T
AKE IT SERIOUSLY
!
C
AN SANCTIONS IN COMBINATION WITH A MORAL APPEAL INDUCECOMPLIANCE
?
UNIVERSITY OF GRONINGEN Faculty of Economics and Business MSc Human Resource Management
SUBMITTED: June, 9th 2016
LISA TEUBER l.teuber@student.rug.nl
SUPERVISOR: Dr. Laetitia B. Mulder Department of Human Resource Management & Organizational Behavior
University of Groningen
ABSTRACT
Non-compliance with rules is common in everyday life and policy makers routinely impose sanctions to discourage such behavior. The prevailing assumption is that punishments render these actions less attractive and hence people refrain from it. However, sanctions can also backfire and reduce compliance by undermining people’s moral concerns. On the other hand, sanctions might also evoke moral concerns, which cause people to be intrinsically motivated to comply. I hypothesize that combining a sanction with a moral appeal will increase the likelihood that the sanction is interpreted as retributive and hence evokes moral concerns. Therefore, people will be motivated by their own morality to refrain from deviant behavior and compliance will be increased. The hypotheses were tested in a 2 (sanction vs. no sanction) x 2 (moral appeal vs. no moral appeal) field experiment at the research lab of the University of Groningen by randomly assigning participants to the four conditions. Nonetheless, the outcomes of this field experiment did not confirm the proposed hypotheses. The sanction caused participants to show deviant behavior instead and the moral appeal did not have any moderating effect on sanction and moral concerns. This might have been due to the suboptimal installation of the sanction and the moral appeal in the experiment.
TABLE OF CONTENTS 1. Introduction ... 2 2. Theory ... 5 Sanctions ... 5 Moral Appeal ... 7 3. Methodology ... 11
Design and participants ... 11
1. INTRODUCTION
Understanding people’s compliance with rules, laws, and social norms is crucial for societies as well as organizations because non-compliant or even deviant behavior is common in everyday life. While there is a downward trend for major crimes such as homicide or violent robbery in Europe, minor crimes such as domestic burglary are on the rise (Eurostat, 2013) and minor convictions even seem part of people’s lives: people exceed speed limits, illegally download movies or litter on streets. There is also an increasing body of literature focusing on deviant workplace behavior (e.g. Dalal, 2005) such as workplace aggression and violence, bullying or damaging company property (e.g. Chiaburu & Harrison, 2008; Bennett & Robinson, 2000; Bowling & Beehr, 2006; Porath & Pearson, 2013). Policy makers and organizations routinely impose penalties to discourage such behavior. The prevailing assumption is that punishments will render the undesired actions less attractive (Cooter, 1998; McAdams, 2000; Williams & Hawkins, 1986) and various experimental studies have proven that sanctioning systems can indeed successfully increase compliance (e.g. Caldwell, 1976; Gächter, Renner, & Sefton, 2008; Rand, Dreber, Ellingsen, Fudenberg, & Nowak, 2009).
Nevertheless, sanctions are not always effective and research suggests that there is also a dark side to sanctions (e.g. Fehr & Gächter, 2000; Gneezy & Rustichini, 2000; Tenbrunsel & Messick, 1999). It has been shown that sanctions can transform decisions about moral behavior into economic considerations in which people calculate their costs and benefits related to defection or compliance. Sometimes, sanctions can even decrease compliance and undermine people’s moral considerations altogether (Frey & Oberholzer-Gee, 1997; Gneezy & Rustichini, 2000; Mulder, Verboon, & de Cremer, 2009; Tenbrunsel & Messick, 1999).
concerns have shown to be a strong motivator for people to comply with rules (e.g. Cialdini, Demaine, Sagarin, Barret, Rhoads, & Winter, 2006; Cooter, 1998; Reno, Cialdini, & Kallgren, 1993).
Therefore, the question arises which conditions determine a sanction’s effectiveness. I argue that sanctions are more effective at increasing compliance, when their message convinces people about the immorality of their behavior. Following Kantian Deontology (1785), individuals are not always motivated by self-interest, but by their moral principles and duties. More anecdotal evidence from everyday life shows that people also adhere to moral commands such as “reduce garbage, buy organic produce, consume fair-trade products, be a team player”. These commands become moral duties for people’s lives and a sanction has shown to be effective at communicating these duties to people (Cooter, 1998; McAdams, 2000; Mulder, 2008; Thøgersen, 2003; Williams & Hawkins, 1986). Considering the motivational effect of moral norms, I argue that a sanction that evokes people’s moral concerns might effectively steer their behavior. Generally, people are inclined to define themselves as moral beings and act accordingly (Aquino & Reed 2002; Blasi 1983, 1984). If a sanction convinces people that their behavior is morally wrong, people would refrain from it out of intrinsic conviction. This would make people adhere to rules even though they might not be monitored, the fine is not costly, or the sanction would be removed altogether (Mulder & Nelissen, 2010, Wenzel, 2004). Moreover, this would encourage people to base their actions on their moral standards instead on the extrinsic sanction system. This in turn would limit the likelihood that moral concerns are undermined causing economic considerations of costs and benefits instead (Mulder, 2008). Following from this, I expect that appeals to people’s moral norms contribute to a sanction’s effectiveness at increasing compliance.
moral appeals in comparison to sanctions on increasing tax returns from people (e.g. Fellner, Traxler, & Sausgruber, 2013; McGraw & Scholz, 1991; Schwartz & Orleans, 1967) and research from the field of behavioral economics focused on increased levels of cooperation in social dilemma games due to moral appeals in comparison to other incentives (e.g. Chen, Pillutla & Yao, 2009; Dal Bó & Dal Bó, 2014; Pruckner & Sausgruber, 2013). Hence, prior research has mainly focused on measuring the effectiveness of moral appeals and other incentives separately from one another. However, research has yet failed to study the effectiveness of a sanction in combination with a moral appeal on increasing people’s compliance.
2. THEORY Sanctions
Modern societies regulate people’s behavior through various external incentives and sanctions are a common mean to do so. Research has shown that sanctions simply make a certain action less attractive and thus steer people in the desired direction, because normally individuals want to avoid the costly punishment (Cooter, 1998; McAdams, 2000; Williams & Hawkins, 1986). Therefore, sanctions can be effective at encouraging people to comply with existing rules (Balliet, Mulder, & van Lange, 2011; Caldwell, 1976; Gächter, Renner, & Sefton, 2008; Rand, Dreber, Ellingsen, Fudenberg, & Nowak, 2009; Yamagishi, 1986).
into economic considerations about costs and benefits of compliance, which caused the parents to rather pay the fine and instead keep up their morally unacceptable behavior. I argue that a sanction might induce compliance due to its deterring effect on people. However, one could still risk negative side effects such as economic considerations, which may cause the sanction to backfire.
Moreover, there is also substantive research arguing that sanctions also work in the opposite way and evoke private moral concerns about people’s own behavior instead of economic considerations about costs and benefits (Cooter, 1998; McAdams, 2000; Mulder, 2008; Thøgersen, 2003; Williams & Hawkins, 1986). Harsanyi (1969) noted that people’s behavior cannot solely be explained by self-interest for economic gains or losses, but also by the need for social acceptance through following society’s moral norms. These norms have shown to be a strong motivator for people to comply with society’s rules and also to judge trespassing these rules as morally wrong (Bicchieri, 2006; Cialdini et al., 2006; Cooter, 1998; Mulder, 2008; Reno, Cialdini, & Kallgren, 1993). Indeed, literature on tax evasion shows that most people pay costly taxes even though detection of evasion is rather unlikely and that they feel the duty to do so (Torgeler & Schneider, 2005; Scholz & Pinney, 1995).
