic status, which did not differ between overachievers and underachievers.”

In addition, we have now added Table 3 to describe these variables for the 3 group of voters.

Table 3. Socio-demographic characteristics of participants reporting voting for Mélenchon, Hamon or Macron (in the first round of the election) in our dataset.

First-round vote Age (SD) Gender

(% women) Occupation

(% White Collar) Occupation

(% Student)

Mélenchon 31.21 (12.42) 0.62 0.29 0.45

Hamon 32.57 (14.59) 0.67 0.33 0.45

Macron 37.22 (16.75) 0.45 0.47 0.32

Third, the findings may be biased by the fact that tasks are not incentivized as they would be in an ideal scenario. In particular, I worry that the lack of impact on most of the voters might be the consequence of the fact that the questions are asked hypothetically. A final question is raised as to whether one can believe that participants might have a desire to please who developed the study. However, there are not enough information on the structure of the survey, so it is very difficult to evaluate the validity of the research design.

Indeed, our measures are based on choices that were not incentivized. We have now added one paragraph in the discussion to highlight this difference between our study and the study of Deffains et al., and the role it may have played in our results.

“Before discussing further our work, we would like to highlight one difference between our study and that of Deffains and colleagues that might be key in explaining this partial discrepancy. In Deffains et al., participants’ redistribution choices in the dictator game had real consequences on the payoffs of other players, whereas in our paradigm redistribution choices were only hypothetical. It has been proposed (e.g. Camerer & Hogarth, 1999) that in the absence of incentives, participants might try to please the experimenter or conform to some social norms, e.g. by being generous in dictator games. It is not clear to us why this “desirability bias” would lead to the specific interaction between Status and First-round vote. Besides, since our experiment was conducted online and responses were anonymous, such an explanation in terms of desirability bias seems unlikely to us. Nevertheless, it is possible that participants who reported voting for Hamon or Macron are more sensitive to the presence of real life incentives, and that a true implementation of their redistribution choices was a necessary feature to obtain the effect of Status. By contrast, it could be that Mélenchon voters are less sensitive to the presence of real incentives, and would exhibit the effect of Status even in the absence of incentives. To evaluate these possibilities, further research would need to compare redistribution choices with and without incentives, for the different groups of voters.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

14 Nov 2019

PONE-D-19-26429R1

How symbolic success affects redistributive behavior? An experiment based on the 2017 French presidential election

PLOS ONE

Dear Dr. Berthet,

Thanks for your email. Please resubmit the paper by uploading the correct files.

We would appreciate receiving your revised manuscript by Dec 29 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Valerio Capraro

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

Thanks for your email. Please resubmitted the paper attaching the correct files.

[Note: HTML markup is below. Please do not edit.]

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

14 Nov 2019

Editor’s letter :

I have now collected one review from one expert in the field. Unfortunately, I was unable to find a second reviewer. However, since this review is very detailed and thorough, I have opted for making a decision on your manuscript with only one review. As you will see, the reviewer is quite positive, but has several major comments. Therefore, I would like to invite you to revise your manuscript according to the reviewer's comment. Additionally, after reading your manuscript, I would like to add one more comment. I have myself done a lot of work on redistributive behavior, which I think is quite relevant to your discussion regarding the determinants of redistributive behavior, so you might want to have a look at it (of course, citing this work is not a requirement). In Capraro & Rand (2018) and Tappin & Capraro (2018) we found that redistributive behavior in the trade-off game is driven by moral preferences for doing the right thing and, consequently, it very much depends on how the decision problem is framed; in Capraro (2019), I found that women are more likely than men to choose an equitable distribution over an efficient one.

Looking forward for the revision.

References

Capraro V (2019) Gender differences in the equity-efficiency trade-off. Available at https://ssrn.com/abstract=3386124.

Capraro V, Rand DG (2018) Do the right thing: Experimental evidence that preferences for moral behavior, rather than equity and efficiency per se, drive human prosociality. Judgment and Decision Making 13, 99-111.

Tappin BM, Capraro V (2018) Doing good vs. avoiding bad in prosocial choice: A refined test and extension of the morality preference hypothesis. Journal of Experimental Social Psychology 79, 64-70.

We thank you for bringing this work to our attention. We now refer to Capraro & Rand (2018) in our introduction. In addition, we now include gender as a covariate in our regression analysis. We have found indeed that women redistributed more than men, but this did not change our main results. Regarding this gender effect, we refer to Capraro (2019).

