Nudging individuals' creativity using social labeling.

Simple instructions have been shown to robustly influence individual creativity, which is key to solve local problems. Building on social labeling theory, we examine the possibility of nudging individual's creativity using "creative" and "not creative" labels. Study 1 showed that subjects labeled as "creative" or "not creative" performed better in a creative task than unlabeled subjects and established the moderating effect of self-perceived creativity. Among subjects scoring low on self-perceived creativity, those labeled as "creative" performed better than those labeled as "not creative". Conversely, among subjects scoring high on self-perceived creativity, those labeled as "not creative" tend to perform better than those labeled as "creative". Study 2 and Study 3 further explored the psychological mechanisms at play in both cases: specifically, Study 2 showed that applying a "creative" label has the ability to increase creative self-efficacy through self-perceived creativity, whereas Study 3 demonstrated that applying a "not creative" label has the ability to increase individual creativity performance through a higher involvement in the creative task.

Introduction

Potential antecedents of collaborators’ creativity, i.e. their ability to produce ideas that are both original and useful [1] [2], include collaborator’s education level, learning orientation and job self-efficacy, supervisor’s support and expectations, job complexity and creativity requirements, a favorable organizational climate or shared knowledge of who knows what [3] [4] [5] [6] [7]. Still, these drivers of individual creativity require such deep organizational changes that exploring how managers can activate individual creativity in a more direct and frugal way is a central issue. In this perspective, understanding whether self-perception based techniques, such as social labeling [8], can indirectly sustain collaborators’ individual creativity contributes to our knowledge of essential psychological processes within organizations.

Social labeling is “a persuasion technique that consists in providing a person with a statement about his or her personality or values (i.e. social label) in an attempt to provoke behavior that is consistent with the label” [9]. Bem’s re-attribution mechanism [10] explains the persuasion at stake in social labeling. Although self-observation of an individual’s own behavior provides the clearest information to make self-attribution of dispositional properties [11], labels provided by others can also be informative about a person’s traits and values and lead to their re-attribution to the self [12]. Scholars have theorized that a labeled individual appears to internalize the values or personality traits associated with the label as representative of his or her basic self-perception, leading to changes in subsequent behaviors [13] [14] [15]. To date, however, no research has specifically interrogated whether social labeling may play a role in enhancing creative self-efficacy perceptions, hence boosting individual creativity. The question of whether labeling an individual as “creative” (or “not creative”) supports individual creativity remains open.

To explore the impact of labeling a collaborator as “creative” (or “not creative”) on individual creativity performance, we conducted three experiments that all began with a self-assessment creativity questionnaire to provide a credible basis for random labeling. A first study shows that participants labeled as “creative” or “not creative” people performed better in the creative task than those in the control condition. Going further, our results suggest a moderating effect of self-perceived creativity. Further interrogating these initial results, two follow-up studies investigated the mechanism at stake when assigning individuals to a “creative” or “not creative” label (versus no label). Study 2 shows that labeling individuals as “creative” enhances successively their self-perceived creativity and their creative self-efficacy, explaining why subjects scoring low on self-perceived creativity would perform better when labeled as “creative”. Study 3 suggests that labeling individuals as “not creative” activates psychological reactance and enhances subjects’ involvement in the creative task, hence individual creativity performance. These findings thus contribute to the literature on creativity management by offering relevant insights on new ways to nudge collaborators’ creativity.

Boosting individual creativity within organizations

Ideation is usually a demanding activity that requires time and cognitive efforts, while remaining a highly uncertain activity with a risk of failure. It is therefore paramount that individuals should have the confidence to remain persistent in creative ideation, thus developing creative self-efficacy, defined as the belief that one has the knowledge and skills to produce creative outcomes [6]. And it is the managers’ job to shape the organization's culture [16] likely to disseminate key assumptions, beliefs, and values regarding creativity, thus providing psychological safety [17] and impacting individual ideation. Direct extrinsic stimuli from managers to employees–such as managerial instructions to “be creative”–have been shown to have a concrete impact on performance in tests of creative thinking [18] [19] [20]. Specifically, they may lead to more creative ideas, but to a lower number of ideas [21], although sometimes generating psychological reactance [22]. Those results are not always replicated and remain questionable [23], calling to further research on the issue of indirect stimuli provided by managers to foster individual creativity.

Direct effect of the “creative” label

Social labeling techniques have proven effective to elicit a variety of desirable behaviors (e.g. eco-friendly consumerism, scoring better in math, signing a petition, giving to charity) from a collection of targets such as children, students or adults (e.g. [24] [25]). As such, they could play as relevant indirect stimuli provided by managers to enhance individual creativity performance.

In line with the literature, we contend that labeling an individual as “creative” could alter his or her self-perception, leading to a view of the self as creative, as soon as the label appears plausible. To build a plausible label, experimenters usually rely on people’s previous statements [12] [26] [27] or recent behavioral evidence [9] [28]. When the label does not appear as plausible, persuasion knowledge is activated and the label is rejected [29] [9]. As a conclusion, the persuasive intent behind the label must be masked for the self-perception to be altered among adults. Such a modification of self-perception is then likely to boost his or her creative self-efficacy. Framed in terms of psychological empowerment, general self-efficacy is found to influence the motivation and ability to engage in, and cope with, specific tasks [30]. Similarly, creative self-efficacy enhances the motivation and ability to engage in, and cope with, creative tasks, resulting in enhanced individual creativity performance, as proposed (e.g. [30] [31]) and consistently confirmed (e.g. [3] [5] [6] [7] [32]) by the literature.

Based on the above, we argue that the application of a “creative” label will be effective if the subjects do not perceive the persuasion attempt behind the label. We therefore hypothesize the following:

Hypothesis 1: Subjects labeled as “creative” display higher creativity performance than unlabeled subjects, i.e. a control group.

Direct effect of the “not creative” label

Social labeling effectiveness relies on the re-attribution to the self of the qualities stressed in the label, whether the label relates to positive or negative qualities [13] [8]. While re-attribution to the self of positive qualities consistently enhances positive behaviors (e.g. [9] [28]), re-attribution to the self of negative qualities has a less consistent influence on subsequent behaviors [12]. On the one hand, subjects could re-attribute a negative label to their negative qualities, and behave accordingly. Goldman and colleagues [14] showed that subjects labeled as unhelpful were less likely to comply with a request, suggesting that the negative label could have activated the perception of actually being a rather unhelpful person. This negative influence of negative labels is in line with the interactionists’ perspective on deviance, whose central contribution to labeling theory is that treating an individual as if he or she were deviant leads that person to share this perception of him/herself and preserves his or her deviant behavior (e.g. [33]). On the other hand, subjects could re-attribute a negative label to their negative qualities, and make efforts to disprove it. To explain this, research suggests that negative labels could make the individual more sensitive to the consequences of being associated with negative qualities, the way he or she presents himself, and the way he or she is perceived, hence his or her self-image [8]. Steele [34], for example, showed that subjects labeled as individualistic are more likely to give some time to a charity in order to restore their self-esteem.

Negative labels can thus lead to two contradictory outcomes: confirmation of the negative label vs. resistance to the negative label. The acceptance or rejection of the label may depend on the ability of the subsequent requested behavior to effectively contradict the label [35]. Further testing this explanation by examining congruence between the label and the requested behavior, Guéguen [25] confirmed that subjects are more likely to reject a negative label when the label is linked to the nature of the subsequent requested behavior, thus restoring their self-esteem. In contrast, when the requested behavior has no link with the basis for the negative label, subjects have no opportunity to behave in a way that directly contradicts the negative label, thus behave in a way that confirms the label. Following, asking individuals to participate in a creative task clearly appears as a relevant opportunity to contradict a “not creative” label and demonstrate creativity.

Hence, a “not creative” label is expected to erode collaborators’ self-esteem in terms of creativity, but should actually encourage them to invest efforts to display high creativity performance in a subsequent creative task in order to restore it. We therefore propose the following hypothesis:

Hypothesis 2: Subjects labeled as “not creative” display higher creativity performance than unlabeled subjects, i.e. a control group.

Hypotheses 1 and 2 propose that labeling a collaborator as “creative” or “not creative” will both result in an increase in individual creativity performance. These hypotheses are not mirror images of each other, being based on different psychological mechanisms, and are both expected to be supported.

Moderating effect of self-perceived creativity

The effectiveness of social labeling depends on the actual existence of a sufficient and clear self-perceptions in labeled subjects [24] [36] and on the label’s ability to modify self-perceptions [13] [14] [15]. Social labeling effectiveness should therefore depend on the way individuals initially perceive themselves as regards the targeted quality. Actually, a “positive” label should provide more room for modification among people scoring low on the targeted quality than higher scorers, and thus have more effect on the former than the latter group. A “negative” label, in contrast, should provide less room for modification among people scoring low on the targeted quality than higher scorers, and thus have more effect on the latter than the former group. Following this rationale, “creative” and “not creative” labels are likely to influence individual creativity performance differently depending on their ability to modify the way subjects initially perceive themselves in terms of personal creativity. A “creative” label is particularly likely to enhance self-perception, creative self-efficacy, and individual creativity performance, among subjects initially perceiving themselves as lowly creative. In contrast, a “not creative” label is particularly likely to erode self-perception, motivate self-esteem restoration and enhance individual creativity performance, among subjects initially perceiving themselves as highly creative.

We therefore expect that a “creative” label should be more effective among subjects scoring low on self-perceived creativity, and a “not creative” label among subjects scoring high on self-perceived creativity. Our final hypothesis is thus that self-perceived creativity moderates the influence of the positive or negative nature (“creative” vs “not creative”) of the label on individual creativity performance.

Hypothesis 3: The effectiveness of a “creative” vs. “not creative” label on subjects’ creativity is moderated by subjects’ self-perceived creativity, such that:

a “creative” label generates higher creativity performance among subjects scoring low on self-perceived creativity,

whereas a “not creative” label generates higher creativity performance among subjects scoring high on self-perceived creativity.