Sanctions are said to communicate to people what society’s social rules and laws are and provide a moral norm to people about what is socially disapproved of (McAdams, 2000). Feldman (2009) for example showed that society’s laws evoke moral concerns among people in the sense that they perceive trespassing such a rule as immoral. Indeed, due to the norm-expressive function of a sanction system, people are inclined to punish perpetrators, because their behavior is seen as morally unacceptable and against society’s norms (Cooter, 1998; Williams & Hawkins, 1986).
deserts’ (Darley & Pittman, 2003; Mulder, 2008, 2009). When a sanction is interpreted as compensatory, people engage in extrinsic motivated business calculations and could decide to pay the costs of the fine rather than complying with the rule (Kurz, Thomas, & Fonseca, 2014; Mulder, 2009). Contrarily, when a sanction is interpreted as retributive, it communicates to people that their behavior is morally unacceptable and that the sanction is installed to restore the injustice caused. In this case, people show compliance, because they become aware of the moral implications of their behavior and moral concerns are evoked (Darley & Pittman, 2003; Kurz, Thomas, & Fonseca, 2014; Mulder, 2008, 2009). For example, Kurz et al. (2014) showed that a sanction imposed on latecomers, which was framed in a retributive manner in comparison to a compensatory sanction, significantly evoked moral concerns about being late and also increased levels of punctuality.
Nevertheless, the question arises, how sanctions could be effectively framed in a retributive way, so that its negative effects are limited and moral considerations about one’s behavior are increased instead.
Moral Appeal
Additionally, McGraw and Scholz (1991) showed that an appeal to tax payers’ moral obligations lead people to evaluate the tax law from a social perspective. People more strongly perceived the law as socially acceptable and fair for the poor. In contrast, the sanction on tax evasion induced a more self-interested evaluation, letting people focus more on their individual gains and losses. Additionally, scholars from the field of behavioral economics, which traditionally rejects the assumption of people as mere rational agents (e.g. Kahneman & Tversky, 2000; Baron, 2000; Rabin, 1998), showed that a moral appeal significantly increased pro-social behavior in social dilemma games and was further effective at preventing people from cheating (Chen, Pillutla & Yao, 2009; Dal Bó & Dal Bó, 2014; Mazar, Amir, & Ariely, 2008; Pruckner & Sausgruber, 2013).
However, research on the effectiveness of moral appeals to induce compliance is ambiguous. Neither McGraw and Scholz (1991) nor other studies from the field of tax evasion could reassert the findings that a moral appeal significantly increased compliance (Blumenthal, Christian, & Slemrod, 2001; Fellner, Traxler, & Sausgruber, 2013; Tittle & Rowe, 1973; Torgler, 2013; Wenzel & Taylor, 2004).
The ambiguous results of the effectiveness of moral appeals on increasing compliance suggest that a moral appeal alone might not be a reliable incentive. However, a moral appeal might be effective at increasing moral concerns amongst people. Consequently, I assume that a sanction and a moral appeal may strengthen each other at evoking moral concerns. This is due to the fact that the message of the moral appeal might render the moral aspects of the sanction more salient to people. This might cause them to rather interpret a sanction as retributive instead as a mere compensation for their behavior.
participants increased and could also be sustained over several rounds of their experiment. However, the increased levels of cooperation in their experiment might have resulted from the fact that people improved their behavior after having received a costly punishment and then valued compliance as more beneficial in the subsequent rounds of the experiment. Moreover, in real life, sanctions are imposed from above, for example by governments or senior managers in organizations, and rarely installed or given by peers. This makes it much more relevant to test the effect of sanctions and moral appeals in the context of an authority installing the sanction, and to test whether the authority is able to prevent people from trespassing in the first place. Nevertheless, Dal Bó and Dal Bó’s research (2014) shows the beneficial effect of combining a sanction with a moral appeal to induce compliance. I expect that under the condition of a present moral appeal, the sanction will more likely evoke moral concerns amongst people. Moreover, I assume that moral concerns will mediate the conditional relationship between sanction, moral appeal, and compliance and hence lead to increased levels of compliance. However, I expect that when a moral appeal is absent, the sanction will directly increase compliance among people. That is likely due to its deterring effect on people, who want to avoid the costly punishment. Nevertheless, I hypothesize that in this case moral concerns amongst people will not be evoked.
Hypothesis 1: When a moral appeal is present, the sanction will more likely increase moral concerns, compared to when a moral appeal is absent.
Hypothesis 2a: When a moral appeal is absent, the sanction will directly increase compliance without evoking moral concerns.
Figure 1 visualizes the expected relationships of my research displays the proposed
conceptual research model.
3. METHODOLOGY Design and participants
The hypotheses were tested as a field study at the Faculty of Economics and Business of the University of Groningen. Participants were students who had signed up as participants to an experiment called “HR decision-making, memory, and personality”. Nevertheless, the study was actually on health and HR selection decisions. Participant’s effort as well as attention during the experiment on health served as an operationalization of compliance for this field study. The participants took part in the experiment either as part of their undergraduate program in exchange for four research points, or for a compensation of eight euros. The data was collected within two weeks.
For this 2 (sanction vs. no sanction) x 2 (moral appeal vs. no moral appeal) field study 258 students (118 men, 140 women; Mage = 21.64 years, SDage = 2.57) participated.
Participants were randomly assigned to each condition (Nsanction = 63; NMoral Appeal=61; NCombined=65; NControl= 64). The sample consisted of students from different nationalities, 160
Dutch (61.8%), 20 German (7.7%), 11 Chinese (4.2%), and 67 others (26.3%). The participants also came from different study programs, 178 Bachelor (68.7%), 45 Master (17.4%), 23 Pre-Master (8.9%), and 12 others (4.6%). 140 participants took part for money as compensation (54.1%) and 118 participated for research points (45.6%). Six participants, who indicated their suspicion that the study on health and HR selection decisions additionally tested the effect of sanction and moral appeal during the experiment, were excluded from all further analyses.
Procedure
not obtain any information about this field study. Participants were seated in separate computer cubicles. The communication with other participants was prohibited. All further instructions for the experiment appeared on the computer screens. Participants were randomly assigned to each condition and the introductory message at the beginning of the experimental session was manipulated. The first welcoming message was the same for all four conditions: “Welcome in the FEB research lab! You will now participate in the study ‘Personality,
Memory and HR decision making’.” Only in the moral appeal condition, participants further
read the following: “By participating you will contribute to scientific knowledge about human
decision making in business and economic settings. Results of this research will be published in international scientific journals and also in your study books. In order to ensure valid and reliable results, it is essential that not only we as researchers develop sound research methods, but also that research participants like yourself take their own role seriously.”
