PUNISHMENT OF CHILDREN AND ADULTS 1
A person-centered approach to punishment: Reprimand and avoidance are differentially directed
towards adult and child transgressors
Cindel J.M. White1, Mark Schaller1, Elizabeth G. Abraham2, Joshua Rottman3
1 Department of Psychology, University of British Columbia, 2 Department of Teaching &
Learning, Florida International University, 3 Department of Psychology, Franklin & Marshall
College
Author Note
This work was supported by a University of British Columbia Faculty of Arts Graduate Research Award, and by funds from the Franklin & Marshall College Committee on Grants, and the Social Sciences and Humanities Research Council of Canada (Insight Grant 435-2020-1278). We thank Lysa Adams for illustrating the transgressions and Kelly Minard for research assistance. Correspondence concerning this article should be addressed to Cindel J. M. White, Department of Psychology, University of British Columbia, 2136 West Mall, Vancouver, BC V6T 1Z4. Email: [email protected]
Declarations of interest: none PUNISHMENT OF CHILDREN AND ADULTS 2
Abstract
Three studies (N = 867) investigated how adults’ and children’s punitive responses to moral transgressions differ depending on whether transgressors are adults or children. Adults judged the transgressions of fellow adults as substantially more wrong, and as more worthy of avoidance and punishment, than identical actions performed by children. This difference was partially mediated by the perception that adults’ actions are considered to be more wrong, more harmful, and stranger than children’s identical actions, and by greater anxiety about the negative consequences of confronting adults about their bad behavior. Despite viewing children’s actions as less wrong, adults were more likely to reprimand children than adults who engaged in identical behavior, and this difference became more pronounced when statistically controlling for the wrongness and strangeness of actions. Adults’ nurturant tendencies towards children, as well as their perceptions of children’s moral character as more changeable, also predicted relatively greater reprimand and less avoidance of child transgressors. These differences between reprimand and avoidance of child and adult transgressors was robust to the type of transgression
(including harm- and purity-related norms), several individual differences, and a global pandemic. In contrast, 4- to 9-year-old children were equally likely to avoid and reprimand adult and child transgressors, suggesting that different processes are engaged when adults judge children compared to when children evaluate their own peers. Together, these findings indicate how diverse responses to moral transgressions are differentially adapted for norm violators of different ages.
Keywords: Moral Judgment; Punishment; Character; Avoidance; Reprimand; Children PUNISHMENT OF CHILDREN AND ADULTS 3
A person-centered approach to punishment: Reprimand and avoidance are differentially directed
towards adult and child transgressors
1 Introduction
Punishment of norm violations is widespread across human societies. Responses to transgressions come in many varieties, such as physical aggression and other forms of retaliation that directly impose a penalty on transgressors, confrontation and rebukes that aim to encourage transgressors to atone for their misdeeds, and exclusionary tactics (e.g., avoidance, gossip) that indirectly punish transgressors by lowering their social standing (Boehm, 2012; Cushman, 2015).
What determines how people will respond to a moral transgression? This paper presents three studies demonstrating that the age of the observer and the age of the transgressor—specifically, whether they are an adult or a child—is an important determinant of how people will respond to norm violations. We show that, even when performing identical actions with identical outcomes, adults’ and children’s transgressions elicit distinct appraisals and emotional responses, resulting in differing desires to reprimand, avoid, and punish. These results are consistent with broad patterns of functional differences between the relationships that adults have with children and with other adults.
1.1 Why do people punish?
Punishment signals the condemnation of a transgressor’s action, and thus regulates interpersonal relationships and enforces social norms (Cushman et al., 2019; Ho et al., 2019;
Kriss et al., 2016). Thus, punishment can benefit individuals and groups by encouraging future prosocial behavior (Boyd et al., 2003; dos Santos & Wedekind, 2015; Marlowe & Berbesque,
2008) and deterring future harm to oneself (Delton & Krasnow, 2017; Krasnow et al., 2012,
2016). Publicly punishing norm violators can also have reputational benefits for the punisher, PUNISHMENT OF CHILDREN AND ADULTS 4 when it communicates that the punisher disapproves of the transgression and values prosocial, normative conduct (Jordan et al., 2016; Jordan & Rand, 2019). These goals can be variously accomplished by relying on many different varieties of direct and indirect punishment.
Choosing which response is most appropriate in a given situation can be partly explained by features of the transgressive action itself. More severe norm violations—those perceived to be more harmful, more intentionally malicious, and generally more wrong—are likely to elicit greater punishment of all sorts (Boehm, 2014). Different types of actions may also elicit different types of responses. For example, actions that are harmful, anger-provoking, and a potential direct threat to oneself may be more likely to elicit direct confrontation and physical aggression towards the transgressor, as the most direct way to deter future transgressions against oneself, whereas violations of standards of bodily and spiritual purity, and other disgust-eliciting actions, may be more likely to elicit avoidance as a means to prevent contamination (physical and metaphorical) by the offender (Kemper & Newheiser, 2018; Kupfer & Giner-Sorolla, 2017;
Rottman et al., 2018; Yoder et al., 2016). However, there is good reason to think that features of the transgressor and features of the potential punisher are also important determinates of how people will respond to norm violations, due to the relative costs and benefits of different types of punishment. Just as moral judgments about goodness and badness can be understood as judgments of persons, rather than simply judgments of actions and their consequences (Uhlmann et al., 2015), judgments about the appropriate form for punishment can be also understood by considering the identities of and relationships between punishers and the people who commit moral transgressions. PUNISHMENT OF CHILDREN AND ADULTS 5
1.2 Person-centered influences on direct and indirect punishment
Direct punishment—including confronting the transgressor, telling them that they did something wrong, and imposing costs for misbehavior in the form of physical violence, financial sanctions, or other withholding of resources—is often the most effective way to ensure that the transgressor (a) knows that they did something wrong and (b) is disincentivized to offend in the future (Delton & Krasnow, 2017; Krasnow et al., 2012, 2016). Therefore, past research has found that direct punishment (and approach-oriented emotions like anger) is most likely to be deployed when people are most motivated to change the transgressor’s behavior (Tybur et al.,
2019), such as when transgressions were committed against oneself (Molho et al., 2017, 2020;
Pedersen et al., 2018) or one’s kin (Lopez et al., 2019) rather than strangers. Direct punishment also poses a greater risk to the punisher, if the transgressor decides to retaliate. Fear of retaliation deters confrontation of norm violators (Balafoutas & Nikiforakis, 2012) and direct punishment is therefore more likely when punishers have more power and transgressors have less power (Molho et al., 2020), including factors such as social status and physical strength, such as when punishers are men rather than women (Balafoutas & Nikiforakis, 2012) and when norm violators are women rather than men (Balafoutas et al., 2014).
Many of the costs of direct punishment can be mitigated by instead engaging in indirect forms of punishment, such as avoiding the transgressor and excluding them from one’s social network. Indirect punishment (and avoidance-related emotions such as disgust) have therefore been found to be more common when punishers have less power and value offenders less
(Molho et al., 2020; Tybur et al., 2019), such as when responding to transgressions against third parties rather than oneself or kin (Lopez et al., 2019; Molho et al., 2017; Pedersen et al., 2018), when the punisher places less value on their relationship with the transgressor, and when there is PUNISHMENT OF CHILDREN AND ADULTS 6 a high threat of future exploitation from the transgressor (Burnette et al., 2012; McCullough et al., 2013, 2013; Smith et al., 2019). Avoidance and expressions of disgust towards the transgressor therefore provide protection from the transgressor, by limiting future social interactions, while additionally serving as a low-cost signal of disapproval of the transgressor’s behavior (Kupfer & Giner-Sorolla, 2017; Yoder et al., 2016).
1.3 Differences between responses to adults’ and children’s norm violations
These descriptions of flexibility in functional forms of punishment imply differences in response tendencies for transgressions committed by children as compared to adults. Previous research has investigated variation in moral value according to age (Goodwin & Landy, 2014; Li et al., 2010), indicating that transgressive actions are differentially appraised depending on whether they are performed by adults or children. Furthermore, adults’ norm transgressions are likely to be perceived as weirder, more harmful, more deliberate and under voluntary control, more anger- and disgust-evoking than children’s norm transgressions (White & Schaller, 2018), implying that adults’ actions are more likely to be punished than identical actions performed by children. The precise form of punishment (reprimand vs. avoidance) may then depend on what type of action was performed (e.g., a violation of harm and fairness norms vs. purity norms), regardless of who performed the action. However, as described above, the different considerations underlying direct and indirect punishment imply further divergences between responses to adults and children. Specifically, we predict that children are more likely than adults to be reprimanded for their transgressions, while adults are more likely than children to be avoided after transgressing. Because directly confronting and reprimanding a transgressor can be effective but poses a greater risk of retaliation than does avoidance and other indirect forms of punishment, adult observers are likely to perceive young children as generally less threatening PUNISHMENT OF CHILDREN AND ADULTS 7 than other adults, therefore reducing the risk of reprimand and the benefits of avoiding children.