We now report our regression analysis as follows:

“Noteworthy, to evaluate whether our results were robust to changes in model specification, we conducted a new regression analysis, in which we added gender, age and the 5 self-report measures as covariates. This regression revealed that fatalism, views on income equality, and economic patriotism were significant predictors of redistributive behaviors. In addition, we found that gender affected redistribution, with women redistributing more than men, replicating previous findings (e.g. Capraro, 2019; Corneo & Grüner, 2002). We note that the main effect of the First-round vote did not reach significance in this new analysis, but critically the interaction between Status and First-round vote remained significant (Table 2). When examining the effect of Status separately for the 3 groups of voters, again adding the 5 self-report measures as covariates, we found that redistributive behavior was affected by Status only for Mélenchon voters (F(1, 210) = 6.03, p = .0149, ηp 2 = .0279), but not for Hamon voters (F(1, 112) = 2.78, p = .098, ηp 2 = .0242) or Macron voters (F(1, 157) = .790, p = .376, ηp 2 = .0050). replicating our main finding.”

We have also modified Table 2 accordingly. The new Table 2 is as follows:

ANOVA table for redistributive behavior in the disinterested dictator game. The different factors included in the model are the effects of Gender, Age, self-report measures (attitude towards Income inequality, Economic patriotism, Fatalism, Attitude towards France, Political position), Status in the experiment (Overachiever vs. Underachiever), First-round vote and the interaction between Status and First-round vote.

ηp 2 S.S. d.f. F p

Gender 0.0097 628 1 4.81 .029 *

Age 0.0006 36 1 0.28 .598

Income inequality 0.0135 879 1 6.73 .010 **

Economic patriotism 0.0139 906 1 6.94 .009 **

Fatalism 0.0482 3260 1 24.97 ***

Attitude France 0.0030 193 1 1.48 .225

Political position 0.0018 119 1 0.91 .340

Status 0.0029 187 1 1.43 .232

Vote1 0.0066 425 2 1.63 .197

Status:Vote1 0.0148 967 2 3.70 .025 *

Residuals 64360 493

Reviewer #1: Referee’s report - Manuscript ID PONE-D-19-26429

This paper uses experimental data to examine how redistributive preferences are shaped by the experience of success and political views. This experiment was implemented during the recent presidential election in France. It finds that redistributive behavior is affected primarily by political opinions e.g. Mélenchon voters redistribute more than other groups. In addition, the paper shows that redistributive behavior is influenced by the exogenous manipulation of experiences of success. Interestingly, this effect is not homogenous across political groups and only the Mélenchon voters were significantly affected by the status.

The paper is well-written and competently conducted. The literature review is fine although I would suggest referring also to the literature investigating experimentally if and how information about actual inequality (and political position) affects policy preferences. Overall, this paper investigates a relevant question and I enjoyed reading it.

We thank the reviewer for these positive comments.

Regarding the comment about the experimental literature, we have now expended the third paragraph of introduction to include some references showing that providing information about actual inequality affects support for redistribution. The paragraph now reads as follows:

“Surveys have shown that such beliefs about the determinants of inequality are not homogeneous across the population (e.g. Cozzarelli, Wilkinson, Tagler, 2001). Relatedly, support for redistributive policies varies across social groups defined by race, gender, age or socioeconomic status (Keely & Tan, 2008). In the United States, whites are more averse to redistribution than blacks, even after controlling for individual characteristics such as income, education, etc. (e.g., Gilens, 1999; Alesina & La Ferrara, 2005). Past upward mobility also decreases the support for redistribution (e.g. Guillaud & Sauger, 2013; Alesina & La Ferrara, 2005). Some of these observations have been confirmed by experimental data. For instance, when participants are presented with mock news articles reporting high (vs. low) rates of social mobility, their tolerance for inequality increased (Shaffir, Wiwad, Aknin, 2016). Providing American adults with factual information about the rise of inequalities in the United States (vs. control information) increased their beliefs that economic inequalities are due to structural rather than individual factors and increased support for redistribution (McCall et al., 2017; Boudreau & MacKenzie, 2018).”

Despite that, I am not sure about the results. I have a number of concerns and I turn to discuss each of these issues in the remainder of this report.

First, the sample is heavily biased to left position and is far to be representative of the overall population. Although, I appreciate that authors are well aware about this limitation, I still think that this is a big issue. The worry is that one would expect the left-wing participants to possibly have quite different preferences from the general population. How much we can take from this work and extend to other groups or even other contexts? Moreover, we have also to take in consideration the timing of the survey. Indeed, voters usually receive more information in the period of elections and may react to them more strongly - at least in the short term.