Participants in both, the moral appeal and sanction condition, read the following: “We ask you
to do this research in a concentrated and serious way. This means: read the instructions well and pay close attention to the questions, turn mobile devices off, and don’t be distracted by any other activities.” For participants in the moral appeal condition, the following statement
was included: “Only then the results will be useful. You can now start the experiment.” Instead in the sanction condition, participants read the following message: “During this study,
it will randomly be checked whether participants pay attention and do the experiment in a serious and concentrated way. If there is evidence that participants fail to do so, these participants will not be compensated. You can now start the experiment.” In the combined
condition of moral appeal and sanction, participants read an introduction with the combination of first the moral appeal and then the sanction. In the control condition, participants only read the following: “Welcome in the FEB research lab! You will now participate in the study
‘Personality, Memory and HR decision making’. You can now start the experiment.” After the
uncontrollable determinants of health or health measurements for governments, which served as a manipulation for the study on health and HR selection decisions. Next, participants were instructed to write a summary about the main arguments of the article on a piece of paper and slip it under the door of their computer cubicle. The researchers collected the summaries afterwards. Subsequently, participants were asked to answer an attention check (refer to
Appendix A for details on the attention checks), which was testing whether they were paying
close attention to the instructions. The directions stated that differences in how people feel, their previous knowledge and experience, and their environment could affect choices. Further, it was stated that thus we would be interested in information about the participant. Specifically, we would be interested in whether the participant had actually taken time to read the instructions and in order to show that he/she paid attention, participants were instructed to ignore the question below about how he/she was currently feeling and instead check the ‘none of the above’ option and write the word ‘none’ in the text box. Afterwards, participants completed several HR selection tasks related to the experiment on health and HR decisions. One further attention check stating: “To show that you are paying attention, skip this item and
do not tick a box.” was hidden in a grid of statements about people’s opinions on the
controllability of health. In addition, a third attention check tested people’s attention after having completed several grids of items relating to people’s opinion about health in general. The instructions stated that some people claimed that the sign of their horoscope would also influence how thoroughly people read texts and that this question aimed at studying this. People were instructed to ignore the question about what astrological sign they were and instead were asked to indicate their favorite sport in the text box. Afterwards, participants were asked to answer several items concerning personal attitudes towards health. In this grid of items, a forth attention check item was hidden stating the following: “In order to show that
you are paying attention don’t tick any box.” Subsequently, participants read the fifth
of the screen. Do not click on the scale items that are labeled from 1 to 9. This is just to screen out random clicking.” In the end, participants were asked to indicate their moral
concerns with regard to taking the experiment seriously. This measurement scale was followed by manipulation check items and at the end of the study participants were instructed to complete the personality measures.
Measurement
Manipulation checks. For the moral appeal manipulation check, participants indicated on a 7-point Likert-type scale ranging from 1 (totally disagree) to 7 (totally agree) whether in the beginning of the experiment: “A moral appeal was made to do the experiment in a serious
and concentrated way.” and whether “It was explained why it is important to do the study in a serious and concentrated way.”. These two items formed a rather reliable scale, α=.64. The
sanction manipulation check included two items whether “It was explained that participants
might get punished in case there is evidence that they didn’t do the experiment in a serious and concentrated way.” and if “After the experiment it will be checked randomly whether participants did the experiment in a serious and a concentrated way.”. However, the two
items failed to form a reliable scale, α= .28, so each item was tested separately.
Second, the five attention checks served as a measurement of how serious participants paid attention during the experiment. Oppenheimer, Meyvis, and Davidenko (2009) noted that participants who do not take experiments seriously will not bother to pay close attention to the instructions or questions on a study. Attention checks therefore reveal who pays attention and who does not. The attention checks (see Appendix A) were varied and adapted to fit this experiment according to Oppenheimer, Meyvis and Davidenko’s, (2009) ‘Blue Dot Task’. However, the measure of attention checks failed to provide a reliable scale for the measure of compliance (α= .47). Nevertheless, as the number of attention checks passed does likely indicate the amount of attention paid, a collapsed variable was created. It served as an index variable of how many of the five attention checks participants had passed ranging from 0 (passed none) to 5 (passed all).
Third, the first attention check in the experiment served as a measure of compliance. By including attention checks, one could risk that participants become aware of the monitoring system and thus change their behavior once they noticed the first attention check. Ellsworth (2010) noted that researchers therefore risk priming respondents about the aim of the manipulation. Further, I assumed that the manipulations were still fresh and hence more effective for the very first attention check in the experiment. Thus, the dichotomous dependent variable of passing vs. failing the first attention check was included.
Moral concerns.1 The scale for moral concerns was based on a measure of Mulder (2008) and Feldman (2009) and adapted to the behavior of doing the experiment in a serious way (e.g. “It is one’s duty as a participant to do the experiment in a serious way.”, “It goes
1 A principal component analysis based on a forced principle with Kaiser rotation was conducted on the initial
against my moral principles not to do the experiments in a serious way”, or “It is morally wrong not to do the experiments in a serious way.”). All six items were measured on a 7-point
Likert-type scale ranging from 1 (totally disagree) to 7 (totally agree) and yielded a highly reliable scale α= .85.
Form of compensation. The form of compensation was measured through one dichotomous variable (research points vs. money) since I assumed that how people got compensated might influence their reaction towards the sanction system. Indeed, previous research has noted that merely activating the concept of money (e.g. touching money) causes people to increase self-interested behavior and engage in economic considerations (e.g. Reutner & Wänke, 2013).
Conscientiousness. Additionally, a measure for personality traits was included, as I expected that the personality trait conscientiousness (being thorough and orderly, act dutifully) (Ozer & Benet-Martinez, 2006) might have an additional moderating effect upon participants’ effort and attention during the experiment. The 44-item Big Five Inventory was used with a Likert-type scale ranging from 1 (totally disagree) to 5 (totally disagree) (John, Donahue, & Kentle, 1991; John, Naumann, & Soto, 2008; mean r= .83). The nine items forming the conscientiousness scale yielded a reliable measure α= .74.