Reprimanding children may also be viewed as typically more effective than reprimanding adults, due to the perception that children’s actions were less intentionally-transgressive and that their moral character is undeveloped and more malleable (White et al., 2020; White & Schaller, 2018) than the moral character of adult transgressors. This may be especially true when controlling for the perceived wrongness of actions: When the wrongness of children’s actions is equated with the wrongness of adult’s actions, people may be especially likely to reprimand children more than adults.
In addition to these patterns based on the transgressor’s age, features of the observer should moderate tendencies toward reprimand and avoidance. For example, individuals who are generally high in interaction anxiety or disgust sensitivity may be more likely to avoid both adult and child transgressors, and individuals who are dispositionally nurturant and protective towards children may be less likely to avoid and more likely to reprimand child transgressors.
Responses are also likely to differ depending on whether observers are adults or children themselves. For example, when children misbehave, adults may be more likely to give them a break, whereas children may view fellow children and adults as both capable of controlling their actions and responsible for behaving well, meaning that both children and adults deserve similar punishments for the same behavior. Additionally, to the extent that punishment functions to ensure partner control and inform partner choice, it may be preferentially directed toward members of one’s peer group (who serve as potential cooperative partners). Thus, condemnation may be elevated when transgressors are similar in age to the punisher. PUNISHMENT OF CHILDREN AND ADULTS 8
1.4 Overview of studies
In three studies, we test how punitive responses to norm violations differ depending on features of transgressive actions, features of transgressors, and features of observers. Study 1 tested how adult participants respond to adult transgressors compared to child transgressors who engage in violations of harm-related and purity-related norms. (Study 1 also tested the replicability of primary results before and after the onset of a global pandemic.) Study 2 extended these findings to a new sample of adult participants and tested whether 4- to 9-year-old child participants also respond differently to child vs. adult transgressors. Study 3 tested for several mediators that could explain why adult participants respond differently to adult and child transgressors. Hypotheses, methods, and analysis plans were preregistered on the Open Science
Framework prior to data collection
(https://osf.io/mbn3y/?view_only=0a238742af9a4a5aad04fbab6d7b92e5), and all data and analysis scripts are available at the same link. In the Methods and Results sections (or in accompanying Supplementary Materials) we report all manipulations and measures completed by participants, as well as preregistered sample sizes and preregistered data exclusion decision- rules. In addition, we identify any deviations from preregistered analysis plans.
2 Study 1
In Study 1, adults evaluated a series of transgressions—including actions that did or did not cause harm to another person, and actions that did or did not involve potential pathogens— that were described as being performed by children (4–9 years old) or adults (25–50 years old).
Participants evaluated the transgressor’s actions and then reported their willingness to actively punish, directly reprimand, or avoid the transgressor. This study established general differences between how adults respond to adults’ and children’s transgressions, and demonstrated that PUNISHMENT OF CHILDREN AND ADULTS 9 evaluations of actions alone (e.g., the greater strangeness of adult’s norm violations) or evaluations of actors alone (e.g., the greater capacity for adults to control their behavior) are insufficient to explain the different methods of punishing adult and child transgressors. Instead, information about transgressions and transgressors are evaluated in tandem to produce a unique punitive response.
2.1 Methods
2.1.1 Participants
Timepoint A.
We planned to recruit a sample of approximately 350 adult participants, located in the
United States, from Amazon’s Mechanical Turk (MTurk). Based on power analyses via simulation conducted using the SIMR package in R (Green & MacLeod, 2016), a sample size of
350 should be able to detect reasonably-small two-way (b = .20) or three-way interactions (b
= .25) between transgressor age and vignette types with >85% power. To increase power, all variables were manipulated within-subjects. According to additional power analyses conducted using the pwr package in R, this sample size should also have enough power to detect reasonably-small correlations (r = 0.15) between variables of interest with >80% power. Based on preregistered criteria, we excluded 98 participants who failed one or more attention check criteria (inability to recall one thing that an adult or child did in one of the scenarios; incorrect response to the question “Please select Always as your answer to this question” or a CAPTCHA; participants who reported that they did not pay attention, did not take the survey seriously, or were otherwise distracted while completing the survey). This resulted in a final sample size of 331 participants, 51% of whom were parents (parental status did not moderate any of the effects described below; all sample demographics are provided in Supplementary Materials). All data were collected in April 2019. PUNISHMENT OF CHILDREN AND ADULTS 10
Timepoint B: Longitudinal follow-up.
All participants from Timepoint A were asked to complete this survey again one year later, in March/April of 2020, to assess whether the pattern of results changed after the novel coronavirus (COVID-19, an infectious respiratory disease) was declared a global pandemic. On the date that this survey was launched, most schools and other public gatherings in the United
States had recently been closed, and COVID-19 cases had been rapidly spreading in the USA and around the world. These numbers increased exponentially over the course of running the survey, with the USA eventually having more confirmed cases than any other country in the world.
Timepoint B included 157 participants who responded at both timepoints, and did not significantly differ from the entire Time 1 sample in terms of gender, age, education, income, parenthood, or dispositional parental tenderness or disgust sensitivity. In general, Timepoint B was moderately worried that they would get sick themselves or make someone else sick because of COVID-19 (M = 52.71, SD = 27.85, on a scale from 0 - 100), and 76% of participants reported that travel plans, special events, or their child’s school had been cancelled, although very few participants had been infected themselves (n = 2), or had friends or family members who tested positive (n = 10). This sample therefore allows us to test whether our results are replicable in a context where the threat of potential disease is highly salient, and in which public health officials recommend avoidance of strangers and people who might be infected with
COVID-19. This longitudinal follow-up was not preregistered, but the procedure was identical to that pre-registered for Timepoint A. PUNISHMENT OF CHILDREN AND ADULTS 11
2.1.2 Materials and Procedure
Vignettes describing transgressions
Each participant evaluated eight different vignettes depicting transgressions, that manipulated within-subjects (a) whether the transgression was committed by an adult (aged 25,
30, 35, 40, 45, or 50 years old) or a child (aged 4, 5, 6, 7, 8, or 9 years old), (b) whether the transgression involved potential pathogens (e.g., “While David is alone at home, he smears a handful of his own poop all over his bathroom wall”) or did not involve pathogens (e.g., “Brian says mean things about how an overweight person looks”), and (c) whether the transgression involved a victim (e.g., “Jennifer passes some food to her friend with her bare hands, even though she hasn't washed her hands all week long”) or no victim (e.g., “When camping, Emily draws a picture and writes her name on a big, pretty rock with permanent markers”) who might be affected by the transgression. The presence of pathogens and potential victims were manipulated orthogonally to create 4 different transgression types: Pathogen and no victims, victims and no pathogens, both pathogens and victims, and neither pathogens nor victims.
Vignettes were chosen from a broader pool of 7 scenarios per condition (each depicting a different target individual), the order of presentation of vignettes was randomized across participants, and the order of items was randomized within each type of question. All stimuli are available in the Supplementary Materials.
Evaluations of transgressions and transgressors
Participants made several evaluations of each vignette, all on 5-point rating scales. These questions were: “how wrong is [Transgressor]’s behavior?”, “how angry does this make you feel?”, “how grossed out does this make you feel?” and “do you think that [Transgressor] is usually a good person or a bad person?” These items were combined into a composite measure PUNISHMENT OF CHILDREN AND ADULTS 12 of action wrongness (Cronbach’s α = .73). Participants then rated whether the transgressor has various mental capabilities (is able to “do things on purpose,” “come up with smart ideas,” “tell right from wrong,” “understand how others are feeling,” “feel pain,” and “feel joy,” composite: α
= .89).
Punishment of transgressors
Participants next reported how they would respond to the transgression. Based on exploratory factor analyses (see Supplementary Materials), these responses were grouped into three composite variables: Two items assessed reprimand of the transgressor (“I would try to explain to him [her] why it is better to behave well,” “I would try to explain to him [her] why his behavior was wrong”, α = .90), two items assessed vengeance (“I would yell, shout, or scream at him [her]”, “I would want to slap or hit him [her]”, α = .82), and four items assessed avoidance of the transgressor (“I would try to stay as far away from him [her] as possible,” “I would not want to be his [her] friend,” “I would feel embarrassed to have lunch with him [her],”
“I wouldn’t want to be on his [her] team when playing a game,” α = .92).
Transgression strangeness
Two items assessed perceptions of the transgression’s strangeness: “How strange would it be for other children [adults] to do what [Transgressor] did?” and “Is [Transgressor] behaving very differently from most people her age?” (composite α = 0.95).