We are indeed aware of this limitation due to the bias in our sample, which was already acknowledged in the discussion. In the revised manuscript, we have made more explicit the possibility that right-wing or far-right voters could have exhibited a different behavior.

We now also highlight that the timing of the survey is an important part of the specific context of our study. Generalization to other populations and other contexts is always an issue in experimental studies. However, we believe that since it is an empirical issue, it can be tackled in future work. We have added the following paragraph in the discussion section to address this point:

“In addition, one could argue that a second limitation of the present work is related to the specific timing of the study, which took place during the French presidential election. This specific timing was chosen on purpose for two reasons. One reason was to benefit from the increased interest towards political topics at this time. The other reason was to probe voters’ redistributive behavior at a time that constitutes an important step in the democratic process. However, we acknowledge that it is possible that voters’ behavior in our study is unusual, because of this unusual timing. Voters may receive more information in the context of an election, and they may react more strongly to information delivered in this context. Whether our results would generalize to another context unrelated to a particular election thus remains an open empirical issue.”

Second, there are not enough information about groups characteristics. Are participants in the overachiever or underachiever groups different in their characteristics? Did the exclusions of some observations affect the balancing properties of the sample? How much people voting Hamon are different from those voting Macron, for example? These are very important information to understand the validity of the final results.

We thank the reviewer for this comment. We have clarified in the discussion that group characteristics did not differ between Mélenchon voters and Hamon voters (Macron voters were older and more male on average), and between overachievers and underachievers.

“Explanations of this finding in terms of age, sex, or socio-economic status are unlikely in our dataset as Mélenchon voters and Hamon voters did not differ significantly on these variables (Table 3). In addition, we verified that our Status manipulation was truly random with respect to age, sex, or socio-economic status, which did not differ between overachievers and underachievers.”

In addition, we have now added Table 3 to describe these variables for the 3 group of voters.

Table 3. Socio-demographic characteristics of participants reporting voting for Mélenchon, Hamon or Macron (in the first round of the election) in our dataset.

First-round vote Age (SD) Gender

(% women) Occupation

(% White Collar) Occupation

(% Student)

Mélenchon 31.21 (12.42) 0.62 0.29 0.45

Hamon 32.57 (14.59) 0.67 0.33 0.45

Macron 37.22 (16.75) 0.45 0.47 0.32

Third, the findings may be biased by the fact that tasks are not incentivized as they would be in an ideal scenario. In particular, I worry that the lack of impact on most of the voters might be the consequence of the fact that the questions are asked hypothetically. A final question is raised as to whether one can believe that participants might have a desire to please who developed the study. However, there are not enough information on the structure of the survey, so it is very difficult to evaluate the validity of the research design.

Indeed, our measures are based on choices that were not incentivized. We have now added one paragraph in the discussion to highlight this difference between our study and the study of Deffains et al., and the role it may have played in our results.

“Our main result is that redistributive behavior is influenced by the exogenous manipulation of Status only in a subgroup of participants, specifically those who reported voting for Mélenchon. Therefore, our study partially replicated the findings of Deffains and colleagues (2016). This partial discrepancy between our study and that of Deffains might be due to incentives. In Deffains’ study, participants’ redistribution choices in the dictator game had real consequences on the payoffs of other players, whereas in our paradigm redistribution choices were only hypothetical. It has been proposed (e.g. Camerer & Hogarth, 1999) that in the absence of incentives, participants might try to please the experimenter or conform to some social norms, e.g. by being generous in dictator games. However, in our experiment, participants redistributed less than in Deffains’ study: our mean allocation to A was 60.08 while the corresponding value in Deffains’ study would be 57.56. Besides, it is not clear to us why this desirability bias would lead to the specific interaction between Status and First-round vote. Furthermore, since our experiment was conducted online and responses were anonymous, such an explanation in terms of desirability bias seems unlikely to us. Nevertheless, it is possible that incentives might have influenced our results independently of the desirability bias. Participants who reported voting for Hamon or Macron might be more sensitive to the presence of real life incentives than Mélenchon voters. Thus, incentivizing redistribution choices might be a necessary feature to obtain the effect of Status in Hamon or Macron voters, whereas Mélenchon voters would exhibit the effect of Status even in the absence of incentives. To evaluate these possibilities, further research would need to compare redistribution choices with and without incentives, for the different groups of voters.”