4. RESULTS Preliminary analysis
Inter-correlations. Table 1 displays the means, standard deviations, and inter-correlations of all independent and dependent variables. Notably, the presence of a sanction showed a significant negative correlation with the effort participants put into writing the summary r= -.20, p< .01, but not with the number of attention checks participants had passed
r= .04, p= .50. The presence of a moral appeal was not significantly correlated with effort r=
.06, p= .38 nor with any other variable. Moral concerns were significantly positively correlated with the number of attention checks participants had passed r= .35, p< .01, as well with the first attention check in the experiment r= .26, p< .01. Additionally, the form of compensation significantly negatively correlated with the number of attention checks passed,
r= -.13, p< .05. Sanction was significantly positively correlated with agreeableness r= .13, p<
Table 1: Pearson zero order correlations, means and standard deviations
Variables M SD 1 2 3 4 5 6 7 8 9 10 11 12
1 Gender 1.5 .50 .10 -.10 .04 .11 .08 .15* .08 .14* -.03 .05 .22**
2 Control Dummy Article 1 .34 .475 -.51** -.08 .06 .07 .10 .00 .03 -.05 -.03 .11 3 Control Dummy Article 2 .34 .473 .08 -.01 -.11 -.18** .02 .05 -.00 -.01 -.07
4 Moral Appeal -.01 1.00 -.00 -.07 -.08 -.07 .06 -.01 -.02 .01 5 Sanction -.02 1.00 .02 .13* .03 -.20** .04 .01 -.00 6 Moral Concerns 5.65 .89 .30** .21** .12 .35** .26** .13* 7 Agreeableness 3.57 .44 .21** .03 .15* .14* -.05 8 Conscientiousness 3.44 .53 .11 .12 .05 .08 9 Effort 3.37 1.12 -.01 .06 .10
10 Attention Checks passed 4.55 .81 .67** -.13*
11 1st Attention Check .91 .29 -.07
12 Form of compensation .54 .50 Note. *p < .05 (two-tailed), **p < .01 (two-tailed)
Main analysis
Manipulation checks. To test whether the two manipulations were successful, I tested whether participants’ judgments regarding the message of the introduction to the experiment differed between the sanction and moral appeal conditions. A 2 (sanction vs. no sanction) x 2 (moral appeal vs. no moral appeal) ANOVA by the moral appeal manipulation check revealed a significant main effect of moral appeal F(3,249)= 45.32, p< .001. Participants in the moral appeal condition (M= 5.38, SD= 1.17) felt more than participants in the non-moral appeal condition (M= 4.27, SD= 1.45) that a moral appeal was presented at the beginning of the experiment to take the experiment seriously. Thus, the moral appeal manipulation was successful. However, there was also a significant main effect of the moral appeal manipulation check on sanction F(3,249)= 5.56, p< .05. Participants in the sanction condition (M= 5.01, SD= 1.37) felt more than participants in the non-sanction condition (M= 4.63, SD= 1.46) that they were given a moral appeal in the beginning of the experiment. There was no significant interaction effect found.
Since the sanction manipulation check items had failed to form a reliable scale, I tested the two items separately. A 2 (sanction vs. no sanction) x 2 (moral appeal vs. no moral appeal) ANOVA by the sanction manipulation check item “In the beginning of the experiment
it was explained that participants might get punished in case there is evidence that they didn’t do the experiment in a serious and concentrated way.” revealed a significant main effect of
(sanction vs. no sanction) x 2 (moral appeal vs. no moral appeal) ANOVA by the sanction manipulation check item “After the experiment it will be checked randomly whether
participants did the experiment in a serious and a concentrated way.” revealed a significant
main effect of sanction, F(3,249)= 9.84, p< .01. Participants in the sanction condition (M= 4.90, SD= 1.28) felt more than participants in the non-sanction condition (M=4.36, SD=1.43) that after the experiment, it would be randomly checked whether they did the experiment in a serious way. No other significant effects were found. Overall, I could hence assume that the sanction manipulation was successful.
Moral concerns. I hypothesized that when a sanction is combined with a moral appeal, the sanction would be more likely to evoke moral concerns among participants about their behavior, compared to when a moral appeal is absent. (Hypothesis 1).
However, a 2 (sanction vs. no sanction) x 2 (moral appeal vs. no moral appeal) ANOVA by moral concerns did not confirm this assumption. There was no significant main effect of sanction on moral concerns F(3,249)= 0.10, p= .76 and also no significant main effect of moral appeal on moral concerns F(3,249)= 1.13, p= .29. Additionally, there was no significant interaction effect of sanction and moral appeal on moral concerns F(3,249)= 0.00,
p= .95. Hence, the hypothesis was not supported.
Compliance. I hypothesized that a sanction alone would be directly positively related to compliance of participants to take the experiment seriously due to its deterring effect that participants would want to avoid the costly punishment. Thus, the sanction would increase participant’s effort and the number of attention checks passed as well as participants’ attention on the first attention check in the experiment (Hypothesis 2a).
from what was expected, so that participants in the sanction condition put significantly less effort into the task compared to participants in the non-sanction condition. No other significant effects were found. The post-hoc Tukey HSD test showed that the moral appeal-only condition differed significantly from the sanction-appeal-only condition, so that participants in the moral appeal condition significantly put more effort into the summary task than participants in the sanction condition. Nonetheless, the control condition and also the combined condition were not significantly different from the other two conditions (Table 2).
Second, in order to test whether the sanction affected participants’ attention during the experiment, a 2 (sanction vs.no sanction) x 2 (moral appeal vs. no moral appeal) ANOVA by number of attention checks passed was conducted (refer to Table 2 for means and standard deviations). However, the assumption that the sanction would be directly positively related to the amount of attention was not confirmed and the analysis revealed no significant effects. Additionally, the post-hoc Tukey HSD test showed that none of the four conditions were significantly different from one another (Table 2).
Table 2: Means and standard deviations (reported in parentheses) between groups DV= Attention Checks Mo ra l A pp ea l Sanction No Yes No 4.57 ( .68) 4.54 ( .76) Yes 4.45 ( .94) 4.62 ( .84) Note. *Tukey HSD p< .05
Third, in order to test whether participants were failing or passing the first attention check in the experiment, using sanction and moral appeal as predictors, a logistic regression with 95% bias corrected confidence intervals based on 1,000 bootstrap samples was conducted (M= .91; SD= .29; Npassed =230, Nfailed= 23). A test of the model was statistically
not significant, indicating that the predictors did not reliably distinguished between failing the attention check or passing it, χ2=6.04, p= .30, with df= 5. Nagelkerke’s R2 = .05 (R2 = .05 for
Step 1, ∆R2 = .01 for Step 2, ∆R2 = .00 for Step 3) indicated no relationship between prediction and grouping. Additionally, no other significant effects of the predictors were found.
Moderated mediation. I hypothesized that only when a sanction is combined with a moral appeal, moral concerns would mediate the relationship between sanction and compliance (Hypothesis 2b). This would mean that through the conditional indirect effect of sanction and moral appeal through moral concerns, levels of effort put in the summary task, attention levels measured through the number of attention checks passed, and also the odds that participants had passed the first attention check in the experiment would increase.
Table 3: Linear model of predictors of compliance (effort & attention checks), with 95% bias corrected and accelerated confidence intervals reported in parentheses, based on 1,000 bootstrap samples
Note. *p < .05, **p < .01; †p< .10, R2
Effort = .03, ∆R2 Effort = .06 for Step 2 R2Attention Checks = .03, ∆R2 Attention Checks = .10 for Step 2
Effort Attention Checks
Predictor b SE β p Predictor b SE β p
Step 1 Step 1
Constant 2.74 [2.27, 3.20] .25 p< .01** Constant -1.61 [2.67, 4.51] .47 p< .01**
Gender .32 [0.05, 0.60] .14 .14 p< .05* Gender -.08 [-0.29, 0.11] .10 -.05 p= .45 Control Dummy Article 1 .16 [-0.17, 0.47] .17 .07 p= .35 Agreeableness .30 [0.06, 0.54] .12 .16 p< .05*
Control Dummy Article 2 -.24 [-0.08, 0.54] .16 .10 p= .14
Step 2 Step 2
Next, the third measure of compliance as passing or failing the first attention check was analyzed through a logistic regression with sanction, moral appeal, and moral concerns as predictors in order to test whether moral concerns were positively related to compliance. The logistic regression was conducted with 95% bias corrected and accelerated confidence intervals based on 1,000 bootstrap samples (Table 4). The test of the model was marginally significant in step 1 (χ2= 5.17, p= .08), but highly significant in step 2 after including the three predictors moral appeal, sanction, and moral concerns (∆χ2= 12.13, p< .01), with df= 1 and Nagelkerke’s R2= .15 (R2= .04 for Step 1, ∆R2 = .11 for Step 2). Moreover, moral concerns
were a significant predictor in the equation, β= .85, p< .01, suggesting a significant main effect of moral concerns: the odds of participants who have high moral concerns passing the first attention check are 2.33 times higher than of participants with low moral concerns.