Fig 1 synthesizes the hypotheses of the conceptual model.

Fig 1

Conceptual model.

We will test our conceptual model in Study 1, before further exploring the psychological mechanisms at play when labeling an individual as “creative” in Study 2, i.e. the activation of creative self-efficacy, or when labeling an individual as “not creative” in Study 3, i.e. an enhanced involvement in the creative task. We will therefore test in two follow-up studies why both “creative” and “not creative” labels may enhance individual creativity performance.

Materials and methods

Study 1

Procedure

The HEC Montreal's Ethic Committee approved this study (approval # 2015–62). Consent was obtained by the participant going through the questionnaire—any participant willing to stop could just opt out of the online questionnaire. Besides data were analyzed anonymously.

Before engaging in the study per se, participants registered their age, gender and education level, as these variables are traditionally considered as potentially influencing individual creativity performance (e.g. [3] [37] [6]).

Then, to ensure effective labeling of participants as “creative” or “not creative”, we had to make our label credible, and so the participants were first asked to complete an initial questionnaire about their self-assessed ability to generate creative ideas. Participants were then randomly assigned to one of the four conditions: a control condition, a “creative” label, a “not creative” label, and a “moderately creative” label. The “moderately creative” label was introduced to rule out the potential effect on individual creativity performance of the mere application of a “creative” or “not creative” label, that could increase the salience or perceived importance of creativity behavior, and drive a person to act accordingly independently of the positive or negative nature of the label. Following, we expect that participants labeled as “moderately creative” will not have a higher creativity performance than unlabeled participants, i.e. a control group.

Concretely, three quarters of participants were shown their supposed positions on a graph in which: (1) the highest-scoring 10% of the population were given the “creative” label, (2) the lowest-scoring 10% were given the “not creative” label, and (3) the rest of the population were given the “moderately creative” label. Members of the three label conditions were told: “Here is a histogram of the distribution of individual creativity scores of a representative sample of the population, allowing you to visualize your own level of creativity. The answers you just gave place you in the shaded column.” In the control condition, participants were simply thanked for their answers and given no feedback regarding their responses to the creativity questionnaire. Fig 2 illustrates the experimental stimuli.

Fig 2

Experimental stimuli.

“Not creative label on the left. “Creative” label on the right.

Of note, we acknowledge that disguising what we were studying deceives subjects but appears necessary to preserve the internal validity of our results. Though it can be considered as a big issue by economists, most social sciences that have been carrying experiments for decades tend to consider it as a legitimate procedure [38].

After filling in the questionnaire and being labeled (or not), participants were asked to suggest a number of creative ideas for completing the “egg task”. This divergent thinking task is commonly used in team-building warm-ups by practitioners [39] or in teaching (e.g. [40]), and has also been used in recent years in experimental studies of creativity (e.g. [41] [42] [43]). The egg task is presented as follows: “Please suggest as many original solutions as possible to the following problem: Make sure that a hen's egg that is dropped from a height of 10 meters does not break.” All participants were given as long as they wanted to write down as many original solutions as they could, with a minimal answering time of 10 minutes. To conclude, we asked the participants who were labeled to recall the nature of the label they were provided with to check the validity of our manipulation. Disregarding the outliers (i.e. the 10% fastest and the 10% slowest respondents), completing the whole survey took between 12 and 22 minutes (mean = 15, s.d. = 2).

For ethical purposes, after completing the whole procedure, participants were informed in the following words that the label was purely randomly applied: “Your creativity score stated at the start of the survey was randomly assigned for the purpose of our research. It does not reflect your actual level of creativity at all.” As for Study 2 and Study 3, Study 1 experimental procedure was approved by one of the authors’ Institutional Review Board.

Measures

Before applying the label, we measured pre-label self-perceived creativity in line with existing literature (e.g. [44] [45] [46]), using 6‐items ranging from 1 "strongly disagree" to 7 "strongly agree" (see Table 1). This measure displays a good reliability (Cronbach’s α = .849).

Table 1

Pre-label self-perceived creativity items.

I like to discover new ways to consider a problem

I like to spend time going beyond the initial perception of a problem

I like to take the measure of a situation by considering it as a whole

I like working on vaguely-defined or emerging problems

I like to use my imagination to generate several ideas

I like to work on exceptional ideas

Individual creativity performance was measured in terms of originality, i.e. the ability to break free of the obvious and commonplace, and generate novel ideas and responses [47]. To measure each participant’s originality, we first used Agogué and colleagues [41] coding matrix. This matrix, which associates each possible answer to the egg task to a specific score on a 5-point scale ranging from 1 (“not at all original”) to 5 (“highly original”), was developed using Amabile’s [48] consensual assessment technique. Each participant was then assigned the Top 2 originality index proposed by Silvia and colleagues [49], which consists of the sum of the scores for his/her two most original responses. Following this approach, people are evaluated by the best level of performance they are able to achieve [2] on a constant number of responses. As such, the Top 2 originality index is a better measure of originality than an average score of originality that would penalize respondents who give a large number of uncreative responses. In the end, the participants’ individual creativity performance scores were between 1 and 8.8 (mean = 5.48, s.d. = 1.39).

Participants

200 subjects recruited from the online panel of a European professional market research institute provided an informed written consent for experimentation before participating in the study (mean age 41, 51% women). The participants were randomly assigned to one of the four conditions (50 participants per condition). Participants did not differ in terms of age (F(3,196) = .190, ns), gender (χ2 = 1.361, ns), education level (χ2 = 6.085, ns), or self-perceived creativity (F(3,196) = 1.058, ns) depending on the group they were assigned to.

Results

To examine whether the application of a social label as “creative” (or “not creative”) influences individual creativity performance, we first conducted an ANOVA considering the manipulation as a between-subjects factor. This ANOVA revealed a significant influence of the manipulation on individual creativity performance (F(3,196) = 2.927, p < .05, ηp2 = .043, ω = 0.79). Corroborating H1, the participants labeled as “creative” performed better in the creative task than those in the control condition (5.81 vs. 5.05, F(1,98) = 6.884, p < .01, ηp2 = .066, ω = 0.83). Also corroborating H2, the participants labeled as “not creative” performed better in the creative task than those in the control condition (5.63 vs. 5.05, F(1,98) = 4.138, p < .05, ηp2 = .041, ω = 0.65). Interestingly, labeled participants’ performance in the creative task did not differ according to the positive or negative nature of the label (F(1,98) = .398, ns). Also, as expected, the participants labeled as “moderately creative” did not perform differently in the creative task than unlabeled participants in the control condition (5.44 vs. 5.05, F(1,98) = 2.206, ns). Of note, we observe the same results when controlling for the participants’ age, gender, education level, and self-perceived creativity. As a relevant ancillary finding, the analysis also showed that self-perceived creativity does not influence individual creativity performance (F(1,195) = .026, ns). These results are plotted in Fig 3.

Fig 3

Results of Study 1.

Direct effects of the labels on individual creativity performance.

Going further, to test whether self-perceived creativity moderates the influence of a social label marking someone out as creative (or not creative) on individual creativity performance, we conducted a floodlight analysis using the procedure recommended by Cadario and Parguel [50]. Half of the participants were considered in this analysis (50 participants labeled as “creative” and 50 participants labeled as “not creative”). Hayes' [51] PROCESS macro (model 1) and 5000 bootstrapped samples were used to determine whether this moderating effect was significant. The manipulation (i.e. a “not creative” label vs. a “creative” label) was included as the independent variable, and individual creativity performance was included as the dependent variable. Though experimental research scholars often discretize quantitative variables when testing moderation, this practice has been challenged over the last 15 years [52], leading some top-tier journal editors to call for its “death” (e.g. [53]). Participants’ self-perceived creativity was thus included in the floodlight analysis as the moderating variable in its continuous form. This analysis revealed a significant interaction effect between the positive or negative nature of the label and participants’ self-perceived creativity (t93 = 2.23, p < .05) controlling for participants’ age, gender, and education level.

Further conditional analyses show that the “creative” (vs. “not creative”) label enhanced individual creativity performance (6.3 vs. 5.2, β = .46, t93 = 2.13, p < .05) among the 18 participants scoring relatively low on self-perceived creativity (score under 4 out of 7). It marginally reduced it (5.8 vs. 6.0, β = -.31, t93 = 1.38, p < .10) among the 18 participants scoring relatively high on self-perceived creativity (score over 6.33 out of 7). Corroborating H3a but only marginally H3b, these results are depicted in Fig 4.

Fig 4

Moderating effect of self-perceived creativity.

Study 2 and Study 3 extend these first findings on the impact of social labeling in creativity management by investigating the psychological mediating mechanisms at play when applicating a “creative” label (Study 2) or a “not creative” label (Study 3), versus no label.

Study 2

Our hypothesis about the effect of labels is based on the idea that labeling an individual alters his or her self-perception. Accordingly, labeling an individual as “creative” should lead to a view of his or her self as creative and result in boosting his or her creative self-efficacy. Fig 5 synthetizes this mediating mechanism.

Fig 5

Conceptual model for Study 2.

Mediating mechanism at play to explain the “creative” label effect.

We now explore the impact of labeling an individual as “creative” (vs. no label) on both his or her self-perceived creativity and creative self-efficacy.

Study 2 builds on Study 1 procedure. We first collected self-perceived creativity as a fake basis to ensure effective labeling of participants as “creative”, and participants had to register their age, gender and education level. Participants were then randomly assigned to one of the two conditions: a control condition vs. a “creative” label (as manipulated in Study 1). After being randomly exposed to one of the two conditions, they were interrogated about their current self-perceived creativity and about the creative self-efficacy they would feel if they were now confronted with a creative task. To conclude, we asked the participants who were labeled to recall the nature of the label they were provided with to check the validity of our manipulation, verified that no participant correctly guessed the purpose of the study and informed them that the label was purely randomly applied.