Individual difference measures
After evaluating the vignettes, participants completed measures of individual differences in the tendency to protect and nurture children (Hofer et al., 2018; composite α = .89) and individual differences in disgust sensitivity (adapted from Viar-Paxton et al., 2015, a measure that was designed for child participants that we used here to be consistent with Study 2, that PUNISHMENT OF CHILDREN AND ADULTS 13 includes child participants; composite α = .74). Finally, they reported other demographic characteristics.
Additionally, at Timepoint B, participants reported their perceived vulnerability to disease in general (Duncan et al., 2009, composite α = .85), and their specific worry about getting sick because of COVID-19 or passing it onto others, and how their lives have recently been impacted by the coronavirus pandemic.
2.2 Results
2.2.1 Timepoint A: Main Preregistered Analyses
General analysis strategy
All analyses were performed as multi-level models that included random intercepts nested within participant, and random intercepts nested within vignette, to account for the repeated-measures nature of the data.1 For these analyses, transgressor age was dummy coded (0
= child transgressor, 1 = adult transgressor) and the presence of a victim or potential pathogens was contrast coded (-0.5 = no victim/pathogens, 0.5 = victim/pathogens). Mental capabilities and strangeness were centered within-participants, and then standardized, and disgust sensitivity and PCAT were standardized across participants.
Separate regressions were performed predicting composite measures of (a) transgression wrongness, (b) avoidance, (c) vengeance, and (d) reprimand. We first (Model 1) investigated differences across the experimentally-manipulated conditions in each vignette, by predicting each dependent variable from transgressor age, pathogen presence, and victim presence, plus all
1 We only preregistered including random intercepts by participant, but we added in random intercepts by vignette to make the analysis consistent with the analysis plan used in Studies 2 and 3. The patterns of results were similar – in magnitude and in inferential implications – for either specification of the models. PUNISHMENT OF CHILDREN AND ADULTS 14 interactions between these variables. These models also controlled for disgust sensitivity,
PCAT, and the interaction between PCAT and transgressor age, to account for individual differences in the tendency to respond to disgust-eliciting stimuli and to respond positively to children (the pattern of results is similar if these variables are not included in the model). Figure
1 displays means across conditions predicted by this model. Second (Model 2), mental capabilities and strangeness were added as covariates to this primary model, to assess the extent to which these variables can account for the differences between responses to adults’ and children’s transgressions. Full results from these models are depicted in Table 1 and depicted in
Figure 1, and simple effects derived from this model are described in the main text.
Model 1: Condition differences, controlling for individual differences
In general, children’s actions were rated as substantially less wrong than adults’ actions, b
= 0.50, 95% CI [0.44, 0.56], p < .001. This difference between judgments about children and adults was a large, statistically significant effect for each of the four transgression types (bs >
0.35), and for each of the four items that made up the composite wrongness measure (wrongness, bad person, grossness, and anger, bs > 0.39, ps < .001).
The preferred punitive responses to transgressors also differed regarding children’s and adults’ actions. As depicted, in Figure 1 and Table 1, participants were substantially more likely to avoid adult transgressors than child transgressors, all ps <.001, and this effect was slightly larger when the transgression included pathogens (with victim: b = 1.00 [0.86, 1.14], without victim: b = 1.13 [0.98, 1.27]) than when not including pathogens (with victim: b = 0.94 [0.79,
1.08], without victim: b = 0.84 [0.70, 0.99]). Participants were also more likely to seek vengeance towards adult transgressors than child transgressors, all ps <.006, and this effect was slightly larger when the transgression included a victim (with pathogens: b = 0.41 [0.29, 0.53], PUNISHMENT OF CHILDREN AND ADULTS 15 without pathogens: b = 0.37 [0.25, 0.50]) than when not including a victim (with pathogens: b =
0.17 [0.05, 0.29], with pathogens: b = 0.32 [0.20, 0.44]). In contrast, participants were substantially less likely to reprimand adults compared to child transgressors (ps < .001), and this difference did not significantly differ across different types of transgressions (ps > .05).
PUNISHMENT OF CHILDREN AND ADULTS 16
Figure 1. Mean tendency to avoid, yell at or hit (vengeance), and reprimand adult and child transgressors, for each type of transgression. Means, and 95% CIs, are predicted from the multi-level models that include random intercepts nested within participants and within vignette scenario, and that control for individual differences in PCAT and disgust sensitivity. A similar pattern is observed if only raw means are plotted. Transgression type: P = pathogens, no victim; P&V = pathogens and victim; V = victim, no pathogens; NP/NV = neither pathogens nor victim.
PUNISHMENT OF CHILDREN AND ADULTS 17
Table 1. Regression model predicting responses to transgressors from experimental conditions, individual differences, mental capabilities and perceived strangeness of the transgression. Avoidance Avoidance Vengeance Vengeance Reprimand Reprimand
+ Mind + Strangeness + Mind + Strangeness + Mind + Strangeness
b [95% CI] p b [95% CI] p b [95% CI] p b [95% CI] p b [95% CI] p b [95% CI] p
Intercept 2.34 2.58 1.66 1.77 4.07 4.16 <.001 <.001 <.001 <.001 <.001 <.001 [2.20, 2.47] [2.46, 2.69] [1.55, 1.78] [1.66, 1.88] [3.95, 4.20] [4.04, 4.28] Transgressor Age 0.98 0.50 0.32 0.11 -0.56 <.001 -0.73 <.001 <.001 <.001 .003 <.001 [0.91, 1.05] [0.43, 0.57] [0.26, 0.38] [0.04, 0.17] [-0.63, -0.49] [-0.82, -0.65] Pathogens 0.26 -0.06 0.06 .50 -0.09 0.10 -0.02 .026 .53 .20 .32 .82 [0.03, 0.49] [-0.24, 0.12] [-0.11, 0.22] [-0.23, 0.05] [-0.10, 0.31] [-0.22, 0.18] Victims 0.04 0.25 0.03 0.13 0.08 0.16 .76 .007 .69 .064 .43 .10 [-0.19, 0.27] [0.07, 0.43] [-0.13, 0.20] [-0.01, 0.27] [-0.12, 0.29] [-0.03, 0.36] Disgust Sensitivity 0.16 0.16 0.04 0.04 0.11 0.10 <.001 <.001 .31 .34 .009 .011 [0.09, 0.24] [0.08, 0.24] [-0.04, 0.13] [-0.04, 0.13] [0.03, 0.19] [0.02, 0.18] PCAT -0.19 -0.18 -0.01 0.00 0.09 0.09 <.001 <.001 .86 .95 .056 .041 [-0.27, -0.10] [-0.26, -0.09] [-0.10, 0.08] [-0.09, 0.09] [-0.00, 0.17] [0.00, 0.18] Transgressor Age * 0.18 0.01 -0.06 -0.14 0.03 -0.05 .010 .92 .33 .015 .73 .48 Pathogens [0.04, 0.32] [-0.11, 0.12] [-0.18, 0.06] [-0.26, -0.03] [-0.12, 0.17] [-0.19, 0.09] Transgressor Age * 0.00 -0.04 0.15 0.13 0.14 0.13 .96 .54 .017 .029 .055 .078 Victims [-0.13, 0.14] [-0.15, 0.08] [0.03, 0.27] [0.01, 0.24] [-0.00, 0.29] [-0.01, 0.27] Pathogens * -0.43 -0.06 -0.31 -0.13 -0.31 -0.13 .068 .75 .065 .37 .14 .50 Victims [-0.89, 0.03] [-0.42, 0.30] [-0.64, 0.02] [-0.40, 0.15] [-0.72, 0.10] [-0.53, 0.26] Transgressor Age * 0.18 0.16 0.11 0.1 0.07 0.06 <.001 <.001 <.001 .001 .058 .084 PCAT [0.11, 0.25] [0.11, 0.22] [0.05, 0.17] [0.04, 0.16] [-0.00, 0.14] [-0.01, 0.13] Transgressor Age -0.18 -0.34 0.19 0.12 0.20 0.16 * Pathogens .19 .005 .13 .31 .17 .28 [-0.46, 0.09] [-0.57, -0.10] [-0.06, 0.43] [-0.11, 0.35] [-0.09, 0.49] [-0.13, 0.44] * Victims Mental capabilities -0.12 -0.09 -0.12 <.001 <.001 <.001 [-0.15, -0.09] [-0.12, -0.06] [-0.16, -0.08] Strangeness 0.55 0.25 0.21 <.001 <.001 <.001 [0.51, 0.59] [0.21, 0.29] [0.16, 0.26] Marginal R2 / .205 / .458 .340 / .612 .032 / .469 .078 / .518 .081 / .364 .113 / .400 Conditional R2
Note. Transgressor age was dummy coded (0 = child transgressor, 1 = adult transgressor, such that estimates in Table 1 reflect the values for child transgressors, with transgressor age interactions indicating the change in these values when the transgressor was an PUNISHMENT OF CHILDREN AND ADULTS 18 adult) and the presence of a victim or potential pathogens was contrast coded (-0.5 = no victim/pathogens, 0.5 = victim pathogens, such that estimates indicate the average effect across both conditions). All models also include random intercepts nested within participants. PUNISHMENT OF CHILDREN AND ADULTS 19
Model 2: Controlling for mental capabilities and strangeness
To investigate possible explanations for these differences between responses of adult and child transgressors, the perceived strangeness of the action and the mental capabilities attributed to the transgressor were added into the regression model. When mental capabilities and strangeness were added to the model, the transgressor age differences in avoidance and vengeance were substantially reduced. To further investigate this effect, mental capabilities and strangeness were investigated as mediators of the transgressor age effect on responses, using an exploratory multilevel structural equation model in lavaan to calculate the indirect/total mediated effects (this model includes all of the same covariates and random intercepts by participant that were present in Model 1).