Table 4: Logistic regression of predictors of compliance (1st attention check), with 95% bias corrected and accelerated confidence intervals reported in parentheses. Confidence
intervals and standard errors based on 1,000 bootstrap samples
Predictor Odds Ratio (95% CI)
b Lower Odds Upper
Step 1 Constant -4.65 [-9.27, -0.33] . Gender -.20 [-1.20, 0.77] .34 .82 1.95 Agreeableness 1.07 [0.23, 1.94]** 1.20 2.91 7.71 Step 2 Moral Appeal -.03 [-1.03, 1.07] .39 .97 2.40 Sanction -.09 [-0.91, 1.19] .44 1.09 2.71 Moral Concerns .85 [0.47, 1.46]** 1.44 2.33 3.79
Note. *p < .05 (two-tailed), **p < .01 (two-tailed), †p< .10
R2 (Nagelkerke)= .04 for Step 1, ∆R2= .11 for Step 2
Model= χ2 (1)= 5.17†, ∆χ2 (2)= 12.13**
(Hayes, 2013; Model 7) (Table 5). First, the mediator variable analysis on moral concerns with effort as outcome variable of the moderated mediation analysis did not reveal any significant effects. However, in the dependent variable model with effort as dependent variable, sanction indeed significantly predicted effort when moral concerns were included in the model, β= -.22, t(247)= -3.60, p< .01. Moreover, moral concerns marginally significantly predicted effort, β= .12, t(247)= 1.80, p= .09. Nevertheless, there was no significant conditional indirect effect of sanction and moral appeal on effort through moral concerns for moral appeal present, β= .00, BCa CI95% [-0.02, 0.04] and for moral appeal absent, β= .00,
BCa CI95% [-0.02, 0.03]. Figure 2 depicts the moderated mediation model of sanction and
moral appeal as predictors of effort, mediated by moral concerns.
Second, the analysis on number of attention checks passed, as second measure of compliance, did not reveal any significant effects in the mediator variable analysis on moral concerns. Moreover, in the dependent variable model with number of attention checks passed as an outcome variable, moral concerns significantly predicted the number of attention checks passed, β= .35, t(242)= 4.34, p< .01. Nonetheless, no other significant effects were found: there was no significant conditional indirect effect of sanction and moral appeal on number of attention checks passed through moral concerns for moral appeal present, β= .01, BCa CI95%
[-0.06, 0.07] and for moral appeal absent, β= -.02, BCa CI95% [-0.08, 0.04]. Figure 3 depicts the
Table 5: Regression results for the conditional indirect effects of sanction and moral appeal on compliance (effort & attention checks) with moral concerns as the mediating variable, with 95% bias corrected and accelerated confidence intervals reported in parentheses
Mediator Variable Model (DV= Moral Concerns)
Effort Attention Checks
Predictor b SE t Predictor b SE t
Constant .00 [-0.12, 0.13] .06 .03 Constant -.00 [-0.12, 0.12] .06 -.04
Gender .07 [-0.05, 0.20] .06 1.13 Gender .04 [-0.08, 0.17] .06 .72
Control Dummy Article 1 .01 [-0.14, 0.16] .08 .13 Agreeableness .29 [0.16, 0.42] .07 4.37**
Control Dummy Article 2 -.09 [-0.23, 0.05] .07 1.22 Sanction (centered) -.02 [-0.14, 0.10] .06 -.27 Sanction (centered) .01 [-0.12, 0.14] .06 .16 Moral Appeal (centered) -.04 [-0.16, 0.08] .06 -.64 Moral Appeal (centered) -.06 [-0.19, 0.06] .06 -.98 Sanction x Moral Appeal .03 [-0.09, 0.16] .06 .50
Sanction x Moral Appeal .00 [-0.12, 0.13] .06 .06
Note. *p < .05, **p < .01; †p< .10 R2
Effort = .02 R2Attention Checks = .10**
Dependent Variable Model (DV= Effort, Attention Checks)
Effort Attention Checks
Predictor b SE t Predictor b SE t
Constant -.02 [-0.14, 0.11] .06 -.25 Constant -.00 [-0.12, 0.12] .06 -.04 Gender .16 [0.04, 0.28] .06 2.56** Gender -.06 [-0.18, 0.06] .06 -.97 Control Dummy Article 1 .08 [-0.06, 0.23] .07 1.11 Agreeableness .05 [-0.06, 0.17] .06 .90 Control Dummy Article 2 .12 [-0.02, 0.25] .07 1.72† Moral Concerns .35 [0.19, 0.50] .08 4.34**
Moral Concerns .12 [-0.01, 0.25] .07 1.80† Sanction .04 [-0.08, 0.17] .06 .71 Sanction -.22 [-0.34, -0.10] .06 -3.60**
Note. *p < .05, **p < .01; †p< .10 R2
Conditional Indirect Effects for Sanction on Compliance through Moral Concerns at values of Moral Appeal
Effort Attention Checks
Effect Boot SE Boot LLCI Boot ULCI Effect Boot SE Boot LLCI Boot ULCI Moral Appeal present .00 .01 -.02 .04 .01 .03 -.06 .07 Moral Appeal absent .00 .01 -.02 .03 -.02 .03 -.08 .04
Figure 2: Moderated mediation model of effort including regression coefficients, the conditional indirect effects and bootstrapped confidence intervals
Third, in order to analyze the conditional indirect effect of sanction and moral appeal through moral concerns on the dichotomous dependent variable of passing or failing the first attention check, a moderated mediation analyses using logistic regression with 95% bias corrected confidence intervals based on 5,000 bootstrap samples was conducted (Hayes, 2013; Model 7) (Table 6). The first mediator variable model analysis revealed no significant effects. Subsequently, the logistic regression of the dependent variable model with first attention check (passed vs. failed) as dependent variable revealed that a R2 (Nagelkerke)= .14, indicating that there was hardly any relationship between prediction and grouping. Furthermore, sanction was no significant predictor in the equation, β= -.06, p= .48, contrarily to moral concerns, which was a significant predictor, β= .73, p< .01. However, there was no conditional indirect effect of sanction and moral appeal through moral concerns on the first attention check for moral appeal present, β= .03, BCa CI95% [-0.10, 0.19], and neither for
moral appeal absent, β= -1.01, BCa CI95% [-0.19, 0.08].