Pre-label self-perceived creativity was measured using the same 6 items-scale that was initially used in Study 1. To measure creative self-efficacy, we expanded Tierney and Farmer’s [6] (2002) scale asking respondents how much confident they would feel in their ability if they were “to participate in a creative exercise now” using 5 items ranging from 1 "not confident" to 7 "strongly confident": “ability to address successfully a creative task / Answer creatively to a creative task / Propose many solutions to a creative task / Find out-of-the-ordinary solutions to a creative task / Be original.” Finally, to measure post-label self-perceived creativity, we used a two-item general scale ranging from 1 "not at all " to 7 "very much": “Right now, you perceive yourself as a creative person / A person who addresses problems in an original way.”

Our measures proved to be reliable: Cronbach’s α of 894 for the items measuring pre-label self-perceived creativity, Cronbach’s α of 945 for the items measuring creative self-efficacy and correlation of .794 for the two items measuring post-label self-perceived creativity. Table 2 displays the correlation between all the constructs measuring creative self-perceptions for the two conditions.

Table 2

Correlations.

		Pre-label self-perceived creativity	Post-label self-perceived creativity	Post-label creative self-efficacy
No label	Pre-label self-perceived creativity	1	.514	.530
	Post-label self-perceived creativity		1	.733
“Creative” label	Pre-label self-perceived creativity	1	.624	.713
	Post-label self-perceived creativity		1	.822

102 subjects recruited from the online panel of a European professional market research institute provided an informed written consent for experimentation before participating in the study (mean age 37, 46% women). The participants were randomly assigned to one of the two conditions (48 participants in the control condition, 54 participants in the “creative” label condition after discarding participants who either didn’t fill the questionnaire entirely or couldn’t recall correctly how they were labeled). Participants did not differ in terms of age (F(1,101) = .153, ns), gender (χ2 = 2.387, ns), education level (χ2 = 1.023, ns), or self-perceived creativity pre-label (F(1,74) = .903, ns) depending on the group they were assigned to.

The aim of this second study is to examine whether the application of a social label as “creative” influences self-perceived creativity and creative self-efficacy. More precisely, Study 2 aims at testing whether self-perceived creativity mediates the influence of a “creative” label (vs. no label) on creative self-efficacy.

To do so, we used Hayes' [51] PROCESS macro (model 4) and 1000 bootstrapped samples. The analysis first shows that the application of the label significantly increases post-label self-perceived creativity (5.2 vs. 4.5 in the “no label” condition, β = .65, p < .01, ηp2 = .067, ω = 0.85). Following, post-label self-perceived creativity transfers to creative self-efficacy (5.5 vs. 4.9 in the “no label” condition, β = .78, p < .01). A bias-corrected 95% confidence interval for the indirect effect ranges from 0.13 to 0.91, with a point estimate of 0.51. Since the confidence interval does not include 0, this pattern provides evidence of an indirect effect on post-label self-perceived creativity: the “creative” label enhances self-perceived creativity, which–in turn–enhances creative self-efficacy. This mediation is indirect-only as the direct effect of the “creative” label on creative self-efficacy is not significant (β = .06, ns). Table 3 displays the results of this analysis.

Table 3

Mediation analysis.

	Post label self-perceived creativity	Creative self-efficacy
No label vs. “Creative” label	.65*	.06
Post-label Self-perceived creativity		.78*
Direct effect of the label on creative self-efficacy IC (95%)		[-.27;.40]
Indirect effect of the label on creative self-efficacy IC (95%)		[.13;.91]

* p < .01. Results do not change when controlling for participants’ pre-label self-perceived creativity.

To conclude, Study 2 shows that the application of a “creative” label has the ability to modify creative self-efficacy through self-perceived creativity, which is in line with social labeling theory. In other words, an individual who is labeled as being creative consequently has a higher perception of his or her creativity and therefore has a higher confidence in his or her capacities to perform creatively, explaining why subjects labeled as “creative” were found to display higher creativity performance than unlabeled subjects in Study 1.

Study 3

Our hypothesis about the effect of the “not creative” label is based on the idea that labeling an individual alters his or her self-perception. Accordingly, labeling an individual as “not creative” should both lead to a view of his or her self as not creative and encourage him or her to invest efforts to display high creativity performance in a subsequent creative task in order to restore it. Fig 6 synthetizes this mediating mechanism.

Fig 6

Conceptual model for Study 3.

Mediating mechanism at play to explain the “not creative” label effect.

To test these mediating mechanisms, we explore the impact of labeling an individual as “not creative” (vs. no label) on his or her self-perceived creativity, involvement in the creative task and individual creativity performance.

Study 3 builds on Study 1 and Study 2 procedure. We first collected self-perceived creativity as a fake basis to ensure effective labeling of participants as “not creative”, and participants had to register their age, gender and education level. Participants were then randomly assigned to one of the two conditions: a control condition vs. a “not creative” label. We then interrogated the participants about their current self-perceived creativity and invited them to participate in a creative task–to come-up with creative ideas for a brick (no minimal time). To conclude, we asked the participants who were labeled to recall the nature of the label they were provided with to check the validity of our manipulation, verified that no participant correctly guessed the purpose of the study and informed them that the label was purely randomly applied.

Pre-label self-perceived creativity was measured using the same 6 items-scale that was used in Study 1 and Study 2 (Cronbach’s α of 899). To measure post-label self-perceived creativity, we also used the same 2 items-scale that was used in Study 2 (correlation of .731). To measure involvement in the creative task, we monitored the time spent on the task. Finally, individual creativity performance was measured, as in Study 1, using Silvia and colleagues’s Top 2 originality index [50]. Its scores were between 0 and 14 (mean = 5.34, s.d. = 3.46).

Subjects recruited from the online panel of a European professional market research institute provided an informed written consent for experimentation before participating in the study (mean age 42, 58% women). They were randomly assigned to one of the two conditions. 180 participants were considered in the analyses (108 participants in the control condition, 72 participants in the “not creative” label condition after discarding participants who either didn’t fill the questionnaire entirely, didn’t produce any idea or couldn’t recall correctly how they were labeled). Participants did not differ in terms of age (F(1,178) = .420, ns), gender (χ2 = .034, ns), education level (χ2 = 6.323, ns), or self-perceived creativity pre-label (F(1,178) = .036, ns) depending on the group they were assigned to.

The aim of Study 3 is to examine whether self-perceived creativity and the involvement in the creative task mediate the influence of a “not creative” label on individual performance creativity. To do so, we used Hayes' [51] PROCESS macro (model 4) and 1000 bootstrapped samples. The analysis first shows that the application of the “not creative” label significantly reduces post-label self-perceived creativity (4.4 vs. 4.9 in the “no label” condition, β = -.59, p < .01, ηp2 = .059, ω = 0.92) and increases time spent on the task (138 vs. 93 seconds in the “no label” condition, β = 51.77, p < .05, ηp2 = .018, ω = 0.43). Following, both post-label self-perceived creativity (β = .34, p < .05) and time spent on the task (β = .0058, p < .01) transfer to individual performance creativity, which scores at 5.7 for the participants labeled as “not creative” vs. 5.1 for the unlabeled participants. A bias-corrected 90% confidence interval for the indirect effect ranges from -0.42 to -0.02, with a point estimate of -0.20 for post-label self-perceived creativity and from 0.04 to 0.56, with a point estimate of 0.30 for the time spent on the task. Since the confidence interval does not include 0, this pattern provides evidence of a marginally significant indirect effect on individual performance creativity: the “not creative” label reduces self-perceived creativity but enhances the involvement in the creative task, which–in turn–modifies individual performance creativity. This mediation is partial as the direct effect of the “not creative” label on individual performance creativity is also marginally significant (β = .76, p < .10). Table 4 displays the results of this analysis.

Table 4

Mediation analysis.

	Post label self-perceived creativity	Involvement in the creative task	Individual performance creativity
No label vs. “Not creative” label	-.59**	51,77*
Post-label Self-perceived creativity			.34*
Involvement in the creative task			.0058**
Direct effect of the label on individual performance creativity IC (90%)			[.02;1.51]
Indirect effect via post-label self-perceived creativity IC (90%)			[-.42;-.02]
Indirect effect via the involvement in the creative task IC (90%)			[.04;.56]

** p < .01

* p < .05. Of note, when controlling for participants’ pre-label self-perceived creativity, the indirect effect via post-label self-perceived creativity is no more significant.

To conclude, Study 3 shows that the application of a “not creative” label damages self-perceived creativity and enhances involvement in the creative task. As the second mechanism appears stronger than the first one, which disappears when controlling for pre-label self-perceived creativity, this could explain why, on the whole, subjects labeled as “not creative” displayed higher creativity performance than unlabeled subjects (β = .87, p < .05, ηp2 = .019, ω = 0.46), corroborating Study 1 findings.

Discussion

In this paper, we explore the relevance of social labeling techniques to foster individual creative behaviour. We show that labeling collaborators as “creative” or “not creative” has the potential to enhance their individual creativity performance, and that the effectiveness of labeling depends on the way individuals initially perceive themselves in terms of creativity. These findings make a number of relevant contributions for both academics and practioners involved in innovation management.

Theoretical contributions

This research contributes to the literature on social labeling. In showing the positive effect of a positive “creative” label on creative behaviors, it replicates previous findings observed in psychology and social marketing in an innovation management setting, thus extending the boundaries of social labeling’s relevance to the organizational field. In demonstrating the moderating effect of self-perceived creativity, it also offers a theoretical validation of social labeling theory itself, whose basic mechanism involves an alteration of individuals’ self-perception. Furthermore, in showing the positive effect of a negative “not creative” label, it corroborates the congruence hypothesis formulated by DeJong [35] and Guéguen [25], according to which a negative label only generates reactance when the label and the requested behavior are unconnected, such that the requested behavior offers no relevant opportunity to contradict the negative label. Additionnaly, Study 2 and Study 3 provide convincing empirical evidence about the mechanism at play behind the effect of each type of label.