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

6 Dec 2019

PONE-D-19-26429R2

How symbolic success affects redistributive behavior? An experiment based on the 2017 French presidential election

PLOS ONE

Dear Dr. Berthet,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Jan 20 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Valerio Capraro

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

The referee still has some major comments. Please do your best to address their comments.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: (No Response)

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Referee’s report - Manuscript ID PONE-D-19-26429R2

Many thanks for the answers. However, I think that there is still something in need of clarification. In particular, my main concerns are about the exclusions of some observations and the balancing property of the sample. Please find below my comments to the previous answers.

--

REFEREE: First, the sample is heavily biased to left position and is far to be representative of the overall population. Although, I appreciate that authors are well aware about this limitation, I still think that this is a big issue. The worry is that one would expect the left-wing participants to possibly have quite different preferences from the general population. How much we can take from this work and extend to other groups or even other contexts?

AUTHOR/S: We are indeed aware of this limitation due to the bias in our sample, which was already acknowledged in the discussion. In the revised manuscript, we have made more explicit the possibility that right-wing or far-right voters could have exhibited a different behavior. Generalization to other populations and other contexts is always an issue in experimental studies. However, we believe that since it is an empirical issue, it can be tackled in future work.

REFEREE R2: Many thanks for your response. However, I`m still not sure about this point. The current title is a bit misleading. I would suggest you to make more explicit that your analysis focuses on left-wing voters. This also requires some additional effort in terms of framing the discussion. Sometimes, it seems that you are referring to the general population.

I also agree that representativity is always an issue in experimental studies. But, we run experiment to understand something more about preferences, behaviors etc. My main concern with this study refers to the fact that the sample is totally biased to the left and in particular to the “French left” during the time of the 2016 election … So, what can I learn from it? What in terms of value added to the existing literature? This something that should be explained better in the paper.

REFEREE: Second, there are not enough information about groups characteristics. Are participants in the overachiever or underachiever groups different in their characteristics? Did the exclusions of some observations affect the balancing properties of the sample? How much

people voting Hamon are different from those voting Macron, for example? These are very important information to understand the validity of the final results.

AUTHOR/S: We thank the reviewer for this comment. We have clarified in the discussion that group characteristics did not differ between Mélenchon voters and Hamon voters (Macron voters were older and more male on average), and between overachievers and underachievers.

REFEREE R2: I think that some answers are missing. I would suggest to address these points in the next version of the paper. Was the randomization before or after the exclusion of the observations? In my understanding, the randomization was before that some observations were excluded. Did the exclusion of these observations affect the balancing properties of the sample? It is always not good to exclude observations. If you do that, you should show that this will not affect your final results. My main concern is still: Are participants in the overachiever or underachiever groups different in their characteristics? author/s should show a table with these statistics and not only reporting two lines saying … “In addition, we verified that our Status manipulation was truly random with respect to age, sex, or socio-economic status, which did not differ between overachievers and underachievers.”

REFEREE: Third, the findings may be biased by the fact that tasks are not incentivized as they would be in an ideal scenario. In particular, I worry that the lack of impact on most of the voters might be the consequence of the fact that the questions are asked hypothetically. A final question is raised as to whether one can believe that participants might have a desire to please who developed the study. However, there are not enough information on the structure of the survey, so it is very difficult to evaluate the validity of the research design.

AUTHOR/S: Indeed, our measures are based on choices that were not incentivized. We have now added one paragraph in the discussion to highlight this difference between our study and the study of Deffains et al., and the role it may have played in our results.

REFEREE R2: I`m fine with the discussion on the incentives but, again, the author/s did not reply to my final comment (or provide more information) about the potential desire of participants to please who developed the study.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

10 Dec 2019

Reviewer #1: Referee’s report - Manuscript ID PONE-D-19-26429R2

Many thanks for the answers. However, I think that there is still something in need of clarification. In particular, my main concerns are about the exclusions of some observations and the balancing property of the sample. Please find below my comments to the previous answers.

REFEREE: First, the sample is heavily biased to left position and is far to be representative of the overall population. Although, I appreciate that authors are well aware about this limitation, I still think that this is a big issue. The worry is that one would expect the left-wing participants to possibly have quite different preferences from the general population. How much we can take from this work and extend to other groups or even other contexts?