Table 6: Logistic regression results for the conditional indirect effects of sanction and moral appeal on compliance (1st attention check) with moral concerns as the mediating
variable, with 95% bias corrected and accelerated confidence intervals reported in parentheses
Mediator Variable Model (DV= Moral concerns) Predictor b SE t Constant -.00 [-0.12, 0.12] .06 -.03 Gender .02 [-0.10, 0.15] .06 .40 Agreeableness .30 [0.17, 0.43] .07 4.48** Sanction (centered) .00 [-0.12, 0.12] .06 .04 Moral Appeal (centered) -.04 [-0.16, 0.08] .06 -.68 Sanction x Moral Appeal .04 [-0.08, 0.17] .06 .70
Conditional Indirect Effects for Sanction on Compliance through Moral Concerns at values of Moral Appeal
Effect Boot SE Boot LLCI Boot ULCI Moral Appeal present .03 .07 -.10 .19 Moral Appeal absent -1.01
.07 -.19 .08
Note. 1st Attention Check was coded (1= passed, 0= failed)
Additional measures
Form of compensation. I assumed that the form of compensation (research points vs. money) has an effect on people’s reactions towards the sanction system. A 2 (sanction vs. no sanction) x 2 (research points vs. money) ANOVA by effort did not confirm this assumption. Thus, no significant effects were found.
A second 2 (sanction vs. no sanction) x 2 (research points vs. money) ANOVA by number of attention checks passed revealed a marginally significant main effect of form of compensation on attention checks passed, F(6,246)= 3.54, p= .06. Participants who took part in the experiment for research points marginally significantly passed more attention checks (M= 4.66, SD= .63) than participants who took part for money as compensation (M= 4.45,
SD= .92). However, no further significant effects were found. Dependent Variable Model – Logistic Regression
(DV= 1st Attention Check) Predictor b SE z Constant 2.61 [2.07, 3.15] .27 9.49** Gender .07 [-0.39, 0.52] .23 .29 Agreeableness .30 [-0.16, 0.77] .24 1.29 Moral Concerns .73 [0.30, 1.16] .22 3.35** Sanction -.06 [-0.51, 0.40] .23 -.25 Note. *p < .05, **p < .01; †p< .10 R2 (Nagelkerke)= .14
Thereafter, the third measure of compliance, defined as passing or failing the first attention check, was analyzed through a logistic regression with sanction, and form of compensation as predictors, in order to test whether the form of compensation affected the model. The logistic regression was conducted with 95% bias corrected and accelerated confidence intervals based on 1,000 bootstrap samples. However, the test of the model was only marginally significant in step 1 (χ2= 5.17, p= .08) and insignificant in step 2 after including sanction and form of compensation as predictors into the model (∆χ2= 1.13, p= .54) as well as insignificant in step 3 including the interaction of sanction and form of compensation into the model (∆χ2= .01, p= .92), with df= 1 and Nagelkerke’s R2= .06 (R2= .04 for Step 1, ∆R2 = .02 for Step 2, ∆R2 = .00 for Step 3). The results indicated no relationship between prediction and grouping. Additionally, no further significant effects were found.
For low levels of the moderator conscientiousness there was a significant negative relationship between sanction and effort, β= -.32, t(245)= -3.48, p< .01, as well as for mean levels of conscientiousness (M= 3.44), β= -.22, t(245)= -3.60, p< .01. The negative effect of sanction decreased significantly due to increased levels of conscientiousness up to the point when participants’ conscientiousness scores were .71 above the mean (M= 3.44 + .71= 4.15),
β= -.15, t(245)= -1.97, p= .05. For people with higher levels of conscientiousness, sanction
did not decrease effort significantly anymore, β= -.13, t(245)= -1.39, p= .17. Figure 4 depicts the interaction.
In the next step, the dependent variable number of attention checks passed was analyzed (Table 9). However, no significant effects were found.
Table 7: Linear model of predictors of compliance (effort), with 95% bias corrected and accelerated confidence intervals reported in parentheses. Confidence intervals and standard
errors based on 1,000 bootstrap samples
Predictor Effort
b SE β p
Step 1
Constant 2.65 [2.13, 3.11] .25 p<.01**
Gender .36 [0.10, 0.65] .14 .16 p<.01**
Control Dummy Article 1 .25 [-.08, 0.57] .17 .11 p=.14
Control Dummy Article 2 .27 [-0.04, 0.59] .16 .12 p=.11 Step 2
Sanction -.23 [-0.36, -0.09] .07 -.20 p<.01**
Moral Appeal .08 [-0.05, 0.21] .07 .07 p=.25
Conscientiousness (centered) .13 [-0.01, 0.27] .07 .12 p=.07†
Step 3
Sanction x Moral Appeal .07 [-0.07, 0.20] .07 .06 p=.33
Sanction x Conscientiousness .13 [-0.02, 0.26] .07 .11 p=.07†
Moral Appeal x Conscientiousness .10 [-0.04, 0.24] .07 .09 p=.15
Step 4
Sanction x Moral Appeal x Conscientiousness .09 [-0.05, 0.23] .07 .08 p=.20
Note. *p < .05 (two-tailed), **p < .01 (two-tailed), †p< .10
Table 8: Regression results for the moderation effects of sanction and conscientiousness on effort, with 95% bias corrected and accelerated confidence intervals, based on 1,000
bootstrap samples
Conditional Effect for Sanction on Effort at values of Conscientiousness Moderator Effect Boot SE t Boot LLCI Boot ULCI -.99 (low level) -.32 .09 -3.48** -.49 -.14 .00 (average) -.22 .06 -3.60** -.34 -.10 .71 (point of sig.) -.15 .08 -1.97* -.31 .00 .99 (high level) -.13 .09 -1.39 -.30 .05 Note. *p < .05, **p < .01
Figure 4: Interaction plot of the moderation effect of sanction and conscientiousness on effort
Table 9: Linear model of predictors of compliance (attention checks), with 95% bias corrected and accelerated confidence intervals reported in parentheses. Confidence intervals
and standard errors based on 1,000 bootstrap samples
In the next step, the third dichotomous dependent variable of passing or failing the first attention check was analyzed through a logistic regression with moral appeal, sanction, and conscientiousness as predictors with 95% bias corrected and accelerated confidence intervals, based on 1,000 bootstrap samples (Table 10). The test of the model was marginally significant in step 1 (χ2= 5.17, p= .07), but not significant in step 2 (χ2= 5.37, p= .98) and step 3 (χ2= 6.49, p= .77), until step 4, which was significant (χ2= 10.89, p< .05) with df= 1 and Nagelkerke’s R2= .09 (R2= .04 for Step 1, ∆R2 = .01 for Step 2, ∆R2 = .01 for Step 3, ∆R2 = .03 for Step 4). Moreover, the two-way interactions were no significant predictors. However, the threeway interaction of sanction x moral appeal x conscientiousness was significant, β= -.52, p< .05 (Table 10). Figure 5 depicts the three-way interaction between sanction, moral appeal, and conscientiousness. It indicates that the probability of passing the first attention
Predictor Attention Checks
b SE β p Step 1 Constant 3.68 [2.77, 4.62] .46 p< .01** Gender -.10 [-0.29, 0.09] .10 -.06 p= .33 Agreeableness .28 [-0.01, 0.18] .12 .15 p< .05* Step 2 Sanction .01 [-0.10, 0.12] .06 .01 p= .84 Moral Appeal .00 [-0.11, 0.11] .05 .00 p=.94 Conscientiousness (centered) .08 [-0.01, 0.18] .05 .10 p= .12 Step 3
Sanction x Moral Appeal .07 [-0.03, 0.18] .05 .09 p= .16
Sanction x Conscientiousness -.06 [-0.15, 0.03] .05 -.08 p=.24
Moral Appeal x Conscientiousness -.06 [-0.16, 0.04] .05 -.07 p=.26
Step 4
Sanction x Moral Appeal x Conscientiousness -.07 [-0.17, 0.03] .05 -.09 p=.18
Note. *p < .05 (two-tailed), **p < .01 (two-tailed), †p< .10
check is the highest when there is no sanction, but when the moral appeal is present, and when participants score high on conscientiousness. However, when a sanction is present, the probability decreases when a moral appeal is present and when participants score high on conscientiousness. When the moral appeal is absent, the probability of success of both sanction present vs. sanction absent and high vs. low scores of conscientiousness is rather stable on a rather low level.