Though demonstrating the significant positive effect of the “not creative” label on individual creativity performance, Study 1 and Study 3 also display specific marginal effects to be discussed. In Study 1, an insufficient sample size when testing the moderating influence of self-perceived creativity may be responsible for this result. Still, Study 3 has provided additional data showing that the application of a “not creative” label enhances subjects involvement in the creative task, but only among those scoring high on self-perceived creativity. This external evidence (not presented here to keep the paper consistent and reasonable in size) adds confidence in the moderating influence of self-perceived creativity in the case of the “not creative” label. In Study 3, the marginal effect we found applies to the mediation by time spent on the creative task. As behavioral effects are often more challenging to study, demonstrating the mediating role of subjects’ involvement in the influence of the “not creative” label on creative behavior already provides a relevant finding. In the end, the marginal effects we found in Study 1 and Study 3 do not undermine our general and pioneering contribution while still calling for further research to more finely understand social labeling techniques.

Beyond, this research also contributes to the literature on creativity. First, it shows the efficiency of operational devices using self-perception mechanisms to avoid forms of psychological reactance that could negatively affect individual creativity performance [22]. Positive labels, far from being perceived as an external influence, are received as the external recognition of the individual’s internal dispositions, and as such do not generate reactance. Second, this research sheds light on the psychological foundations of the Pygmalion process described by Tierney and Farmer [7], where creative self-efficacy mediates the effects of supervisor expectations on individual creativity performance. Third, this research also offers insights on using social labeling as a more frugal and less invasive way for managers to trigger creative self-efficacy than existing procedures (e.g. [6]). Social labeling techniques do not require managers to engage in any concrete creative process, nor to be creative themselves. This research thus contributes to the existing literature on the role of leadership for creativity, investigating minimal-resource strategies to increase individual creativity performance [54] [55]. It also calls for further research on the potential of nudges [56] in creativity management, since direct stimulation of creativity such as feedbacks from managers has been proven to have mixed results [45] [57].

Last, from the perspective of the innovation management literature, this research helps understand the “champion” phenomenon, in which individual intrapreneurs are recognized as developing unusual skills of innovation in their organizations [58] [59]. Such people may perceive themselves as highly creative, but are labeled as “not creative” by being assigned to operational teams: rather than losing self-esteem and work motivation and underperforming, they “overperform” to contradict the label [4] [60]. Labels already exist in most organizations. Large organizations, even industrial entities and institutions often implicitly convey positive stereotypes about specific members such as industrial designers [61], bioinformaticians [62] or web-designers and visual artists [63], and negative stereotypes about other members (e.g. buyers, lawyers or accountants). Establishing routines and processes highlights this separation between creativity and operational decision-making [64]. Further research should explore whether implicit “creative” and “not creative” labels have the same effects as explicit ones.

Managerial implications

First of all, our findings have implications for directly managing individual creativity in one-to-one settings between a manager and his or her employees. Creativity is a key factor in exploratory development of disruptive ideas [65] and innovation in highly competitive markets [66]. As such, the ability to boost employees’ creativity is more and more expected from managers. As social labeling appears as a frugal tool to influence behaviours, our findings suggest that such techniques can be used to foster creativity in situations where individual creativity is expected but does not benefit from much structural support (e.g. in small organizations such as start-ups or SMEs, or whenever an urgent creative answer is required, even in large companies).

Yet, social labels need to be credible to mask any persuasive intent and be effective [9]. This raises the question of how a “creative” or “not creative” label could be credible in an organization. Praising creative efforts, even if they are unsuccessful, has been proposed by Tierney and Farmer [7], but may lack credibility unless it is backed up by a record of previous achievements. Other sources of plausibility may lie in the reference to past achievements or in the identity of the label giver. Ideally, the label giver should be legitimate: this does not mean that he or she needs to display tremendous creativity, but that he or she has recognized skills in terms of creativity evaluation and selection. Organizational structure, and its implicit labeling system, could be perceived as more credible than any HR or communication department, which would not be recognized as a legitimate label giver. Being impersonal in nature, processes should be under less suspicion of persuasive intent than a manager, whose personal approaches can always be questioned.

Interpreting the paradoxical effects of positive and negative labels on creativity, several managerial recommendations finally result from our research. The first one is to remind operational employees–who may not perceive themselves as being creative in their occupation–that they too are creative, if innovative behaviors are to be expected subsequently. Low scorers on self-perceived creativity will then perform better in terms of creativity. On the other hand, for individuals having a high belief in their own creativity, psychological reactance is bounded when a negative label is applied, and our research suggests that the “not creative” label should be applied to them to elicit the desired creative behavior [25] [35]. One may object that such a managerial act may cause relation conflicts within a team. However, the study by Jung and Lee [67] showed that for invidiuals who value harmonious relationships in the workplace, conflictual relationalship may contrigger the perception of the problematic situation from a new standpoint and harness individual creative behavior to overcome the situation. This clearly calls to further explorate the role of negative labels on creativity, integrating moderating factors such as placing great value on creating hamonious relationships.

Limitations and further research

The present research is not without limitations. First, the nature of the label we used was quite simple: we used labels that relate to the level of creativity of individuals in a broad sense, by labeling individuals as “creative”, “moderately creative” or “not creative”. Building on the specific cognitive mechanisms inherent to creativity, such as associative thinking [68], further research could thus investigate the influence of more elaborate social labeling, specifically focusing on the individual ability to perform one particular cognitive mechanism. For instance, building on the very recent concept of idea linking that has been shown to stimulate creativity, labels such as “you are a person who is very good at using aspects of early ideas as input for subsequent ideas in a sequential manner” could provide fruitful insights on the managerial levers to boost individual creativity through social labeling.

Second, our experimental protocol was built on a label automatically provided by the computer to our participants, leaving a blind spot regarding the nature of the existing relationship between managers and employees in real-life settings. Indeed, leader-member exchange has been consistently shown to positively influence creativity [69], specifically when creative self-efficacy mechanisms are at stake [70], calling for further investigation regarding the moderating influence of leader-member relationships on the effectiveness of social labeling for creativity.

Third, we didn’t test the persistence of the different labels’ effects: we measured the influence of labels on self-perceived creativity, creative-self efficacy, and creativity performance just after applying a label but did not examine its lasting influence. As such, further research could investigate the long-lasting influence of social labeling on creativity, specifically in the case of negative labels that may, in the long-term term, influence work identity and task motivation.

Last, we should stress that we investigated only one part of the creative process, namely individual idea generation, calling to extend our findings both on the consequent phases of the creative process, namely idea evaluation and selection, and on more collective settings.

Data for Study 1.

(SAV)

Click here for additional data file.

Data for Study 2.

(SAV)

Click here for additional data file.

Data for Study 3.

(SAV)

Click here for additional data file.

24 Jul 2019

PONE-D-19-16350

Nudging individuals’ creativity using social labeling : an experimental study

PLOS ONE

Dear Dr Agogué,

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 the reviewers' comments below.

We would appreciate receiving your revised manuscript by Sep 07 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.

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

Additional Editor Comments (if provided):

I have now collected two reviews from two experts in the field. Both reviewers recommend major revisions. Therefore, I would like to invite you to revise the paper according to the reviewers' suggestions. Needless to say that all comments must be addressed. Particular attention, however, should be given to the issue raised by both reviewers regarding the smallness of the sample (especially in light of the current replicability crisis) and to the issue raised by Reviewer 1 regarding deception (please explain exactly which parts of the experiments involved deception and why you think this is not a problem for your experiments). Moreover, I would like to mention that social labels have been recently used also to impact people's decisions in economic games (https://www.sciencedirect.com/science/article/pii/S0022103116302098; https://www.sciencedirect.com/science/article/pii/S0022103118302841; http://journal.sjdm.org/17/171107/jdm171107.pdf; http://journal.sjdm.org/19/190107/jdm190107.pdf). Note that I am the author of some of these papers. My decision on this manuscript will obviously be independent of whether you will decide to include or not these papers in your reference list. I have just thought that you might find them relevant.

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. 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

Reviewer #2: Partly

**********

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

Reviewer #1: Yes

Reviewer #2: 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

Reviewer #2: 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

Reviewer #2: 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: Did you do a power test to determine whether your 76-subject study n was large enough to be statistically representative of the population? 76 people does not sound like a big sample.

Can you explain what alpha means in the following sentences?

Our measures display good reliability indicators: Cronbach’s α of 874 for the items measuring pre-label self-perceived creativity, Cronbach’s α of 941 for the items measuring creative self-efficacy and correlation of .803 for the two items measuring post-label self- perceived creativity.

I am not sure about PLOS-One, but I know that most experimental economics journals does not allow for deception. You clearly used deception in this experiment by randomly assigning people to the “creative” and “non-creative” groups. If PLOS-One is fine with deception, then this is not an issue. However, it is a big issue with experimental economists.

I would like to see demographic and other sample characteristics for the two samples of the control vs. treatment for each study. Are there any demographic differences? With a total n of 76, I would like to know if there are differences other than just the creativity score.

Again, with study 2, did you do a power test on your n=200 participants to see if this was a sufficient number of subjects?

Again, can you show the demographic characteristics of the control vs treatments for study 2?

Reviewer #2: This study explores whether social labels could influence creative behavior.

Major issues:

In the introduction, researchers suggested that “a labeled individual appears to internalize the values or personality traits associated with the label.” Later, authors argue that for negative labels there is a different mechanism “participants could re-attribute a negative label to their negative qualities, and make efforts to disprove it” As it stands now, it seems that labels influence behavior via different mechanisms, internalization, and resistance to the negative label. While it could be the case, neither study 1 nor study 2 directly measure internalization or resistance to internalize the label. Thus, overall, there is no compelling case that empirically demonstrates either of mechanisms.