The mediation model predicting avoidance is depicted in Figure 2. Adult transgressors were perceived to possess greater mental capabilities and their transgressions were viewed as much stranger than were children’s transgressions. Greater perception of mental capabilities predicted less avoidance, resulting in an overall negative indirect effect, b = -0.019, 95% CI = [-
0.03, -0.01], total effect: b = 0.50 [0.42, 0.57]. That is, transgressor age predicted less avoidance via mental capabilities, which therefore cannot explain why adult transgressors were avoided more than child transgressors. Transgression strangeness also predicted greater avoidance, and the greater strangeness of adult’s transgressions did partially mediate the association between transgressor age and avoidance, indirect effect: b = 0.48 [0.43, 0.53], total effect: b = 0.99 [0.93,
1.06]. However, a large tendency to avoid adults more than children remained even after accounting for transgression strangeness, indicating other untested factors leading to greater avoidance of adults.
PUNISHMENT OF CHILDREN AND ADULTS 20
Figure 1. Coefficients from the path model testing whether the greater tendency to avoid adult, compared to child, transgressors is mediated by the perceived strangeness of the actions and mental capabilities of the transgressor. The model also controls for whether the transgression involved pathogens or a victim, participants’ level of disgust sensitivity and PCAT, and random intercepts by participant, as in the primary preregistered models.
0.90*** Strangeness 0.53***
Adult (1) vs. 0.52*** Child (0) Avoidance Transgressor
0.14*** Mental Capabilities -0.13***
Mental capability attribution also did not explain the elevated inclinations to enact vengeance against adult transgressors or the elevated tendencies to reprimand child transgressors.
Greater mental capabilities predicted less willingness to seek vengeance toward the transgressor, b = -0.10, p < .001, but did not explain much of the relationship between transgressor age and vengeance tendencies, indirect effect: b = -0.01 [-0.02, -0.01], total effect: b = 0.09 [0.02, 0.16].
Greater mental capabilities also predicted less reprimand, b = -0.13, p < .001, but did not explain much of the relationship between transgressor age and reprimand tendencies, indirect effect: b =
-0.02 [-0.03, -0.01], total effect: b = -0.73 [-0.82, -0.65]. Therefore, differences in mental capability attribution do not explain the differences in responses to adult and child transgressors.
Perceived strangeness of the transgression predicted greater willingness to enact vengeance against adults, b = 0.25, p < .001, as well as the reduced tendency to reprimand adults, b = 0.20, p < .001. Strangeness partially mediated the greater willingness to yell at or hit adult transgressors, indirect effect, b = 0.22 [0.18, 0.26], total effect: b = 0.33 [0.26, 0.39], indicating PUNISHMENT OF CHILDREN AND ADULTS 21 that the greater strangeness of adults’ transgressions is part of the reason why participants were more inclined to yell at or hit adults. However, strangeness did not explain the tendency to reprimand children more than adults, as indicated by the positive indirect effect of transgressor age on reprimand via strangeness, indirect effect: b = 0.18 [0.13, 0.23], total effect: b = -0.54 [-
0.62, -0.46]. In other words, adults’ actions were stranger, and stranger actions were more likely to be reprimanded, but this cannot explain why participants were especially likely to reprimand child transgressors. When controlling for the strangeness of the action, participants were especially likely to report that they would reprimand children more than adults.
2.2.2 Timepoint B: Longitudinal follow-up
All participants were invited to complete the survey again in the spring of 2020, to explore whether the pattern of responses to norm violations (including some involving potential pathogens) differed amid the heightened salience of pathogen threats due to the COVID-19 pandemic. In this sub-sample, the overall pattern of results replicated the patterns of responses at the first timepoint: Participants were much more likely to avoid, b = 0.81 [0.71, 0.91], more likely to seek vengeance toward, b = 0.28 [0.20, 0.37], and less likely to reprimand, b = -0.63 [-
0.73, -0.52], adults compared to child transgressors, ps < .001. Additionally, we replicated the findings from Timepoint A that controlling for strangeness and mental capabilities somewhat reduced this tendency to avoid adults, b = 0.37 [0.27, 0.47], p < .001, and to seek vengeance toward adults, b = 0.07 [-0.03, 0.16], p = .19, while not reducing the tendency to reprimand children more than adults, b = -0.68 [-0.80, -0.55], p < .001. There was no significant change between timepoints in tendencies to avoid, seek vengeance against, or reprimand transgressors.
Furthermore, timepoint did not significantly moderate the differences between responses to adults vs. children, nor did individual differences in worries about the coronavirus (getting the PUNISHMENT OF CHILDREN AND ADULTS 22 virus oneself, or spreading it to others) significantly predict any of these relationships (full details are available in the Supplementary Materials). Our main pattern of results is therefore consistent when tested both before and after the onset of a global pandemic.
2.3 Discussion
Study 1 demonstrated a general tendency for adults to reprimand child transgressors more than adult transgressors, and to avoid adult transgressors more than child transgressors. This pattern persisted even after controlling for the perceived strangeness of the behavior, the perceived mental capabilities of the transgressor, and individual differences in disgust sensitivity and nurturant tendencies towards children. Strikingly, these results were consistent across different types of norm violations (which varied in their inclusion of both victims and pathogens), indicating that the age of a transgressor was more important than the nature of a transgressive action in determining punishment responses.
3 Study 2
Study 2 was designed to replicate this pattern of differences in adults’ responses to adult and child transgressors, and to test whether child observers also respond differently to adults compared to children their own age. Whereas adult observers tend to rate children as less in control of their behavior than adults, to rate children’s actions as less wrong and less counter- normative than adults’ actions, and to perceive norm-violating children as less threatening than norm-violating adults, we suspected that these patterns would not hold for child observers. A large body of literature has shown that children are readily willing to judge (Fu et al., 2014;
Smetana et al., 2003; Tisak, 1993), condemn (Helwig et al., 2001; Wainryb et al., 2005), tattle on
(Misch et al., 2018), and punish (Jordan et al., 2014; McAuliffe et al., 2015) their peers when they misbehave, implying that the divergences in adults’ responses to different transgressors will PUNISHMENT OF CHILDREN AND ADULTS 23 not be observed in children’s responses. That is, the age of the observer is likely to be as important as the age of the transgressor in shaping punitive responses.
Because the nature of the action did not meaningfully shape participants’ responses in
Study 1, we reduced the number of action categories from four to two. Rather than treating the presence of victims and the presence of pathogens as orthogonal factors, we grouped items into categories of harm (including victims but no pathogens) and purity (including no victims, and including either pathogens or more abstract forms of contamination). Previous research has indicated that harm and purity may represent distinct normative domains that recruit different forms of cognitive processing (e.g., Graham et al., 2013; Rottman & Young, 2019; Young &
Saxe, 2011) and that yield different forms of punishment (Kemper & Newheiser, 2018).
3.1 Methods
3.1.1 Participants
We aimed to recruit a final sample size of at least 90 child participants and 90 adult participants (after exclusions). A power analysis via simulation, conducted using the simr package in R (Green & MacLeod, 2016), indicated that this sample size would have more than
99% power to detect the difference between reprimand or avoidance of children vs. adults, based on the effect sizes observed in Study 1. All data were collected between July and December of
2019.
Child sample
Ninety-one child participants, aged 4 to 10 years old (Mage = 6.85 years, SD = 1.70 years, range = 4.02 – 10.0 years, 39 girls, 52 boys), were recruited from and tested at a Montessori school in a small city in Pennsylvania or recruited from a participant database and tested at a college laboratory in the same city. An additional 10 children were also tested but were excluded PUNISHMENT OF CHILDREN AND ADULTS 24 from the final sample, as preregistered, due to failures to pass the attention check at the end of the survey.