AUTHOR/S: We are indeed aware of this limitation due to the bias in our sample, which was already acknowledged in the discussion. In the revised manuscript, we have made more explicit the possibility that right-wing or far-right voters could have exhibited a different behavior. Generalization to other populations and other contexts is always an issue in experimental studies. However, we believe that since it is an empirical issue, it can be tackled in future work.

REFEREE R2: Many thanks for your response. However, I`m still not sure about this point. The current title is a bit misleading. I would suggest you to make more explicit that your analysis focuses on left-wing voters. This also requires some additional effort in terms of framing the discussion. Sometimes, it seems that you are referring to the general population.

I also agree that representativity is always an issue in experimental studies. But, we run experiment to understand something more about preferences, behaviors etc. My main concern with this study refers to the fact that the sample is totally biased to the left and in particular to the “French left” during the time of the 2016 election … So, what can I learn from it? What in terms of value added to the existing literature? This something that should be explained better in the paper.

AUTHOR/S R2: We changed the title, and made our focus on left-wing voters more explicit in the abstract and in the main text.

In terms of representativity / general interest of our results, we now mention in the discussion that:

“We note that although our results may not be representative of right-wing voters, one could envision that they would generalize for left-wing voters in some other countries. Indeed, in the last decade, western democracies have seen a polarization of opinions, with a crisis of the traditional parties and a rise of support for extreme populist parties. Examples of populist far-left parties are die Linke in Germany, Podemos in Spain, Siriza in Greece, La France Insoumise (Jean-Luc Mélenchon’s party) in France. According to Rooduijn and Akkerman (2017) these radical left parties have in common that “They do not focus on the ‘proletariat’, but glorify a more general category: the ‘good people’” , contrary to former communist parties and that ”they do not reject the system of liberal democracy as such, but only criticize the political and/or economic elites within that system”. Our results regarding the susceptibility of Mélenchon voters to status manipulation could thus be evaluated and replicated in other countries.”

Regarding how our study contributes to the existing literature, we now highlight that our general finding that Mélenchon voters displayed a different behavior than Hamon and Macron voters connects with recent findings showing that supporters of extreme political groups have different characteristics from those with more moderate views (e.g. Rooduijn, 2018; Hanel et al., 2019). We have re-written the last paragraph in the discussion section to address this point:

“To conclude, our findings revealed that self-reported far-left voters turned out to be the more sensitive to the exogenous manipulation of symbolic success. This leads to three remarks. Firstly, we need further research to better understand to what extent, and in which groups, redistributive behavior can be manipulated through exogenous manipulations of experience of success. In particular, further studies are needed that shall use a proper manipulation of symbolic success and representative samples in terms of political and socio-economic features. Secondly, our findings suggest that the various political groups process information differently, that is, they are not cognitively homogeneous (e.g., Amodio, Jost, Master, Yee, 2007; Rollwage, Dolan, & Fleming, 2018; Clarkson et al., 2015). Finally, and more broadly, the fact that Mélenchon voters displayed a different behavior than Hamon and Macron voters extends recent findings showing that supporters of extreme political groups have different characteristics from those with more moderate views, although they are not necessarily different on socio-demographic variables such as age or level of education (e.g. Rooduijn, 2018). For instance, Hanel, Zarzeczna, and Haddock (2019) reported that extreme (left-wing or right-wing) supporters are usually more heterogeneous than moderate ones in terms of human values and politics-related variables such as attitudes toward immigrants and trust in institutions. In the current social and political context, we believe that understanding further these differences, especially whether some groups are more susceptible to influence than others, appears a worthwhile subject for future research. Using controlled experiments during political elections can be a useful tool in such research.”

The added references are:

Rooduijn, M., & Akkerman, T. (2017). Flank attacks: Populism and left-right radicalism in Western Europe. Party Politics, 23(3), 193-204.

Rooduijn, M. (2018). What unites the voter bases of populist parties? Comparing the electorates of 15 populist parties. European Political Science Review, 10(3), 351-368.

Hanel, P. H. P., Zarzeczna, N., & Haddock, G. (2019). Sharing the same political ideology yet endorsing different values: left- and right-wing political supporters are more heterogeneous than moderates. Social Psychological and Personality Science, 10, 874-882.

REFEREE: Second, there are not enough information about groups characteristics. Are participants in the overachiever or underachiever groups different in their characteristics? Did the exclusions of some observations affect the balancing properties of the sample? How much people voting Hamon are different from those voting Macron, for example? These are very important information to understand the validity of the final results.