I could argue that the internalization happens in both cases: positive label increases self-perception and individuals perform better; negative label decreases self-perception of creativity that makes individuals work harder. (you suggested similar mechanism explaining the moderation later in the paper) If this explanation is the case, there need to be a study that demonstrates this effect. Namely, negative labeling decreases self-perception and that leads to better performance, positive labeling increases self-perception and that leads to better performance than in the control condition.

- The section “Boosting individual creativity within organizations” does not seem to contribute to developing the story. Neither organizational structures nor supervisor advice is studied in this research. I suggest removing or reorganizing these sections in a way that includes explicit connections to the goals and measures of the study.

- The authors do not provide enough evidence that supports the following statement “the application of a “creative” label will be effective if the subjects do not perceive the persuasion attempt behind the label.” Why individuals who perceive the persuasion attempt behind the label will not be empowered by it? There is literature suggesting that if people know about persuasive effects of nudges they are still affected by them (e.g. Loewenstein G, Bryce C, Hagmann D, Rajpal S. Warning: You are about to be nudged.)

- It is possible that “a “creative” label enhances self-perception of creativity while a “not creative” label erodes self-perception. The moderation analysis highlights this possibility. However, to have evidence for this statement, it is important to incorporate the measures of self-perception before and after the manipulation to demonstrate that with a positive label the self-perception goes up, while with “negative label” self-perception goes down.

- Study 1

- Samples size in study one is 76 people which suggests that the study is probably underpowered. Did you do power calculation and a sample size calculation?

- There was randomization in this study, why age, gender, education level, and self-perceived creativity are added as controls? There was no difference in these measures between groups, therefore these controls should be removed from the analysis

- The goal of the study was to show that the label influences participants’ self-perception of creativity, which was achieved. Why there was an analysis of self-efficacy? If it highlights the mechanism of how labels influence performance, why there is no measure of self-efficacy in Study 2? I would not state in the discussion that this research contributes to self-efficacy literature as there is no association between self-efficacy and the study conditions or performance.

- Neither design of this study nor analysis speaks to this hypothesis. While this study could be a part of the research package, it does not directly speak to any of the stated hypothesizes. The measure of self-perception demonstrates rather that the manipulation worked as researchers planned it.

Study 2

- Where the participants' removal in this study? Did you have the same manipulation checks as in the previous one? Did participants remember the labels correctly?

- Covariates should not be included in the experimental study, especially that there were no differences in these variables between conditions.

- Please report degrees of freedom in moderation study, it seems underpowered, especially for the continuous measure.

- While moderation analysis demonstrates the possibility of the highlighted in the introductions relationships, the further mediation analysis is needed to provide further evidence (see my comments above)

- I don’t see the value in following Hayes process with ANCOVA, you could remove it from the study and visualize your observations directly from the Process Model 1 procedure.

OVERALL

- I recommend re-organizing the description of the mechanism/s via which positive and negative labels influence behavior.

- Recommend additional study that has exact measures to test the proposed mechanisms (see my comments above)

Minor issues

- While I can see that the work of Bem, 1972 might be related to social labeling, It would be great if authors could elaborate on the connection between self-perception theory and social labeling.

- I am not sure what you mean here “In facts, for the self-perception to be altered, the persuasive intent behind the label must be masked, explaining why experiments often rely on people’s previous statements…”

**********

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

Reviewer #2: 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.

1 Oct 2019

Nudging individuals’ creativity using social labeling: an experimental study

Manuscript: PONE-D-19-16350

Reviewer #1

Dear Reviewer #1,

We sincerely thank you for your very detailed and relevant comments on our manuscript.

To facilitate your review of our answers and changes, we discuss your comments and explain how we have integrated each of them into this new version of our manuscript directly below each comment. Your original comments are reproduced in bold; our responses follow each comment; the modified text appearing in the new version of our manuscript is copied in blue.

Did you do a power test to determine whether your 76-subject study sample was large enough to be statistically representative of the population? 76 people does not sound like a big sample.

We thank you for this comment. Originally educated in psychology, we more often use the criterium of the effect size to evaluate whether results are important enough. However, we do acknowledge that the power of the test appears today as a more robust criterium.

To better satisfy the power criterium, we collected more data to complete Study 1 sample (please note that we now present Study 1 as Study 2). Precisely, we asked our market research institute to propose our study to 30 more panelists. These panelists were exposed to the “creative” label condition, which was previously underrepresented (n=28 versus n=48 for the control condition). Out of these 30 more respondents, 26 checked the manipulation regarding the label and were included in the database we provide along with our new submission (these new respondents are identified as 101, 102… 126 in the database). Following, the section dedicated to the presentation of the results of the study has been fully rewritten in our new submission to consider our data collection extension. Precisely, it shows that our results stay the same but that the addition of respondents allowed an increase in the power of our test, which reaches now 84,5% as now explicitly appears in the manuscript:

To do so, we used Hayes' (2012) PROCESS macro (model 4) and 1000 bootstrapped samples. The analysis first shows that the application of the label significantly influences post-label self-perceived creativity (�=.65, p<.01, ηp2=.067, ω=0.85).

Can you explain what alpha means in the following sentences? “Our measures display good reliability indicators: Cronbach’s α of 874 for the items measuring pre-label self-perceived creativity, Cronbach’s α of 941 for the items measuring creative self-efficacy and correlation of .803 for the two items measuring post-label self-perceived creativity.”

The Cronbach’s alpha measures internal consistency, i.e. how closely related a set of items are as a group. It is a measure of scale reliability. As the Cronbach’s alpha is quite a classical indicator in statistical test theory, we do not propose to add more explanation about what it means in the manuscript. If we did not understand your comment properly, please be more specific, we will be more than happy to answer in a more appropriate way.

I am not sure about PLOS-One, but I know that most experimental economics journals does not allow for deception. You clearly used deception in this experiment by randomly assigning people to the “creative” and “non-creative” groups. If PLOS-One is fine with deception, then this is not an issue. However, it is a big issue with experimental economists.

You are right mentioning that we designed our experiment as a deception experiment and also perfectly right mentioning that deception is a big issue in experimental economics. As mentioned by Tyler and Amodio in 2015 the Oxford Handbook of Experimental Economic Methodology, “there is no subject that provokes so much emotion among economists as the use of deception in an experiment”.

Tyler, T. R., & Amodio, D. M. (2015). Psychology and economics: areas of convergence and difference. Handbook of experimental economic methodology. Oxford University Press, Oxford, 181-196.

Still, deception experiments are more than frequent in psychology experiments and have been regarded as legitimate by many social sciences that have been using experiments for decades.

Deception experiments are not used by psychologist for amoral reasons, but to ensure internal validity. Precisely, our manipulation allows to draw conclusion that could not be drawn otherwise. We can not imagine that it would have been possible to tell our subjects, as an introduction of the experiment, that we were about to put them in different fake conditions regarding their personal creativity. Such a disclosure would clearly have modified the way our subjects would have presented themselves and the way they would have responded to our label.

Tyler and Amodio (2015) identify a specific problem with deception experiments as they can “spoil the subject pool” for subsequent experiments. Though this argument could be relevant when carrying experiments using internal student samples, it is less relevant when considering the external sample of a market research institute. As a matter of fact, we have been using the same market research institute for years and always ask our subjects what they thought was the actual objective of the study they had just participated in and none of them (in all our data collections) correctly guessed it. Following, we are pretty confident on the fact that deception experiments conducted on very large pool of respondents and not on very restricted local pool do not “spoil the sample”.

Another problem is more moral and regards the fact to put subjects in scenario that make them believe for few minutes that they are not creative. To circumvent this problem, we used a clear debriefing statement at the end of the experiment to make clear that all subjects are. As an addition, we would also like to mention that all subjects participating in our studies gave an active consent before participating and that our protocols were all validated by one of the researchers’ IRB.

A written in the first version of our manuscript:

For ethical purposes, after completing the whole procedure, participants were informed in the following words that the label was purely randomly applied: “Your creativity score stated at the start of the survey was randomly assigned for the purpose of our research. It does not reflect your actual level of creativity at all.” As for Study 2 and Study 3, Study 1 experimental procedure was approved by the first author’s Institutional Review Board.

In the new version, we add a specific comment on deception (see the sentence in bold characters in the previous paragraph) to fully address your comment.

Of note, we acknowledge that disguising what we were studying deceives subjects but appears necessary to preserve the internal validity of our results. Though it can be considered as a big issue by economists, most social sciences that have been carrying experiments for decades tend to consider it as a legitimate procedure (Tyler & Amodio, 2015).

I would like to see demographic and other sample characteristics for the two samples of the control vs. treatment for each study. Are there any demographic differences? With a total n of 76, I would like to know if there are differences other than just the creativity score.

We provide statistics about demographic differences as follows in Study 1:

102 subjects recruited from the online panel of a European professional market research institute provided an informed consent for experimentation before participating in the study (mean age 37, 46% women). The participants were randomly assigned to one of the two conditions (48 participants in the control condition, 54 participants in the “creative” label condition after discarding participants who either didn’t fill the questionnaire entirely or couldn’t recall correctly how they were labeled). Participants did not differ in terms of age (F(1,101)=.153, ns), gender (χ²=2.387, ns), education level (χ²=1.023, ns), or self-perceived creativity pre-label (F(1,74)=.903, ns) depending on the group they were assigned to.

Again, with study 2, did you do a power test on your n=200 participants to see if this was a sufficient number of subjects?

For Study 2 (now Study 1), the tests we carried show that the size of our effects are between medium and large. The power of our tests mostly appears over .70 and for H1 over .80, as detailed in the new version of our manuscript:

To examine whether the application of a social label as “creative” (or “not creative”) influences individual creativity performance, we first conducted an ANOVA considering the manipulation as a between-subjects factor. This ANOVA revealed a significant influence of the manipulation on individual creativity performance (F(3,196)=2.927, p<.05, ηp2=.043, ω=0.79). Corroborating H1, the participants labeled as “creative” performed better in the creative task than those in the control condition (5.81 vs. 5.05, F(1,98)=6.884, p<.01, ηp2=.066, ω=0.83). Also corroborating H2, the participants labeled as “not creative” performed better in the creative task than those in the control condition (5.63 vs. 5.05, F(1,98)=4.138, p<.05, ηp2=.041, ω=0.65).