Adult sample
Adult participants were initially recruited from any willing parents of the child participants tested in the laboratory setting (n = 37); additional adult participants who reported that they were parents were recruited from MTurk (n = 53) to reach the desired sample size.
(Responses did not meaningfully differ, in magnitude nor inferential implications, between participants recruited through MTurk and the children’s parents.) An additional 37 adult participants, tested via MTurk, were excluded based on the same attention check questions used in Study 1.
3.1.2 Materials and Procedure
Vignettes describing transgressions
Each participant evaluated four different vignettes depicting transgressions, that manipulated within-subjects (a) whether the transgression was committed by an adult or by a child, and (b) whether the transgression involved a harm-related transgression or a purity-related transgression (see Figure 3 for examples). The pool of harm-related transgressions included 6 of the 7 victim/no pathogen scenarios used in Study 1, and the pool of purity-related transgressions included 3 victimless scenarios that potentially involved pathogens and 3 victimless scenarios that did not, adapted from the set of scenarios used in Study 1. Adult transgressors’ ages were described as being between 28 and 63 years old (depending on the particular vignette). For child participants, the age of child transgressors was matched to be the same as the participant’s age.
For adult participants, the child transgressors’ ages were described as 4, 5, 6, 7, 8, or 9. Each vignette was accompanied by a hand-drawn cartoon image of the described action, modified to PUNISHMENT OF CHILDREN AND ADULTS 25 include either an adult or a child transgressor (see Figure 3 for examples). The order of vignette presentation, the age of the target, and the content of each scenario was randomized across participants.
Figure 3. Examples of vignettes and images used as stimuli in Study 2.
Tim is a 6-year-old boy… Tim is a 52-year-old man…
Tim yells at the family's dog when it gets in his way.
Daniel is a 5-year-old boy… Daniel is a 63-year-old man…
Daniel likes to eat earthworms that he finds on the ground.
Evaluations of transgressions
After reading each vignette, participants evaluated the perceived wrongness of the action
(“How wrong is [Transgressor’s] behavior?”), whether they would reprimand the transgressor
(“Would you try to tell [Transgressor] why his [her] behavior was wrong?”), whether they would PUNISHMENT OF CHILDREN AND ADULTS 26 avoid the transgressor (“If there was only one seat left on the bus, and it was right next to
[Transgressor], would you sit next to him [her]?”), and the strangeness of the transgression
(“How weird would it be for other kids [grown-ups] to do what [Transgressor] did?”).
Individual difference measures
After evaluating the vignettes, participants completed measures of individual differences in disgust sensitivity (Viar-Paxton et al., 2015), mental capabilities attributed to average adults and children (i.e., most grown-ups [kids] are able to “…do things on purpose,” “…tell right from wrong,” and “…understand how others are feeling”), and reported other demographic characteristics. We had preregistered including these variables in the following analyses, but the disgust sensitivity scale did not reach an acceptable level of reliability in any sample (αs < .52), and the mental capability scale was not reliable among child participants (αs < .35); therefore, we do not include these variables in the main analyses reported below.
3.2 Results
3.2.1 Analytic strategy
All analyses were performed as multi-level models that included random intercepts nested within participant (to account for the repeated-measures nature of the data) and random intercepts nested within vignette scenario (to account for variability in stimuli). Transgressor Age and Participant Age were each dummy coded (0 = child, 1 = adult) and Transgression Type was contrast coded (-0.5 = purity, 0.5 = harm). Transgression Strangeness was centered within- participants, and then standardized. Separate regressions were performed predicting (1) wrongness judgments, (2) avoidance, and (3) reprimand of the transgressor. We first (Model1) investigated differences across the experimentally-manipulated conditions in each vignette according to participant age, and then (Model 2) strangeness was added as a covariate to this PUNISHMENT OF CHILDREN AND ADULTS 27 primary model. Full results from these models are depicted in Table 2 and Figure 4, and simple effects derived from this model are described in the main text.
3.2.2 Wrongness
In contrast to our predictions and to the Study 1 findings, adults’ and children’s transgressions were rated as similarly wrong, by both adult participants, bharm = 0.16 [-0.08,
0.40], p = .19, bpurity = 0.06 [-0.18, 0.30], p = .65, and child participants, bharm = 0.01 [-0.22,
0.25], p = .90, bpurity = -0.01 [-0.25, 0.22], p = .91 (full details are available in the Supplemental
Materials).
3.2.3 Avoidance
Adult participants were much more likely to avoid adult transgressors than child transgressors, b = 0.37 [0.20, 0.53], p < .001. Participants were more likely to avoid transgressors when their actions were perceived as stranger, and controlling for strangeness reduced the difference between avoidance of adult transgressors and child transgressors, b = 0.29
[0.13, 0.46], p < .001. This mediation was further probed using a structural equation model
(SEM) in lavaan, including only adult participants (and only including random intercepts by participant, not random effects by vignette, due to constraints in lavaan’s ability to account for random effects). Adult participants rated adults’ transgressions as much stranger than children’s transgressions, b = 0.54, p < .001, and strangeness predicted greater avoidance of transgressors, b
= 0.19, p < .001. Strangeness therefore partially mediated adults’ greater tendency to avoid adults more than child transgressors, indirect effect: b = 0.10 [0.05, 0.17], total effect, b = 0.31
[0.16, 0.47].
In contrast, child participants were equally likely to avoid adult transgressors and child transgressors, b = -0.07 [-0.23, 0.10], p = .44. Additionally, child participants rated adults’ and PUNISHMENT OF CHILDREN AND ADULTS 28
children’s transgressions as equally strange, bharm = 0.20 [-0.06, 0.46], p = .14, bpurity = -0.22 [-
0.48, 0.05], p = .11, and controlling for strangeness did not alter their tendency to avoid adult and child transgressors, b = -0.06 [-0.23, 0.10], p = .44. The age of child participants did not moderate these null relationships (see Supplementary Materials for full details).
3.2.4 Reprimand
Adult participants were less likely to reprimand adult transgressors than child transgressors, b = -0.32 [-0.51, -0.13], p = .001. They were also more likely to reprimand actions that were perceived as stranger, and controlling for strangeness increased the difference between reprimands of adult transgressors and child transgressors, b = -0.44 [-0.62, -0.25), p < .001. A test of this mediation (using SEM) indicated that adults’ transgressions were rated as stranger than children’s transgressions, and strangeness predicted reprimand, b = 0.32, p < .001, indirect effect: b = 0.17 [0.10, 0.25], total effect, b = -0.36 [-0.51, -0.18], but strangeness did not explain the greater tendency to reprimand child than adult transgressors: An even larger tendency to reprimand child vs. adult transgressors remained after controlling for strangeness, b = -0.53, p
< .001.
Child participants were equally likely to reprimand child transgressors and adult transgressors overall, although there was an interaction by transgression type such that children were equally likely to reprimand across target ages for purity-related transgressions, b = -0.10 [-
0.48, 0.27], p = .59, but less likely to reprimand adults for harm-related transgressions, b = -0.28
[-0.54, -0.02] p = .038. Exploratory analyses indicated that there was an interaction between child’s age and this transgressor age effect (see supplementary materials for full details), such that older children were more likely to reprimand children than adults for harm-related transgressions, whereas younger children responded to adults and children more similarly. PUNISHMENT OF CHILDREN AND ADULTS 29
Controlling for the strangeness of transgressions did not meaningfully change these transgressor age effects, bharm = 0.34 [0.04, 0.65], p = .026, bpurity = -0.07 [-0.44, 0.29], p = .69.
Figure 4. Adult and child participants’ mean tendency to avoid and reprimand adult and child transgressors, for each type of transgression. Means, and 95% CIs, are predicted from the preregistered multi-level models that included random intercepts nested within participants and within vignette scenario.
PUNISHMENT OF CHILDREN AND ADULTS 30
Table 2. Regression model predicting responses to transgressors from experimental conditions, individual differences, mental capabilities and perceived strangeness of the transgression.