AUTHOR/S: We thank the reviewer for this comment. We have clarified in the discussion that group characteristics did not differ between Mélenchon voters and Hamon voters (Macron voters were older and more male on average), and between overachievers and underachievers.

REFEREE R2: I think that some answers are missing. I would suggest to address these points in the next version of the paper. Was the randomization before or after the exclusion of the observations? In my understanding, the randomization was before that some observations were excluded. Did the exclusion of these observations affect the balancing properties of the sample? It is always not good to exclude observations. If you do that, you should show that this will not affect your final results. My main concern is still: Are participants in the overachiever or underachiever groups different in their characteristics? author/s should show a table with these statistics and not only reporting two lines saying … “In addition, we verified that our Status manipulation was truly random with respect to age, sex, or socio-economic status, which did not differ between overachievers and underachievers.”

AUTHOR/S R2: We have updated Table 3 and added Table 4 to provide more information about our groups as well as the statistics for the comparison between overachiever and underachiever. We have moved these two tables up in the main text so they are now Table 1 and Table 2.

Table 1

Socio-demographic characteristics of participants reporting voting for Mélenchon, Hamon or Macron in the first round of the election in our dataset. Participants who reported another vote are pooled together in this table, and were not analyzed further in the present study.

First-round vote N Age (SD) Gender

(% women) Occupation

(% White Collar) Occupation

(% Student) Status

(% overachiever)

Mélenchon 219 31.21 (12.42) 0.62 0.29 0.45 0.46

Hamon 121 32.57 (14.59) 0.67 0.33 0.45 0.57

Macron 166 37.22 (16.75) 0.45 0.47 0.32 0.47

Other 120 32.40 (14.65) 0.46 0.23 0.46 0.43

Table 2

Comparison of the socio-demographic characteristics of participants in the overachiever and underachiever conditions in our final sample (i.e. including only participants who reported voting for Macron, Hamon or Mélenchon).

Condition N Age (SD) Gender

(% women) Occupation

(% White Collar) Occupation

(% Student)

Overachiever 248 33.73 (15.34) 0.54 0.38 0.41

Underachiever 258 33.28 (14.06) 0.60 0.34 0.40

T 0.344 -1.371 0.888 0.187

p 0.731 0.171 0.375 0.852

REFEREE: Third, the findings may be biased by the fact that tasks are not incentivized as they would be in an ideal scenario. In particular, I worry that the lack of impact on most of the voters might be the consequence of the fact that the questions are asked hypothetically. A final question is raised as to whether one can believe that participants might have a desire to please who developed the study. However, there are not enough information on the structure of the survey, so it is very difficult to evaluate the validity of the research design.

AUTHOR/S: Indeed, our measures are based on choices that were not incentivized. We have now added one paragraph in the discussion to highlight this difference between our study and the study of Deffains et al., and the role it may have played in our results.

REFEREE R2: I`m fine with the discussion on the incentives but, again, the author/s did not reply to my final comment (or provide more information) about the potential desire of participants to please who developed the study.

AUTHOR/S R2: Our paragraph also addressed the issue of a potential desirability bias. Specifically, in our previous revision we wrote: “It has been proposed (e.g. Camerer & Hogarth, 1999) that in the absence of incentives, participants might try to please the experimenter or conform to some social norms, e.g. by being generous in dictator games. However, in our experiment, participants redistributed less than in Deffains’ study: our mean allocation to A was 60.08 while the corresponding value in Deffains’ study would be 57.56. Besides, it is not clear to us why this desirability bias would lead to the specific interaction between Status and First-round vote. Furthermore, since our experiment was conducted online and responses were anonymous, such an explanation in terms of desirability bias seems unlikely to us.”

We have now modified this passage to clarify this and better separate the discussion of incentives and the discussion of desirability bias. We now write in the discussion:

“It has been proposed (e.g. Camerer & Hogarth, 1999) that in the absence of incentives, participants might try to please the experimenter or conform to some social norms, e.g. by being generous in dictator games. Could this desirability bias explain our results or the difference between our study and Deffains’ study? We believe that such an explanation is unlikely for several reasons. First, if a desirability bias was more present our study than in Deffains’ study, then we should have observed more redistribution in our participants. However, in our experiment, participants redistributed less than in Deffains’ study: our mean allocation to A was 60.08 while the corresponding value in Deffains’ study would be 57.56. Second, and more generally, it is not clear to us why this desirability bias would lead to the specific interaction between Status and First-round vote. Third, the instructions given to participants (see Appendix A) did not refer to the aim of our experiment, so participants were naïve about our hypothesis. Had they tried to guess our expectations, we would have found an effect of status on fatalism, which we did not observe either in the full sample (p=.52) nor in Mélenchon voters (p=.35), whose redistributive behavior was affected by status however. Finally, our experiment was conducted online and responses were anonymous, so participants have no pressure to please the experimenter or conform to social norms.”