Table 10: Logistic regression of predictors of compliance (1st attention check), with 95% bias corrected and accelerated confidence intervals reported in parentheses. Confidence
intervals and standard errors based on 1,000 bootstrap samples
Predictor Odds Ratio (95% CI)
b Lower Odds Upper
Step 1 Constant -1.32 [-4.08, 1.66] . Gender -.20 [-1.20, 0.68] .34 .82 1.95 Agreeableness 1.07 [0.23, 1.94]* 1.20 2.91 7.71 Step 2 Moral Appeal .06 [-0.94, 1.07] .44 1.06 2.55 Sanction .07 [-0.91, 1.05] .45 1.07 2.58
Conscientiousness (cent red) .09 [-0.38, 0.55] .70 1.09 1.71 Step 3
Sanction x Moral Appeal .22 [-0.33, 1.15] .80 1.25 1.96
Sanction x Conscientiousness -.03 [-0.67, 0.59] .62 .97 1.54
Moral Appeal x Conscientiousness -.10 [-0.96, 0.60] .57 .90 1.44
Step 4
Sanction x Moral Appeal x Conscientiousness -.52 [-2.05, 0.21]* .36 .60 .98
Note. *p < .05 (two-tailed), **p < .01 (two-tailed), †p< .10
R2 (Nagelkerke)= .04 for Step 1, ∆R2= .01 for Step 2, ∆R2= .01 for Step 3, ∆R2= .03 for Step 4
Figure 5: Three-way interaction effect between sanction, moral appeal, and conscientiousness on probability of passing the first attention check
5. DISCUSSION
The present study aimed at investigating whether a moral appeal could make a sanction more effective at inducing compliance among people. The expectation was that when a sanction is combined with a moral appeal, it would more likely increase people’s moral concerns. I assumed that only when a moral appeal is present, moral concerns would further mediate the conditional indirect relationship of sanction and moral appeal on compliance. However, I hypothesized that a sanction alone would not increase moral concerns, but would be directly positively related to compliance. Nevertheless, this field experiment did not support these hypotheses. In fact, the effect of the sanction on moral concerns was not moderated by the presence of a moral appeal and hence moral concerns among participants were not strengthened. Moreover, there was no conditional indirect effect of sanction and moral appeal through moral concerns on compliance, neither on effort levels, nor on attention levels during the experiment. As far as the sanction did influence compliance, it had a negative rather than a positive effect. The sanction decreased levels of effort participants put into the summary task, but did not have any effect on attention levels of participants.
The current research showed that high levels of moral concerns among participants increased effort as well as the attention of participants. These findings underline the connection between moral considerations about people’s behavior and compliance. This can be explained by existing research on moral norms, because moral norms have been shown to be a strong motivator for people to show sustained morally right actions over time (Cialdini, Demaine, Sagarin, Barret, Rhoads, & Winter, 2006; Reno, Cialdini, & Kallgren, 1993; Cooter, 1998). Furthermore, it also shows that in case moral concerns are effectively strengthened, these moral concerns might ultimately increase compliance.
indeed a sanction is not effectively moderated by a moral appeal in order to induce moral concerns. In line with that, previous research on tax evasion has suggested that including a moral appeal to an enforcement system could cause people to think that the “system alone cannot cope and has to rely on rhetoric tricks instead”, which can cause people to reject the moral message altogether (Bardach, 1989: 62).
Another explanation why a moral appeal might not be effective at moderating the relationship between sanction and moral concerns is that research has shown that excessively framing a message in moral terms can also backfire and cause aversive behavior (Monin, 2007; Täuber & Van Zomeren, 2013; Täuber, Van Zomeren, & Kutlaca, 2015). For example, Mulder, Rupp and Dijkstra (2015) noted that overweight people showed unhealthier eating behavior after being exposed to a message, which claimed that obesity was a moral issue. So framing the sanction about taking the experiment seriously in moral terms might have caused participants to experience reactance towards the message. However, the current study did not reveal any negative interaction effect of sanction and moral appeal on compliance. If the moral appeal caused any negative psychological reactions, it did not become apparent in this case. It might have only rendered the moral message of the appeal ineffective.
disapproved of in society and that trespassing this norm is morally wrong (Cooter, 1998; McAdams, 2000). Hence, the sanction in this study might have stressed the moral norm of taking research experiments seriously, which would explain why participants in the sanction condition thought that a moral appeal was given to them at the beginning of the experiment.
leader (e.g. Mulder & Nelissen, 2010) and their effects upon participants’ acceptance of the sanction in combination with the moral appeal.
stressing the inherent moral message of behavioral disapproval by a moral appeal could increase the likelihood that moral concerns among people are increased.
The current study showed that the sanction was detrimental to participants’ effort but did not affect their attention levels. This could be explained by research about psychological reactance, which has noted that restricting behavior may lead to negative psychological reactions of people (Brehm, 1966). That is, when individuals feel that they are not closely monitored and can easily manipulate or bypass the system, people might show increased deviance (Brehm, 1966). Participants in this experiment may have concluded that the attention checks were monitoring their behavior and hence wanted to avoid the costly punishment, which may explain why the sanction did not decrease attention levels. However, participants might not have attributed the sanctioned behavior to writing the summary and felt that they were not monitored. Thus, the negative psychological reactions became apparent in the effort task. This is also in line with research about intrinsic motivation. Sanctioning systems have shown to communicate to the individual that the responsibility for assuring compliance rests with the system that is monitoring the behavior and not the individual (Pfeffer, 1994). As a result, individuals become less intrinsically motivated to show the desired behavior when they feel that the system is not closely watching (Cialdini, 1996; Frey, 1997; Lingle, Brock, & Cialdini, 1977). This could explain the decreased effort participants showed in the writing task.
the case that not compensating students for their participation in the experiment was judged as unreasonable and unjust. On the one hand, the financial costs might have been judged as too severe; on the other hand, students also participated for research points, which they needed to pass university courses. In the worst case, not getting compensated for doing the experiment would have meant for them not passing their course. This could have caused some participants to experience the sanction as unreasonable severe. Moreover, the results of this experiment also showed that the form of compensation, either money or research points, in combination with the sanction did not have any interaction effect upon participants’ compliance. Hence, it did not matter which form of compensation participants had received, they all decreased their effort similarly.