Avoidance Avoidance Reprimand Reprimand
+ Strangeness + Strangeness b [95% CI] p b [95% CI] p b [95% CI] p b [95% CI] p Intercept 2.98 <.001 2.98 <.001 3.09 <.001 3.08 <.001 [2.71, 3.25] [2.73, 3.22] [2.85, 3.32] [2.87, 3.30] Transgressor Age -0.07 .44 -0.06 .44 -0.12 .20 -0.12 .20 [-0.23, 0.10] [-0.23, 0.10] [-0.31, 0.06] [-0.30, 0.06] Participant Age -0.98 <.001 -0.94 <.001 0.31 .01 0.38 .002 [-1.23, -0.74] [-1.19, -0.70] [0.07, 0.55] [0.14, 0.62] Transgression Type 0.24 .31 0.27 .21 0.41 .056 0.45 .015 [-0.22, 0.71] [-0.15, 0.69] [-0.01, 0.84] [0.09, 0.82] Transgressor Age * 0.43 <.001 0.36 .003 -0.2 .15 -0.32 .016 Participant Age [0.20, 0.67] [0.12, 0.59] [-0.46, 0.07] [-0.58, -0.06] Transgressor Age * 0.17 .32 0.1 .54 -0.31 .099 -0.41 .028 Transgression Type [-0.16, 0.49] [-0.22, 0.43] [-0.69, 0.06] [-0.77, -0.04] Participant Age * -0.3 .073 -0.29 .083 -0.1 .59 -0.07 .69 Transgression Type [-0.63, 0.03] [-0.61, 0.04] [-0.48, 0.27] [-0.44, 0.29] Transgressor Age * -0.05 .83 0.02 .93 0.29 .29 0.39 .14 Participant Age * [-0.52, 0.41] [-0.44, 0.48] [-0.24, 0.82] [-0.12, 0.91] Transgression Type Strangeness 0.13 <.001 0.2 <.001 [0.06, 0.19] [0.13, 0.27] Marginal R2 / Conditional 0.135 / 0.529 0.149 / 0.527 0.040 / 0.332 0.074 / 0.361 R2
Note: Transgressor age was dummy coded (0 = child transgressor, 1 = adult transgressor, such that estimates reflect the values for child transgressors, with transgressor age interactions indicating the change in these values when the transgressor was an adult); participant age was dummy coded (0 = child participant, 1 = adult participant, such that estimates reflect the values for child participants, with participant age interactions indicating the change in these values when ratings were made by an adult); and the type of transgression was contrast coded (-0.5 = purity, 0.5 = harm, such that estimates indicate the average effect across both conditions). All models also include random intercepts nested within participants. PUNISHMENT OF CHILDREN AND ADULTS 31
3.3 Discussion
Replicating the results of Study 1, adult participants reported that they were more likely to avoid adult than child transgressors (an effect that was partially mediated by the greater perceived strangeness of adults’ transgressions) and more likely to reprimand child than adult transgressors. Child participants’ responses did not consistently differ between adult and child transgressors, and child participants rated adults’ and children’s transgressions as similarly strange. Overall, these findings are consistent with the hypothesis that characteristics of the observer, in addition to characteristics of the transgressor, influence responses to norm violations.
Contrary to predictions, however, and contrary to previous findings that observers express more disdain toward members of their peer groups (Rottman et al., 2020), there were no consistent interactions between Transgressor Age and Participant Age, either for wrongness judgments or punitive inclinations.
4 Study 3
Studies 1 and 2 showed that adult participants respond differently to adult and child transgressors. In these studies, transgression strangeness partially accounted for adults’ greater tendency to avoid adult than child transgressors, but substantial variance in avoidance and reprimand tendencies remained. In Study 3—for which we recruited adult participants only—we aimed to replicate and extend the findings from our first two studies by continuing to investigate whether adults’ divergent punitive responses to adults and children can be partly explained by moral evaluations of transgressive actions themselves, such as the extent to which the transgressions are evaluated as wrong, strange, anger-provoking, and gross. In this study, we additionally included evaluations of the degree to which the transgressions were harmful. Study PUNISHMENT OF CHILDREN AND ADULTS 32
3 also tested several additional theoretically-relevant mediators, including perceptions of the transgressor’s mental capabilities (using the mental capability attribution questionnaire from the previous studies as well as a measure of the perceived changeability/stability of the transgressor’s character) and expectations about how confronting the transgressor would make them feel (including anxiety about confrontation and whether other people would view punitive responses positively or negatively). Including these additional variables allowed us to test the hypotheses that reprimand may be preferentially directed towards children because reprimand is expected to be (a) more effective when directed towards children’s more malleable character traits, (b) less threatening, and (c) generally viewed as a socially desirable response. In addition to these judgments about specific transgressions and transgressors, we assessed how several individual differences (dispositional disgust sensitivity, interaction anxiousness, and parental tenderness) predicted punitive response tendencies.
4.1 Methods
4.1.1 Participants
We planned to recruit a sample of approximately 350 adult participants, located in the
United States, from MTurk. Following preregistered criteria, we barred 432 individuals from participating due to failure of an English-language comprehension check at the beginning of the survey. Following the same criteria as in Study 1 (which were also preregistered), 235 participants were later excluded for failing an attention check, providing inappropriate open- ended responses, or admitting that they were distracted during the survey. New participants were recruited until we met the desired sample size. This resulted in a final sample of 355 participants, 63% of whom were parents. All data were collected in April 2020. PUNISHMENT OF CHILDREN AND ADULTS 33
4.1.2 Materials and procedure
Vignettes describing transgressions
Each participant first evaluated four different vignettes depicting transgressions, which manipulated within-subjects (a) whether the transgression was committed by an adult or a child and (b) whether the transgression involved a harm-related transgression or a purity-related transgression. The pool of vignettes was identical to Study 2 but did not include the images that had accompanied the vignettes in the previous study.
Evaluations of transgressions
Participants then reported how “wrong” the action was, how “angry” and “grossed out” it made them feel, how “harmful” the action was, and how “strange” it would be for other adults
[children] to do what the transgressor did. These items were analyzed separately for some analyses, as well as collapsed into a composite measure of transgression wrongness (α = .86).
Desired responses to transgressions
Participants next reported their likelihood of various responses to the transgression. Four items referred to reprimand, e.g., “I would try to explain to him why his behavior was wrong”
(composite α = .90), four referred to avoidance, e.g., “I would feel embarrassed to sit at the same table with him for lunch” (composite α = .91), and four referred to punishment, e.g., “He should be punished for his behavior” (composite α = .91).
Confrontation anxiety
Participants were then asked to imagine how confronting the transgressor would make them feel, and reported whether they would feel “anxious” and “uncomfortable”, and whether they would be “worried about how he [she] would react” and “afraid of his [her] response”
(composite α = .92). PUNISHMENT OF CHILDREN AND ADULTS 34
Predicted response consequences
Participants next reported whether other people would think that they are “a trustworthy, moral person” if these others saw the participant telling the transgressor that they did something wrong (reprimand consequence), avoiding the transgressor (avoidance consequence), and punishing the transgressor (punishment consequence).
Transgressors’ mental capabilities
Participants then rated whether the transgressor has various mental capabilities –– specifically, whether she [he] is able to “do things on purpose,” “come up with smart ideas,” “tell right from wrong,” “understand how others are feeling,” and “control her [his] behavior”
(composite α = .89).
Transgressors’ personality malleability
Four items then assessed the perception that the transgressor will change over time, e.g.,
“James’s moral character is something very basic about him and it can't be changed much” and
“James will continue learning and changing as he gets older” (composite α = .77).
Individual difference measures
After evaluating the vignettes, participants completed measures of individual differences in disgust sensitivity (Tybur et al., 2009, pathogen disgust subscale, α = .86, and moral disgust subscale, α = .92), the tendency to protect and nurture children (Hofer et al., 2018, composite α
= .88), and interaction anxiousness (Leary, 1983, α = .92). Finally, they reported other demographic characteristics and their worry about COVID-19 using the same measure as Study
1 (Timepoint B). PUNISHMENT OF CHILDREN AND ADULTS 35
4.2 Results
4.2.1 Analytic strategy
Preliminary analyses, replicating Studies 1 and 2, confirmed that adults’ actions were rated as less worthy of reprimand, b = -0.45 [-0.56, -0.34] p < .001, substantially more worthy of avoidance, b = 0.66 [0.56, 0.75], p < .001, and a little more worthy of punishment, b = 0.11
[0.01, 0.21], p = .036, than children’s actions (see Figure 5).
To assess possible explanations for these punitive responses, we conducted a series of preregistered multi-level regression models that separately tested how each of the three punitive responses—reprimand, avoidance, and punishment—were predicted by each of the four sets of predictors (evaluations of actions, evaluations of the transgressors’ mental capabilities, perceived consequences of responding, and individual differences). Results are depicted in Tables 3–5 and
Figure 6. In every case possible, analyses were performed as multi-level models that included random intercepts nested within participants (to account for the repeated-measures nature of the data) and random intercepts nested within vignette scenario (to account for variability in stimuli).
Manipulated variables were dummy coded and continuous predictors were standardized prior to analysis.
These preregistered analyses were then followed by exploratory analyses that included all variables in a single model, to assess their combined effect (a similar pattern of mediators was found when each set of variables was analyzed separately). This model was conducted as a structural equation model using lavaan to test which predictors mediated the association between transgressor age (adult vs. child) and reprimand/avoidance/punishment. Random intercepts by vignette were dropped from this path analysis to simplify the model to suit the constraints of PUNISHMENT OF CHILDREN AND ADULTS 36 lavaan, while random intercepts by participants were retained to account for the repeated- measures nature of the data.