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

3 Jan 2020

PONE-D-19-26429R3

How does symbolic success affect redistribution in left-wing voters? A focus on the 2017 French presidential election

PLOS ONE

Dear Dr. Berthet,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Feb 17 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Valerio Capraro

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

The reviewer suggests minor revisions. Please address them at your earliest convenience. I am looking forward for the revision.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: (No Response)

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Referee’s report - Manuscript ID PONE-D-19-26429R3

Many thanks for these answers. As I said in my previous reports, I truly believe that this paper investigates a relevant question and I enjoyed reading it. Over this process, I believe that the paper has improved a lot.

Yet, there is still something in need of clarification.

As I said in my previous email, it is not good to exclude observations. Therefore, I wonder if it is possible to see (and report in the paper – maybe in the annex) the results of the regression including the observations excluded. Also, I would be happy to see if participants in the overachiever or underachiever groups are different not only in their socio – demographic characteristics but also on “Vote 1”.

Reading again the paper, I have also some additional suggestion/recommendation. I think that the paper would improve if all the results discussed in the paper are also summarized using tables. It would be great to have a table summarizing the results discussed between the line 191 and 193. By contrast, I think that the results reported in Table 4 could not be considered valid. Some of the controls might have been affected by the manipulation e.g. fatalism and views on income inequality. Therefore, the inclusion of these variables as controls may introduce some bias in the estimation. In spite of that, I would suggest to replicate the baseline estimation adding only gender and age as controls.

Finally, I think that the paper would benefit for a better explanation of the variable used in the analysis. How is the First-round vote variable used in the regression is defined after the exclusion of some observations?

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

8 Jan 2020

Reviewer #1: Referee’s report - Manuscript ID PONE-D-19-26429R3

REFEREE R3: Many thanks for these answers. As I said in my previous reports, I truly believe that this paper investigates a relevant question and I enjoyed reading it. Over this process, I believe that the paper has improved a lot. Yet, there is still something in need of clarification.

As I said in my previous email, it is not good to exclude observations. Therefore, I wonder if it is possible to see (and report in the paper – maybe in the annex) the results of the regression including the observations excluded.

AUTHOR/S R3: Regarding the exclusion of observations, we mentioned that few observations were excluded due to a technical error in the data collection. In the paragraph “Description of our sample”, we added the following sentence:

“(due to a technical error in the data collection, redistribution choices could only be analyzed for 626 participants).”

We ran the regression including all observations and added the results in the Annex (Appendix A). Following the recommendation of the reviewer below, we ran the regression adding only gender and age as covariates. Moreover, as some categories of First-round vote have few observations (e.g. Arthaud, Poutou, Lassalle, Cheminade, Asselineau, Dupont-Aignan), observations were merged as follows: Arthaud and Poutou votes were merged with Mélenchon (N = 235); Lassalle votes were merged with Macron (N = 168); Cheminade votes were merged with Fillon (N = 36); Asselineau, Dupont-Aignan, and Le Pen votes were merged in the “FarRight” category (N = 19); Blank votes and abstentions were merged in the “NoVote” category (N = 47). We added the following sentence in the main text:

“For completeness, we also report in Appendix A the results of a regression over all participants, including those who reported a different First-round vote.”

Appendix A: ANOVA table for redistributive behavior in the disinterested dictator game including all groups of voters.

ηp 2 S.S. d.f. F p

Gender 0.0113 1074 1 7.006 0.008

Age 0.0039 367 1 2.396 0.122

Status 0.0006 57 1 0.373 0.542

Vote1 0.0683 6878 5 8.971 <0.001

Status:Vote1 0.0138 1313 5 1.713 0.130

Residuals 93848 612

REFEREE R3: Also, I would be happy to see if participants in the overachiever or underachiever groups are different not only in their socio – demographic characteristics but also on “Vote 1”.