We also found a power over .70 for our floodlight analysis as shown hereafter:

To make this interaction effect easier to communicate (Fitzsimons, 2008), we dichotomized participants between higher and lower scorers on pre-label self-perceived creativity, using a median-split. We then carried out a classic ANOVA considering the manipulation (i.e. a “not creative” label vs. a “creative” label) and the dichotomized version of pre-label self-perceived creativity as between-subjects factors. This ANOVA replicated the significant interaction effect previously identified by the floodlight analysis between the positive/negative nature of the label and participants’ self-perceived creativity (F(1,96)=4.898, p<.05, ηp2=.049, ω=0.71). Planned contrast tests also revealed that the “creative” label enhanced individual creativity performance among participants scoring low on self-perceived creativity, who outperformed the “not creative” label group (6.18 vs. 5.45, F(1,49)=6.417, p<.01, ηp2=.116, ω=0.80). However, among participants scoring high on self-perceived creativity, the “not creative” group outperformed the “creative” label group, though in a marginally significant way (5.90 vs. 5.41, F(1,41)=2.387, p<.10, ηp2=.055, ω=0.45). Corroborating H3a but only marginally H3b, these results are depicted in Figure 4.

Again, can you show the demographic characteristics of the control vs treatments for study 2?

We provide statistics about demographic differences as follows in Study 2:

200 subjects recruited from the online panel of a European professional market research institute provided an informed consent for experimentation before participating in the study (mean age 41, 51% women). The participants were randomly assigned to one of the four conditions (50 participants per condition). Participants did not differ in terms of age (F(3,196)=.190, ns), gender (χ²=1.361, ns), education level (χ²=6.085, ns), or self-perceived creativity (F(3,196)=1.058, ns) depending on the group they were assigned to.

As we added a study 3, we provide statistics about demographic differences as well for this third story.

Reviewer #2

Dear Reviewer #2,

We sincerely thank you for your very detailed and relevant comments on our manuscript.

To facilitate your review of our answers and changes, we discuss your comments and explain how we have integrated each of them into this new version of our manuscript directly below each comment. Your original comments are reproduced in bold; our responses follow each comment; the modified text appearing in the new version of our manuscript is copied in blue.

In the introduction, researchers suggested that “a labeled individual appears to internalize the values or personality traits associated with the label.” Later, authors argue that for negative labels there is a different mechanism “participants could re-attribute a negative label to their negative qualities, and make efforts to disprove it” As it stands now, it seems that labels influence behavior via different mechanisms, internalization, and resistance to the negative label. While it could be the case, neither study 1 nor study 2 directly measure internalization or resistance to internalize the label. Thus, overall, there is no compelling case that empirically demonstrates either of mechanisms. I could argue that the internalization happens in both cases: positive label increases self-perception and individuals perform better; negative label decreases self-perception of creativity that makes individuals work harder. (you suggested similar mechanism explaining the moderation later in the paper) If this explanation is the case, there need to be a study that demonstrates this effect. Namely, negative labeling decreases self-perception and that leads to better performance, positive labeling increases self-perception and that leads to better performance than in the control condition.

We sincerely thank you for this comment that appears perfectly legitimate and invited us to enhance our argument and the quality of our manuscript.

In the new version that we are submitting, you will find a whole new data collection, which focuses on the influence of the “not creative” label and demonstrates that when labeled as “not creative”, subjects scoring high on self-perceived creativity display a lower perception of their creativity and an increase in their involvement in the creative task, as measured by the time spent on the task. Please see below:

Study 3

Our hypothesis about the effect of the “not creative” label is based on the idea that labeling an individual alters his or her self-perception. Accordingly, labeling an individual as “not creative” should lead to a view of his or her self as not creative and encourage him or her to invest efforts to display high creativity performance in a subsequent creative task in order to restore it. To test this mediating mechanism, we explore the impact of labeling an individual as “not creative” (vs. no label) on both his or her self-perceived creativity and involvement in the creative task.

Procedure

Study 3 builds on Study 1 and Study 2 procedure. We first collected self-perceived creativity as a fake basis to ensure effective labeling of participants as “not creative”, and participants had to register their age, gender and education level. Participants were then randomly assigned to one of the two conditions: a control condition vs. a “not creative” label. We then interrogated the participants about their current self-perceived creativity and invited them to participate in a fake creative task. To conclude, we asked the participants who were labeled to recall the nature of the label they were provided with to check the validity of our manipulation, verified that no participant correctly guessed the purpose of the study and informed them that the label was purely randomly applied.

Measures

Pre-label self-perceived creativity was measured using the same 6 items-scale that was used in Study 1 and Study 2 (Cronbach’s α of 899). To measure post-label self-perceived creativity, we also used the same 2 items-scale that was used in Study 2 (correlation of .731). Finally, to measure involvement in the creative task, we monitored the time spent on the task.

Participants

202 subjects recruited from the online panel of a European professional market research institute provided an informed consent for experimentation before participating in the study (mean age 42, 58% women). The participants were randomly assigned to one of the two conditions (120 participants in the control condition, 82 participants in the “not creative” label condition after discarding participants who either didn’t fill the questionnaire entirely or couldn’t recall correctly how they were labeled). Participants did not differ in terms of age (F(1,200)=.620, ns), gender (χ²=.021, ns), education level (χ²=5.315, ns), or self-perceived creativity pre-label (F(1,200)=.007, ns) depending on the group they were assigned to.

Results

The aim of Study 3 is to examine whether self-perceived creativity mediates the influence of a “not creative” label on the involvement in the creative task. To do so, we used Hayes' (2012) PROCESS macro (model 4) and 1000 bootstrapped samples but did not find any significant mediation between the construct. However, and as expected, the analysis shows that the application of the “not creative” label (vs. no label) both significantly damages post-label self-perceived creativity (F=12.19, p<.01, ηp2=.057, ω=0.97) and increases the time spent on the task (F=3.03, p<.05, ηp2=.015, ω=0.54).

Going further, to test whether the influence of the “not creative” label (vs. no label) on participants’ involvement in the creative task could depend on pre-label self-perceived creativity, as suggested by findings from Study 1, we conducted a floodlight analysis. Hayes' (2012) PROCESS macro (model 1) and 5000 bootstrapped samples were used to determine whether this moderating effect was significant. The manipulation (i.e. a “not creative” label vs. no label) was included as the independent variable, the time spent on the task as the dependent variable, and participants’ self-perceived creativity in its continuous form as the moderating variable. This analysis revealed a significant interaction effect between the manipulation and participants’ self-perceived creativity (t=1.98, p<.05).

As in Study 1, to make this interaction effect easier to communicate, we dichotomized participants between higher and lower scorers on pre-label self-perceived creativity, using a split at the Johnson-Neyman point (i.e. over 5.50). We then carried out a classic ANOVA considering the manipulation (i.e. a “not creative” label vs. no label) and the dichotomized version of pre-label self-perceived creativity as between-subjects factors. This ANOVA replicated the significant interaction effect previously identified by the floodlight analysis (F(1,198)=4.845, p<.05, ηp2=.024, ω=0.71). Planned contrast tests then revealed that, among participants scoring high on self-perceived creativity, the “not creative” label (vs. no label) significantly enhanced their involvement in the creative task (208’ vs. 97’, F(1,78)=3.804, p<.05, ηp2=.047, ω=0.61). No such influence appeared among participants scoring low on self-perceived creativity (88’ vs. 91’, F(1,120)=.017, ns).

To conclude, Study 3 shows that the application of a “not creative” label damages self-perceived creativity of subjects scoring high on self-perceived creativity and enhances their involvement in the creative task. Among subjects scoring low on self-perceived creativity, it only damages self-perceived creativity. These results could explain why subjects labeled as “not creative” were found to be likely to display higher creativity performance than unlabeled subjects in Study 1.

The section “Boosting individual creativity within organizations” does not seem to contribute to developing the story. Neither organizational structures nor supervisor advice is studied in this research. I suggest removing or reorganizing these sections in a way that includes explicit connections to the goals and measures of the study.

We agree with this observation. We have reframed and tightened the literature section by focusing specifically and direction and indirect stimuli that managers can use to trigger creative behavior.

The authors do not provide enough evidence that supports the following statement “the application of a “creative” label will be effective if the subjects do not perceive the persuasion attempt behind the label.” Why individuals who perceive the persuasion attempt behind the label will not be empowered by it? There is literature suggesting that if people know about persuasive effects of nudges they are still affected by them (e.g. Loewenstein G, Bryce C, Hagmann D, Rajpal S. Warning: You are about to be nudged.)

I am not sure what you mean here “In facts, for the self-perception to be altered, the persuasive intent behind the label must be masked, explaining why experiments often rely on people’s previous statements…”

Thank you for this comment. We really enjoyed the reading of the article you mention about the effects of disclosing the presence of default options and discussed a lot about it.

In the end, we retain the following conclusion made by Loewenstein and colleagues (2015): “our findings demonstrate that default options are a category of nudges that can have an effect even when people are aware that they are in play” (p. 40). It is actually interesting to consider that different categories of nudges could have different effects when they are disclosed. As far as we are concerned in the present manuscript, our nudge is based on self-perception theory and implies deception, which is quite common and legitimate in experimental psychology. Following, disclosing our manipulation would have meant to mention right from the beginning to our subjects that we were about to lie to them when labeling them as “x” or “Y”. In such circumstances, we sincerely do not think that such labeling could have altered their self-perception.