Figure 5. Mean tendency to avoid, reprimand, and punish adult and child transgressors, for each type of transgression. Means, and 95% CIs, are predicted from the multi-level models that included random intercepts nested within participants and within vignette scenario.
4.2.2 Evaluation of transgressions
Adults’ actions were rated as more wrong, b = 0.14 [0.02, 0.26], p = .022, more anger- provoking, b = 0.32 [0.20, 0.45], p < .001, grosser, b = 0.28 [0.14, 0.42], p < .001, more harmful, b = 0.16 [0.04, 0.28], p = .008, and stranger, b = 0.75 [0.62, 0.88], p < .001, than identical actions performed by children. These transgressor age effects did not significantly differ between harm and purity transgressions (ps > .05). PUNISHMENT OF CHILDREN AND ADULTS 37
All of these negative evaluations of actions predicted greater avoidance of the transgressor, and partially mediated the tendency to avoid adult transgressors more than child transgressors, although a large tendency to avoid adult transgressors remained after controlling for these perceptions of the transgression’s wrongness, b = 0.43 (see Table 3). Anger and harmfulness also predicted greater reprimand, although these variables cannot explain the greater tendency to reprimand children: controlling for these variables actually increased the size of the transgressor age effect, such that participants are especially likely to reprimand children, b = -
0.56, after accounting for the anger, grossness, harmfulness, and strangeness of children’s vs. adults’ transgressions (see Table 4). Finally, anger, harmfulness, and strangeness also predicted greater punishment preferences, and mediated the greater tendency to punish adult transgressors
(see Table 5). In general, these effects remained consistent across harm transgressions and purity transgressions.
4.2.3 Perception of transgressors’ mental states
Adult transgressors were rated as having greater mental capabilities than child transgressors, b = 0.40 [0.32, 0.47], p < .001, and adults’ character was rated as less changeable than children’s character, b = -1.12 [-1.20, -1.05], p < .001.
As depicted in Tables 3–5, perceiving transgressors to have greater mental capabilities predicted less avoidance, less reprimand, and less punishment, and mental capabilities had a small negative indirect effect on avoidance and reprimand. Perceiving the transgressor’s character as more changeable predicted less avoidance, b = -0.27, p < .001, and partially mediated the tendency to avoid adult transgressors more than children. Changeable character also predicted greater reprimand, b = 0.20, p < .001, and mediated the tendency to reprimand child transgressors more than adults. After controlling for perceptions of mental capabilities and PUNISHMENT OF CHILDREN AND ADULTS 38 personality stability, participants’ willingness to reprimand adults and children did not significantly differ, b = -0.12, p = .10. Perceptions of character changeability were unrelated to punishment.
4.2.4 Expected consequences of confrontation
The thought of confronting adult transgressors was much more anxiety-provoking than the thought of confronting child transgressors, b = 0.71 [0.62, 0.80], p < .001. Confrontation anxiety predicted greater avoidance, b = 0.33, p < .001, and punishment, b = 0.12, p < .001, of transgressors and partially mediated the greater tendency to avoid adult transgressors compared to child transgressors. Confrontation anxiety did not predict likelihood of reprimand.
Participants were more likely to engage in avoidance, reprimand, or punishment if they perceived reputational benefits from these reactions, bs ranged from 0.28 to 0.47, but this only partially mediated the tendency to avoid adults more than child transgressors. Avoiding child transgressors was perceived to have the least reputational benefits (M = 2.22, SD = 1.31), whereas avoiding adults was perceived to have much greater reputational benefits (M = 2.68, SD
= 1.27), b = 0.46 [0.36, 0.56], p < .001. Participants perceived similar reputational consequences for reprimanding adult and child transgressors, b = -0.04 [-0.14, 0.07], p = .51, and for punishing adults and children, b = 0.07 [-0.03, 0.18], p = .17, and they thought they would be perceived as most moral for reprimanding the transgressor (M = 3.14, SD = 1.26), rather than punishing the transgressor (M = 2.62, SD = 1.33).
We also tested whether confrontation anxiety and expected response consequences moderated the association between transgression wrongness (i.e., a composite of how wrong, anger-provoking, gross, harmful, and strange the action was) and the likelihood of various responses to the transgressor. That is, does the wrongness of transgressions predict PUNISHMENT OF CHILDREN AND ADULTS 39 reprimand/avoidance/punishment more strongly among those who are low in confrontation anxiety, and among those who perceive reputational benefits from these responses? We did not find any evidence in favor of this hypothesis: transgression wrongness predicted greater avoidance, b = 0.37 [0.32, 0.41], p <.001, reprimand, b = 0.40 [0.35, 0.45], p < .001, and punishment, b = 0.35 [0.30, 0.40], p < .001, but this association was not significantly moderated by confrontation anxiety or expected response consequences, all bs < .03, ps > .16, except in the case of reprimand, where the association was in the opposite direction from hypothesized, b = -
0.09 [-0.14, -0.05], p < .001 (see Supplemental Materials for full details). Therefore, the perceived wrongness of the transgression, reputational benefits of responding, and anxiety about confronting the transgressor appear to be independent predictors that have unique mediating effects on the willingness to avoid, reprimand, and punish adults and children.
4.2.5 Individual differences
In addition to the evaluations of specific transgressions and transgressors described above, we investigated several individual differences that might similarly predict responses to transgressions. Trait measures of interaction anxiousness predicted greater avoidance and punishment of transgressors, bs > 0.23, but did not predict reprimand, b = -0.03, p = .62, a pattern analogous to the vignette-specific measures of confrontation anxiety described above.
Moral disgust sensitivity and pathogen disgust sensitivity likewise predicted avoidance and punishment, and also predicted reprimand of adults, bMoral = 0.11 [0.02, 0.21], p = .019, bPathogen
= 0.21 [0.11, 0.31], p < .001, but did not predict reprimand of children, bMoral = -0.04 [-0.14,
0.05], p = .40, bPathogen = 0.06 [-0.04, 0.16], p = .25 (these effects were generally consistent across harm transgressions and purity transgressions, and no other interactions between PUNISHMENT OF CHILDREN AND ADULTS 40 transgressor age and individual differences were significant; see Supplemental Materials for full models).
In contrast to these avoidance-motivating individual differences, parental care and tenderness predicted greater reprimand, b = 0.22, but did not predict avoidance or punishment, bs
< 0.08.
PUNISHMENT OF CHILDREN AND ADULTS 41
Figure 6. Mediation model displaying unstandardized regression coefficients. Bolded values are statistically significant, p < .05.