AUTHOR/S: We verified that Vote 1 and Status were not associated, by running an ANOVA with Status as a dependent variable and Vote 1 as an independent variable (F(2,503)=2.06, p=.13). In addition, we conducted separate tests comparing the proportion of votes for Mélenchon, Hamon and Macron between overachievers and underachievers. These tests indicated no difference between groups. This information about Vote 1 is now included in the table (Table 2).

Condition N Age (SD) Gender

(% women) First-round vote

Mélenchon / Hamon / Macron

Overachiever 248 33.73 (15.34) 0.54 0.41 / 0.28 / 0.31

Underachiever 258 33.28 (14.06) 0.60 0.46 / 0.20 / 0.34

test t = 0.344 χ2 = 1.64 χ2 = 1.10 / 3.67 / 0.29

p 0.731 0.2 0.29 / 0.06 / 0.59

REFEREE R3: Reading again the paper, I have also some additional suggestion/recommendation. I think that the paper would improve if all the results discussed in the paper are also summarized using tables. It would be great to have a table summarizing the results discussed between the line 191 and 193.

AUTHOR/S: We replaced Figure 3 by a new table (Table 4) indicating means and standard errors of the amount of money allocated to A in the Disinterested Dictator Game as a function of First-round vote and Status.

Table 4

First-round vote

Status Mélenchon Hamon Macron

Overachiever 60.5 (1.28) 59.2 (1.30) 64.4 (1.58)

Underachiever 56.7 (0.91) 62.3 (1.86) 63.0 (1.33)

REFEREE R3: By contrast, I think that the results reported in Table 4 could not be considered valid. Some of the controls might have been affected by the manipulation e.g. fatalism and views on income inequality. Therefore, the inclusion of these variables as controls may introduce some bias in the estimation. In spite of that, I would suggest to replicate the baseline estimation adding only gender and age as controls.

AUTHOR/S: We followed the recommendation of the reviewer and we now report the regression analysis adding only gender and age as covariates (Table 5). The results were virtually unchanged.

Table 5

ANOVA table for redistributive behavior in the disinterested dictator game. The different factors included in the model are the effects of Gender, Age, Status in the experiment (Overachiever vs. Underachiever), First-round vote and the interaction between Status and First-round vote.

ηp 2 S.S. d.f. F p

Gender 0.0080 584 1 4.001 .046

Age 0.0051 373 1 2.553 .111

Status 0.0026 188 1 1.291 .256

Vote1 0.0239 1781 2 6.101 .002

Status:Vote1 0.0119 874 2 2.992 .051

Residuals 72701 498

We have re-written the paragraph accordingly:

“Noteworthy, to evaluate whether our results were robust to changes in model specification, we conducted a new regression analysis, in which we added gender and age as covariates (Table 5). This regression revealed that gender affected redistribution, with women redistributing more than men, replicating previous findings (e.g. Capraro, 2019; Corneo & Grüner, 2002). This analysis also indicated a main effect of the First-round vote, and confirmed the interaction between Status and First-round vote. When examining the effect of Status separately for the 3 groups of voters, again adding gender and age as covariates, we found that redistributive behavior was affected by Status only for Mélenchon voters (F(1, 215) = 5.54, p = .020, ηp 2 = .0251), but not for Hamon voters (F(1, 117) = 2.09, p = .151, ηp 2 = .0176) or Macron voters (F(1, 162) = 0.454, p = .502, ηp 2 = .0028) replicating our main finding. For completeness, we also report in Appendix A the results of a regression over all participants, including those who reported a different First-round vote.”

REFEREE R3: Finally, I think that the paper would benefit for a better explanation of the variable used in the analysis. How is the First-round vote variable used in the regression is defined after the exclusion of some observations?

AUTHOR/S:

In the paragraph “Description of our sample”, we have made more explicit the fact that in our analysis, the categories of the First-round vote variable are restricted to Mélenchon, Hamon, and Macron. We added the following sentence:

“Accordingly, in what follows the variable “First-round vote” is a categorical variable with 3 possible values, namely Mélenchon, Hamon, and Macron.”

Submitted filename: Response to reviewers.docx

Click here for additional data file.

30 Jan 2020

How does symbolic success affect redistribution in left-wing voters? A focus on the 2017 French presidential election

PONE-D-19-26429R4

Dear Dr. Berthet,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Valerio Capraro

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

6 Feb 2020

PONE-D-19-26429R4

How does symbolic success affect redistribution in left-wing voters? A focus on the 2017 French presidential election

Dear Dr. Berthet:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Valerio Capraro

Academic Editor

PLOS ONE