To go further, the literature developed on social labeling theory posits that for the self-concept to be altered among adults, the label has to be perceived as plausible. A label is not effective if it is perceived as contradictory to what people think about themselves (Tybout & Yalch, 1980). To avoid such perceptions, experimenters rely on people's actual behaviours or statements about behaviours to build a plausible label. As an illustration, Van der Werff, Steg, and Keizer (2014) justified an eco-friendly (resp. eco-unfriendly) label by using statements related to the 8 most- (resp. 8 least) performed environmental behaviours identified in a pre-test. Interestingly, behaviours or statements about previous behaviours that help justify the label plausibility do not need to be motivated by the trait stressed in the label. In Cornelissen et al. (2007) study, although participants may have selected the most ecological television set for its quality and price, stressing the prosocial dimension of this choice seemed to be sufficient for them to reconsider their original motivations due, at least in part, to pro-environmental dispositions. In addition, and in line with previous research conducted in a more general persuasion context (Baron, Baron, & Miller, 1973; Kumkale & Albarracín, 2004), Cornelissen et al. (2007) showed that social labelling needs a distraction task to be effective. When their protocol did not include any cognitive load, social labelling activated persuasion knowledge (Friestad & Wright, 1994), which led the individuals to reject the label (Burger, 1999).

Baron, R. S., Baron, P. H., & Miller, N. (1973). The relation between distraction and persuasion. Psychological Bulletin, 80, 310–323.

Burger, J. M. (1999). The foot-in-the-door compliance procedure: Amultiple-process analysis and review. Personality and Social Psychology Review, 3, 303–325.

Cornelissen, G., Dewitte, S., Warlop, L., & Yzerbyt, V. (2007). Whatever people say I am, that's what I am: Social labeling as a social marketing tool. International Journal of Research in Marketing, 24, 278–288.

Friestad, M., & Wright, P. (1994). The persuasion knowledge Model: How people cope with persuasion attempts. Journal of Consumer Research, 21, 1–31.

Kumkale, G. T., & Albarracín, D. (2004). The sleeper effect in persuasion: A meta-analytic review. Psychological Bulletin, 130, 143–172.

Tybout, A. M., & Yalch, R. F. (1980). The effect of experience: A matter of salience? Journal of Consumer Research, 6, 406–413.

Van der Werff, E., Steg, L., & Keizer, K. (2014). I am what I am, by looking past the present: The influence of biospheric values and past behavior on environmental selfidentity. Environment and Behavior, 46, 626–657.

Now, we think you are perfectly right on one point. We did not provide enough evidence in the first version of our manuscript to support the following statement “the application of a “creative” label will be effective if the subjects do not perceive the persuasion attempt behind the label.” Therefore, we modified the following paragraph in the new version of our manuscript in order to be more convincing regarding the aforementioned statement:

In line with the literature, we contend that labeling an individual as “creative” could alter his or her self-perception, leading to a view of the self as creative, as soon as the label appears plausible. To build a plausible label, experimenters usually rely on people’s previous statements (Strenta & DeJong, 1981; Tybout & Yalch, 1980; Van der Werff, Steg & Keizer, 2014) or recent behavioral evidence (Cornelissen et al., 2007; Summers, Smith & Walker Reczek, 2016). When the label does not appear as plausible, persuasion knowledge (Friestad & Wright, 1994) is activated and the label is rejected (Burger, 1999; Cornelissen et al., 2007). As a conclusion, the persuasive intent behind the label must be masked for the self-perception to be altered among adults.

It is possible that “a “creative” label enhances self-perception of creativity while a “not creative” label erodes self-perception. The moderation analysis highlights this possibility. However, to have evidence for this statement, it is important to incorporate the measures of self-perception before and after the manipulation to demonstrate that with a positive label the self-perception goes up, while with “negative label” self-perception goes down.

We incorporated this in Study 3 (see above for the details):

The aim of Study 3 is to examine whether self-perceived creativity mediates the influence of a “not creative” label on the involvement in the creative task. To do so, we used Hayes' (2012) PROCESS macro (model 4) and 1000 bootstrapped samples but did not find any significant mediation between the construct. However, and as expected, the analysis shows that the application of the “not creative” label (vs. no label) both significantly damages post-label self-perceived creativity (F=12.19, p<.01, ηp2=.057, ω=0.97) and increases the time spent on the task (F=3.03, p<.05, ηp2=.015, ω=0.54).

Study 1. Samples size in study one is 76 people which suggests that the study is probably underpowered. Did you do power calculation and a sample size calculation?

We thank you for this comment. Originally educated in psychology, we more often use the criterium of the effect size to evaluate whether results are important enough. However, we do acknowledge that the power of the test appears today as a more robust criterium.

To better satisfy the power criterium, we collected more data to complete Study 1 sample (please note that we now present Study 1 as Study 2). Precisely, we asked our market research institute to propose our study to 30 more panelists. These panelists were exposed to the “creative” label condition, which was previously underrepresented (n=28 versus n=48 for the control condition). Out of these 30 more respondents, 26 checked the manipulation regarding the label and were included in the database we provide along with our new submission (these new respondents are identified as 101, 102… 126 in the database). Following, the section dedicated to the presentation of the results of the study has been fully rewritten in our new submission to consider our data collection extension. Precisely, it shows that our results stay the same but that the addition of respondents allowed an increase in the power of our test, which reaches now 84,5% as now explicitly appears in the manuscript:

To do so, we used Hayes' (2012) PROCESS macro (model 4) and 1000 bootstrapped samples. The analysis first shows that the application of the label significantly influences post-label self-perceived creativity (�=.65, p<.01, ηp2=.067, ω=0.85).

Study 1. There was randomization in this study, why age, gender, education level, and self-perceived creativity are added as controls? There was no difference in these measures between groups, therefore these controls should be removed from the analysis

The control variables have been removed from the analysis. Most open, this has improved our results.

Study 1. The goal of the study was to show that the label influences participants’ self-perception of creativity, which was achieved. Why there was an analysis of self-efficacy? If it highlights the mechanism of how labels influence performance, why there is no measure of self-efficacy in Study 2? I would not state in the discussion that this research contributes to self-efficacy literature as there is no association between self-efficacy and the study conditions or performance.

We made sure not to speak about self-efficacy in Study 1 (previous Study 2) and checked that we were not stating that we were contributing to the literature on self-efficacy. Besides, as the addition of Study 3 minimize the role of creative self-efficacy in our whole paper, we also corrected the first paragraph of our introduction not to put too much focus on it.

Study 1. Neither design of this study nor analysis speaks to this hypothesis. While this study could be a part of the research package, it does not directly speak to any of the stated hypothesizes. The measure of self-perception demonstrates rather that the manipulation worked as researchers planned it.

In the new version of our manuscript, we reversed Study 1 and Study 2. Study 1 aims at explicitly testing our conceptual framework. Study 2 and Study 3, introduced following your very relevant comment, aim at exploring further the psychological mechanisms at stake behind our results. This is how it now appears:

We will test our conceptual model in Study 1, before further exploring the psychological mechanism at play when labeling an individual as “creative” in Study 2, i.e. the activation of creative self-efficacy, or when labeling an individual as “not creative” in Study 3, i.e. an enhanced involvement in the creative task.

[…]

Study 2 and Study 3 extend these first findings on the impact of social labeling in creativity management by investigating the psychological mechanisms at play when applicating a “creative” label (Study 2) or a “not creative” label (Study 3).

Study 2. Where the participants' removal in this study? Did you have the same manipulation checks as in the previous one? Did participants remember the labels correctly?

We discarded participants who did not recall the manipulation using the same manipulation checks in our 3 experimental studies. We made sure to make this clear in the new version of the manuscript.

Study 2. Covariates should not be included in the experimental study, especially that there were no differences in these variables between conditions.

The control variables have been removed from the analysis. Most open, this has improved our results.

Study 2. Please report degrees of freedom in moderation study, it seems underpowered, especially for the continuous measure.

We calculated power for the moderation effect as follows:

To make this interaction effect easier to communicate (Fitzsimons, 2008), we dichotomized participants between higher and lower scorers on pre-label self-perceived creativity, using a median-split. We then carried out a classic ANOVA considering the manipulation (i.e. a “not creative” label vs. a “creative” label) and the dichotomized version of pre-label self-perceived creativity as between-subjects factors. This ANOVA replicated the significant interaction effect previously identified by the floodlight analysis between the positive/negative nature of the label and participants’ self-perceived creativity (F(1,96)=4.898, p<.05, ηp2=.049, ω=0.71). Planned contrast tests also revealed that the “creative” label enhanced individual creativity performance among participants scoring low on self-perceived creativity, who outperformed the “not creative” label group (6.18 vs. 5.45, F(1,49)=6.417, p<.01, ηp2=.116, ω=0.80). However, among participants scoring high on self-perceived creativity, the “not creative” group outperformed the “creative” label group, though in a marginally significant way (5.90 vs. 5.41, F(1,41)=2.387, p<.10, ηp2=.055, ω=0.45). Corroborating H3a but only marginally H3b, these results are depicted in Figure 4.

We acknowledge that the power is only medium regarding the influence of the “not creative” label. As we found that this symmetrical influence was clearly of interest, we chose to keep it in the new version of our manuscript and acknowledge this as a limitation in our final discussion.

This research contributes to the literature on social labeling. In showing the positive effect of a positive “creative” label on creative behaviors, it replicates previous findings observed in psychology and social marketing in an innovation management setting, thus extending the boundaries of social labeling’s relevance to the organizational field. In demonstrating the moderating effect of self-perceived creativity, it also offers a theoretical validation of social labeling theory itself, whose basic mechanism involves an alteration of individuals’ self-perception. Furthermore, in showing the positive effect of a negative “not creative” label, it corroborates the congruence hypothesis formulated by DeJong (1979) and Guéguen (2001), according to which a negative label only generates reactance when the label and the requested behavior are unconnected, such that the requested behavior offers no relevant opportunity to contradict the negative label. Additionnaly, Study 2 and Study 3 provide convincing empirical evidence about the mechanism at play behind the effect of each type of label. Still, we acknowledge here that the positive effect of a negative “not creative” label on individual creativity performance only appears marginal in Study 1 and therefore calls for replication.