Nevertheless, the results revealed that personal characteristics of participants could buffer the detrimental effect of the sanction on effort. The higher participants scored on conscientiousness, the more was the negative effect of the sanction on shown effort in the writing task decreased. These findings indicate that participants with the inclination to be very conscientious were less likely to let their performance during the experiment suffer from negative psychological reactions towards the sanction. Previous research has noted that conscientiousness was a strong predictor for less deviant behavior and compliance with rules (Salgado, 2002) and also that conscientiousness was predictive of less workplace deviance, or counterproductive work behavior (Cohen, Panter, Turan, Morse &, Kim, 2014).
Furthermore, the positive relationship between moral appeal and conscientiousness was only apparent as long as there was no sanction present. The likelihood of passing the first attention check decreased severely when a sanction was present, even though participants scored high on conscientiousness. This underlines the suggestion that the detrimental effect of the sanction overruled the positive effect of a moral appeal and of high scores of conscientiousness. In this case, participants could have experienced strong negative psychological reactions, which might have been stronger than their inclination to be receptive to the moral message of the moral appeal.
Statistical Limitations
The weak points of this study are that the sanction manipulation check items failed to form a reliable scale, which could suggest that the manipulation itself was not fully successful. However, the separate tests on each item revealed consistent results that the sanction manipulation indeed showed significant effects. Additionally, the experiment was combined with another excessive study on health and HR decision making, which included its own manipulations into the experimental design. This may have caused the manipulations to become less effective over time, which might have caused the insignificant results. Another limitation is that the dichotomous variable of passing or failing the first attention check did not consist of equal groups (M= .91; SD= .29; Npassed =230, Nfailed= 23), indicting a so called
Practical Implications
This experiment has also practical implications, specifically for the FEB research lab in which this field experiment was conducted. The results suggest that installing a sanction as introductory message of research experiments might have rather detrimental effects on participants’ effort and attention than positive. As a result, this could even worsen the experimental data gathered in the research lab. A moral appeal was further not found to be effective at framing the sanction in a retributive way and hence moral concerns among participants about taking the research experiment seriously were not evoked. Hence it might be reasonable to refrain from installing a sanction in combination with a moral appeal altogether.
Furthermore, this experiment showed that the personality trait conscientiousness buffered the detrimental effect of the sanction on displayed effort. This might be interesting for organizations, as personality testing is a common tool in selection procedures (e.g. Ones, Dilchert, Viswesvaran & Judge, 2007). It could hence be beneficial to focus on hiring individuals with high levels of conscientiousness, as this character trait might buffer a suboptimal installation of a sanctioning system and also limit the possibility of deviant reactions towards the system.
Conclusion
6. REFERENCES
Aquino, K., & Reed, A. 2002. The self-importance of moral identity. Journal of Personality and Social Psychology, 83(6): 1423–1440.
Balliet, D., Mulder, L. B., & van Lange, P. 2011. Reward, punishment and cooperation: A meta-analysis. Psychological Bulletin, 4: 594–615.
Bardach, E. 1989. Moral Suasion and Taxpayer Compliance. Law and Policy, 11: 49–69. Baron, J. 2000. Thinking and Deciding (3rd ed.). New York: Cambridge University Press. Bennett, R. J., & Robinson, S. L. 2000. Development of a measure of workplace deviance.
Journal of Applied Psychology, 85: 349 –360.
Bicchieri, C. 2006. The grammar of society: The nature and dynamics of social norms. New York: Cambridge University Press.
Bilz, K., & Nadler, J. 2009. Law, psychology, and morality. Moral Judgment and Decision Making, 50:101–131.
Blasi, A. 1983. Moral cognition and moral action: A theoretical perspective. Developmental Review, 3(2): 178–210.
Blasi, A. 1984. Moral identity: Its role in moral functioning. In W. Kurtines & J. Gewirtz (Eds.), Morality, moral behavior and moral development. New York: Wiley: 128–139. Blumenthal, M., Christian, C., & Slemrod, J. 2001. Do Normative Appeals Affect Tax
Compliance? Evidence from a Controlled Experiment in Minnesota. National Tax Journal, 54(1): 125–138.
Bowling, N. A., & Beehr, T. A. 2006. Workplace harassment from the victim’s perspective: A theoretical model and metaanalysis. Journal of Applied Psychology, 9: 998–1012.
Brehm, S. S. & Brehm, J. W. 1981. Psychological Reactance: A Theory of Freedom and Control. New York: Academic Press.
Caldwell, M. D. 1976. Communication and sex effects in a five-person prisoners dilemma game. Journal of Personality and Social Psychology, 33: 273–280.
Chen, X. P., Pillutla, M. M., & Yao, X. 2009. Unintended Consequences of Cooperation Inducing and Maintaining Mechanisms in Public Goods Dilemmas: Sanctions and Moral Appeals. Group Processes & Intergroup Relations, 12(2): 241–255.
Chiaburu, D. S., & Harrison, D. A. 2008. Do peers make the place? Conceptual synthesis and meta-analysis of coworker effects on perceptions, attitudes, OCBs, and performance. Journal of Applied Psychology, 93: 1082–1103.
Cialdini, R. B. 1996. Social influence and the triple tumor structure of organizational dishonesty. In D. M. Messick & A. E. Tenbrunsel (Eds.), Codes of conduct. Behavioral research into business ethics: 44–58. New York: Russell Sage Foundation.
Cialdini, R. B., Demaine, L., Sagarin, B. J., Barret, D. W., Rhoads, K., & Winter, P. L. 2006. Managing social norms for persuasive impact. Social Influence, 1(1): 3–15.
Cooter, R. 1998. Expressive law and economics. Journal of Legal Studies, 27(2): 585–608. Cohen, T.R., Morse, L. 2014. Moral character: what it is and what it does. Research in
Organizational Behavior, 34: 43-61.
Cohen, T. R, Panter, A.T, Turan, N., Morse, L. & Kim, Y. 2014. Moral character in the workplace. Journal of Personality and Social Psychology, 107: 943-963.
Dal Bó, E., & Dal Bó, P. 2014. Do the Right Thing: The Effects of Moral Suasion on Cooperation. Journal of Public Economics, 17: 28–38.
Darley, J. M., & Pittman, T. S. 2003. The Psychology of Compensatory and Retributive Justice. Personality and Social Psychology Review, 7(4): 324–336.
Ellsworth, P. C. 2010. The Rise and Fall of the High Impact Experiment. In, M. H. Gonzales, C. Tavris, & J. Aronson (Eds), The Scientist and the Humanist: A Festschrift in Honor of Elliot Aronson, New York: Psychology Press: 79-106.
Eurostat. 2013. Trends in crime and criminal justice, 2010. Statistics in focus, 2013(18).
http://ec.europa.eu/eurostat/statistics-explained/index.php/Crime_trends_in_detail#Crimes_recorded_by_the_police [Accessed on 26.05.2016].
Fehr, E., & Gächter, S. 2000. Cooperation and punishment in public goods experiments. American Economic Review, 90(4): 980–994.
Feldman, Y. 2009. The Expressive Function of Trade Secret Law: Legality, Cost, Intrinsic Motivation, and Consensus. Journal of Empirical Legal Studies, 6(1): 177–212.
Fellner, G., Traxler, C., & Sausgruber, R. 2013. Testing enforcement strategies in the field: Threat, moral appeal and social information. Journal of the European Economic Association, 11(3): 634–660.
Frey, B. 1997. A constitution for knaves crowds out civic virtue. The Economic Journal, 107: 1043- 1053.