However, would you prefer not to keep the focus on the “not creative” label in the manuscript, we would obviously follow your advice.

Study 2. While moderation analysis demonstrates the possibility of the highlighted in the introduction relationships, the further mediation analysis is needed to provide further evidence (see my comments above)

It appeared quite difficult to test the whole mediation in the same experiment. This is why we now provide 3 experiments. The first one demonstrates the effects on individual creativity performance and the two follow-ups investigate deeper the mechanisms at play. We hope that the addition of a new experiment to better explain the effects of labeling will convince you.

Study 2. I don’t see the value in following Hayes process with ANCOVA, you could remove it from the study and visualize your observations directly from the Process Model 1 procedure.

We do agree with the fact that such an analysis is not statistically necessary. Still we find it to be quite relevant to ensure a good communication of our experimental results. This is how we often proceed in our disciplinary field. Would you insist on getting rid of it, we naturally will follow your advice.

While I can see that the work of Bem, 1972 might be related to social labeling, It would be great if authors could elaborate on the connection between self-perception theory and social labeling.

To make it clearer, we do not state anymore that social labeling mechanism relies on Bem’s (1972) self-perception theory, which could be deceptive as social labeling does not require any previous behaviors to modify self-perception. We now insist on the re-attribution of dispositional properties that has been first developed by Bem (1972) and which allows to enlighten how social labeling actually works. This is how it appears in the new version of our manuscript:

Social labeling is “a persuasion technique that consists in providing a person with a statement about his or her personality or values (i.e. social label) in an attempt to provoke behavior that is consistent with the label” (Cornelissen, Dewitte, Warlop & Yzerbyt, 2007, p.279). Bem’s (1972) re-attribution mechanism explains the persuasion at stake in social labeling. Although self-observation of an individual’s own behavior provides the clearest information to make self-attribution of dispositional properties (Ouellette & Wood, 1998), labels provided by others can also be informative about a person’s traits and values and lead to their re-attribution to the self (Strenta & DeJong, 1981).

Submitted filename: Response to the reviewers.docx

Click here for additional data file.

16 Oct 2019

PONE-D-19-16350R1

Nudging individuals’ creativity using social labeling

PLOS ONE

Dear Dr Agogué,

Thank you for submitting your manuscript to PLOS ONE.

While one of the reviewers is happy with your revision, the other one still has some major comments. Moreover, I have noticed that you did not address my comments, which I copy and paste below from my previous decision letter:

"I have now collected two reviews from two experts in the field. Both reviewers recommend major revisions. Therefore, I would like to invite you to revise the paper according to the reviewers' suggestions. Needless to say that all comments must be addressed. Particular attention, however, should be given to the issue raised by both reviewers regarding the smallness of the sample (especially in light of the current replicability crisis) and to the issue raised by Reviewer 1 regarding deception (please explain exactly which parts of the experiments involved deception and why you think this is not a problem for your experiments). Moreover, I would like to mention that social labels have been recently used also to impact people's decisions in economic games (https://www.sciencedirect.com/science/article/pii/S0022103116302098; https://www.sciencedirect.com/science/article/pii/S0022103118302841; http://journal.sjdm.org/17/171107/jdm171107.pdf; http://journal.sjdm.org/19/190107/jdm190107.pdf). Note that I am the author of some of these papers. My decision on this manuscript will obviously be independent of whether you will decide to include or not these papers in your reference list. I have just thought that you might find them relevant.

Looking forward for the revision."

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

Additional Editor Comments (if provided):

While one of the reviewers is happy with your revision, the other one still has some major comments. Moreover, I have noticed that you did not address my comments, which I copy and paste below from my previous decision letter:

"I have now collected two reviews from two experts in the field. Both reviewers recommend major revisions. Therefore, I would like to invite you to revise the paper according to the reviewers' suggestions. Needless to say that all comments must be addressed. Particular attention, however, should be given to the issue raised by both reviewers regarding the smallness of the sample (especially in light of the current replicability crisis) and to the issue raised by Reviewer 1 regarding deception (please explain exactly which parts of the experiments involved deception and why you think this is not a problem for your experiments). Moreover, I would like to mention that social labels have been recently used also to impact people's decisions in economic games (https://www.sciencedirect.com/science/article/pii/S0022103116302098; https://www.sciencedirect.com/science/article/pii/S0022103118302841; http://journal.sjdm.org/17/171107/jdm171107.pdf; http://journal.sjdm.org/19/190107/jdm190107.pdf). Note that I am the author of some of these papers. My decision on this manuscript will obviously be independent of whether you will decide to include or not these papers in your reference list. I have just thought that you might find them relevant.

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: All comments have been addressed

Reviewer #2: (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

Reviewer #2: Yes

**********

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

Reviewer #1: Yes

Reviewer #2: 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: Yes

Reviewer #2: (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

Reviewer #2: 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: (No Response)

Reviewer #2: Thank you for reviewing the manuscript. As before, I applaud your idea and interesting findings specifically with regard to the “non-creative” label. While I enjoy reading the updated version that indeed clearly communicated the framework. There are still changes that need to be made specifically with regard to the analysis of studies 2 and 3.

Abstract

Minor: Study 2 & 3 need to be better described in the abstract. Second to the last sentence is not clear.

Study 1: Please report the sample and the conditions that you included in the moderation analysis. Please remove analysis with dichotomization on page 16, but report the figure directly from PROCESS Model 1 with a continuous variable. To do so, you need to add in the PROCESS the option: “generate data for the plot” (the same relevant for Study 3). Please report conditional effects in Model 1, rather than re-running the analysis in ANOVA

Study 2: Please clarify whether you measured creative self-efficacy, before and after exposing people to labels. Did you control for pre-labeled self-perceived creativity in the mediation analysis?

Could you please provide a table that summarizes correlations between the variables in this study. The table will help to illustrate whether there is a collinearity problem between measures. The table should include pre-labeled self-perceived creativity, post-labeled self-perceived creativity, pre-(?) and post-self-efficacy, and study conditions.

I recommend reporting this study with repeated measure design, in which you have pre and post-self-perceived creativity with conditions as a factor. (See my comment below)

Study 3: The analysis of study 3 is confusing. It needs a picture with a “classic” triangle with a mediator on top. The conclusion as it stands now “both significantly damages post-label

self-perceived creativity (F=12.19, p<.01, ηp2=.057, ω=0.97)” looks like a result of ANOVA analysis but not Hayes Model 4. Please re-run the analysis as I suggested below.

Study 2 & 3 analysis and results are problematic as they stand now. Please state at the beginning that you are planning to test in two different studies conditions with creative and non-creative labels versus control conditions. Both studies need to have identical analyses answering three questions.

Question 1: whether a label changes self-perceived creativity before and after the manipulation. This could be accomplished with a repeated measure design and the conditions as a factor.

Question 2: whether low base self-perceived creativity moderates the effect of conditions on behavior (or its proxy, self-efficacy). In both studies, Hayes model 1 needs to be applied to answer this question. In model 1, conditions need to be included as independent variables, while base self-perception as moderator and dependent variables should be: self-efficacy (a proxy of behavior) in study 2 and a brick-task performance in Study 3. This analysis needs figures reported from Model 1.

Question 3: whether the change in self-perceived creativity influences participants’ behavior (or a proxy of it). A mediation analysis should be used. For study 2: conditions (independent variable)->post-self-perception controlling for pre-self perception (mediator) -> self-efficacy (dependent variable). For study 3: conditions (independent variable)->post-self perception controlling for pre-self perception (mediator) -> brick-task performance (dependent variable). This analysis needs figures.

Minor comment: please keep boosting samples consistent throughout the manuscript; report statistics with degrees of freedom e.g. t (df?)=1.98, p<.05; and confidence intervals.

With regard to the comments:

In study 1, you suggested “to make our label credible, participants were first asked to complete an initial questionnaire about their self-assessed ability to generate creative ideas” is it the same as “pre-label self-perceived creativity measure” in studies 2&3? If no, how did you mask a persuasive intend in study 2&3?

About the last comment: please do not dichotomize variables but report the statistics and figures from Hayes directly.

**********

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: Yes: Harry M. Kaiser

Reviewer #2: 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.

11 Jan 2020

Please see attached document for response to specific reviewer and editor comments.

Submitted filename: Response to the reviewers.docx

Click here for additional data file.

21 Jan 2020

PONE-D-19-16350R2

Nudging individuals’ creativity using social labeling

PLOS ONE

Dear Dr Agogué,

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.

We would appreciate receiving your revised manuscript by Mar 06 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):

One of the reviewers suggests minor revisions before publication. Please address these remaining comments. I am looking forward for the final version.

[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 #2: All comments have been addressed

**********

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 #2: Yes

**********

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

Reviewer #2: 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 #2: (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 #2: 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 #2: The data are well described. I can easily follow the experiments and what was done for the experiment. Thank you for the changes.

Minor comments:

I assume that in Figure 3, error bars show the standard deviation. The figure should show a standard error, it allows readers to estimate the statistical difference from the figure.

Confidence interval 90 is unusual, the conventional way to report a confidence interval of 95 as it corresponds to conventional expectations of p < .05.

It is concerning that when you add pre-label self-perceived creativity in study 3 the indirect effect disappears. However, it could be an issue of low power. Can you please comment on the effect sizes, e.g. what is the mean difference between conditions in post-label self-creativity? and what is the mean difference between conditions in individual creativity performance? That information will allow a reader to estimate the magnitude of the effect of the "non-creative" label.

**********

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 #2: 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.

24 Jan 2020

See attached document

Submitted filename: Response to the reviewers.docx

Click here for additional data file.

28 Jan 2020

Nudging individuals’ creativity using social labeling

PONE-D-19-16350R3

Dear Dr. Agogué,

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-19-16350R3

Nudging individuals’ creativity using social labeling

Dear Dr. Agogué:

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