0.30 Anger 0.14
Grossness 0.22 0.10
0.13 Harmfulness 0.14
0.61 Strangeness 0.11
Transgressor 0.10 Age Avoidance -1.32 -0.16 Personality 0.59 Malleability -0.08
0.90 Mental Capabilities 0.17 0.56 Confrontation Anxiety 0.19 Consequences of avoidance
0.30 Anger 0.20
Grossness 0.21 0.02
0.13 Harmfulness 0.25
0.61 Strangeness 0.03
Transgressor -0.24 Age Reprimand -1.32 Personality 0.17 0.59 Malleability -0.06
0.90 Mental Capabilities -0.07 -0.06 Confrontation Anxiety 0.29 Consequences of reprimand
PUNISHMENT OF CHILDREN AND ADULTS 42
Table 3. Results of preregistered regression models and comprehensive mediation model predicting avoidance, Study 3
Preregistered separate models Exploratory comprehensive model b 95% CI p b p indirect effect total effect
Analysis A
Transgressor Age 0.43 0.34, 0.52 <.001 0.104 .052 Anger 0.16 0.10, 0.21 <.001 0.142 <.001 0.042*** 0.146** Grossness 0.13 0.07, 0.19 <.001 0.099 <.001 0.021** 0.125* Harmfulness 0.14 0.08, 0.20 <.001 0.136 <.001 0.018* 0.121* Strangeness 0.24 0.18, 0.31 <.001 0.105 <.001 0.064** 0.168** Controls: Type of transgression; and interactions with transgression type and all other variables Analysis B
Transgressor Age 0.39 0.26, 0.52 <.001 Mental capabilities -0.15 -0.20, -0.10 <.001 -0.083 <.001 -0.049*** 0.055 Personality Malleability -0.27 -0.33, -0.21 <.001 -0.156 <.001 0.206*** 0.310***
Analysis C
Transgressor Age 0.22 0.13, 0.32 <.001 Confrontation anxiety 0.33 0.28, 0.38 <.001 0.166 <.001 0.149*** 0.253*** Consequences of 0.28 0.23, 0.32 <.001 0.193 <.001 0.108*** 0.211*** avoidance
Analysis D
Transgressor Age 0.78 0.44, 1.12 <.001
PCAT -0.07 -0.17, 0.03 .19 -0.034 .45
Interaction Anxiousness 0.28 0.18, 0.39 <.001 0.251 <.001
Moral Disgust Sensitivity 0.1 0.00, 0.21 .045 0.052 .25 Pathogen Disgust 0.18 0.07, 0.29 .001 0.207 <.001 Sensitivity Controls: Transgressor Age*PCAT + Transgressor Age*Interaction Anxiousness + Transgressor Age*Disgust Sensitivity*Transgression Type *p < .05, **p < .01, ***p < .001 PUNISHMENT OF CHILDREN AND ADULTS 43
Table 4. Results of preregistered regression models and comprehensive mediation model predicting reprimand, Study 3
Preregistered separate models Exploratory comprehensive model b 95% CI p b p indirect effect total effect Analysis A
Transgressor Age -0.56 -0.66, -0.46 <.001 -0.24 <.001 Anger 0.24 0.18, 0.31 <.001 0.20 <.001 0.059*** -0.185** Grossness 0.04 -0.03, 0.10 .295 0.023 .43 0.005 -0.239*** Harmfulness 0.32 0.25, 0.38 <.001 0.25 <.001 0.032* -0.213*** Strangeness 0.00 -0.07, 0.07 .902 0.030 .34 0.019 -0.226*** Controls: Type of transgression; interactions with transgression type and all other variables Analysis B
Transgressor Age -0.12 -0.26, 0.02 .10 Mental capabilities -0.12 -0.17, -0.06 <.001 -0.055 .016 -0.033* -0.277*** Personality Malleability 0.20 0.13, 0.27 <.001 0.17 <.001 -0.224*** -0.468***
Analysis C
Transgressor Age -0.46 -0.56, -0.35 <.001 Confrontation anxiety 0.02 -0.03, 0.08 .369 -0.072 .006 -0.065** -0.309*** Consequences of 0.47 0.42, 0.51 <.001 0.29 <.001 -0.016 -0.260*** reprimand
Analysis D
Transgressor Age -0.43 -0.80, -0.06 .021
PCAT 0.22 0.12, 0.31 <.001 0.21 <.001
Interaction Anxiousness -0.03 -0.13, 0.07 .621 -0.043 .30
Moral Disgust Sensitivity -0.04 -0.14, 0.05 .403 0.021 .60 Pathogen Disgust 0.06 -0.04, 0.16 .247 0.14 <.001 Sensitivity Controls: Transgressor Age*PCAT + Transgressor Age*Interaction Anxiousness + Transgressor Age*Disgust Sensitivity*Transgression Type *p < .05, **p < .01, ***p < .001 PUNISHMENT OF CHILDREN AND ADULTS 44
Table 5. Results of preregistered regression models and comprehensive mediation model predicting punishment, Study 3
Preregistered separate models Exploratory comprehensive model b 95% CI p b p indirect effect total effect Analysis A Transgressor Age -0.06 -0.15, 0.03 .187 -0.045 .42 Anger 0.34 0.29, 0.40 <.001 0.326 <.001 0.097*** 0.053 Grossness -0.02 -0.09, 0.04 .425 -0.068 .012 -0.015* -0.059 Harmfulness 0.21 0.16, 0.27 <.001 0.160 <.001 0.021* -0.024 Strangeness 0.09 0.03, 0.15 .006 0.050 .094 0.030 -0.014 Controls: Type of transgression; interactions with transgression type and all other variables Analysis B
Transgressor Age 0.16 0.03, 0.30 .019 Mental capabilities -0.1 -0.15, -0.04 .001 -0.035 .10 -0.021 -0.065 Personality Malleability 0.00 -0.07, 0.07 .976 0.028 .31 -0.037 -0.082
Analysis C Transgressor Age -0.03 -0.13, 0.06 .515 Confrontation anxiety 0.12 0.07, 0.17 <.001 0.039 .11 0.035 -0.009 Consequences of 0.41 0.36, 0.46 <.001 0.245 <.001 0.019 -0.025 punishment
Analysis D Transgressor Age 0.37 0.02, 0.72 .036 PCAT 0.05 -0.06, 0.15 .358 0.050 .28 Interaction Anxiousness 0.23 0.12, 0.34 <.001 0.177 <.001 Moral Disgust Sensitivity 0.08 -0.02, 0.18 .135 0.094 .040 Pathogen Disgust 0.14 0.03, 0.25 .01 0.154 .001 Sensitivity Controls: Transgressor Age*PCAT + Transgressor Age*Interaction Anxiousness + Transgressor Age*Disgust Sensitivity*Transgression Type *p < .05, **p < .01, ***p < .001 PUNISHMENT OF CHILDREN AND ADULTS 45
4.3 Discussion
Study 3 replicated the patterns of differential avoidance, reprimand, and punishment of adult vs. child transgressors found in Studies 1 and 2, and provided further insights into why adults tend to differentially select each of these punitive responses for transgressors of various ages. These results confirm that there are different psychological predictors for different types of punitive responses: The greater avoidance of adult transgressors was predicted by the tendency to view adults’ actions as more wrong, to view adults’ character as more stable, and to feel anxiety about confronting them than about confronting children. Reprimand was more weakly associated with ratings of transgression wrongness and anxiety about confrontation, but reprimand was partially mediated by the perception that children’s character was more changeable over time, and therefore likely more amenable to change through reprimand. Neither of these patterns was evident when merely measuring the appropriateness of generalized
“punishment,” indicating the importance of measuring specific forms of indirect and direct punishment in order to reveal the divergent psychological profiles of different responses.
5 General Discussion
In three studies, we found that different responses to norm violations—including avoidance, reprimand, and punitive vengeance—have unique psychological profiles that depend on a combination of features of the action, the transgressor, and the observer. Most notably, adult transgressors were more likely to be avoided, and child transgressors were more likely to be reprimanded, for performing identical violations of harm and purity norms. These results contribute to a more nuanced understanding of how punitive responses are employed in different adaptive ways across different contexts. PUNISHMENT OF CHILDREN AND ADULTS 46
5.1 How important are features of the action itself?
Although much previous research has focused on the content of actions as determining various punitive responses, we instead found minimal relevance of the type of transgression that was committed. Our results remained fairly consistent across actions that posed emotional or physical harm to other people and actions that violated standards of bodily or spiritual purity, and this was true both before and after infectious disease threat became increasingly salient amid the coronavirus pandemic. Differential perceptions of the actions themselves can provide part of the explanation: The heightened tendency to avoid adults was partially mediated by adults’ actions being perceived as generally worse (stranger, more harmful, grosser, more anger-provoking, and overall more wrong), whereas these same actions were perceived as less severe norm violations when performed by children. But this is only part of the story.
Substantial differences between evaluations of adults and children remained even when controlling for these emotional reactions and appraisals of the transgressive actions. Appraisals of actions were especially limited in their ability to explain why people would reprimand children. Overall, more severe norm violations were more likely to be reprimanded, but children’s transgressions were viewed as less wrong and more likely to be reprimanded. In fact, the preference to reprimand children became even larger when statistically controlling for the strangeness of transgressions (i.e., when children’s actions were adjusted to be as strange as adults’ actions). Differential responses to norm violations can be more readily explained by looking beyond the features of the transgressive actions and instead focusing on characteristics of the transgressor and the punisher. PUNISHMENT OF CHILDREN AND ADULTS 47
5.2 Person-centered influences on direct and indirect punishment
The preference to reprimand children more than adults was partially explained by the perception that children’s moral character is less stable than that of adults (Study 3).
Reproaching a transgressor by explaining why their behavior was wrong is a potentially risky response, because it involves interacting with and directly confronting the transgressor without any threat of violence, which may substantially increase the danger of retaliation and future harm from the transgressor. However, this risk may be offset by future benefits, particularly if reprimand is actually effective at changing behavior and deterring future transgressions.
Consistent with this line of reasoning, we found that the greater tendency to reprimand children was mediated by the perception that children’s moral character is more malleable than adults’ moral character –– and thus presumably more amenable to changing as a result of reprimands.
Perceiving that adult transgressors have more stable immoral character traits partially explained the greater tendency to avoid adult transgressors. Avoidance was also independently explained by anxiety about confronting the transgressor, either as a state or trait measure. The overall pattern shows that if a transgressor is perceived as incapable of changing their bad behavior, or if confronting them seems too risky, then the safest course of action is to avoid them, rather than approaching them to explain why their behavior was wrong; as a result, this pattern of avoidance is more often directed towards adults than children.
These different sets of predictors show how reprimand and avoidance are flexibly deployed based on the costs and benefits of confronting different types of transgressors (see
Balafoutas et al., 2014; Lopez et al., 2019; Molho et al., 2017, 2020; Pedersen et al., 2018; Tybur et al., 2019). Notably, these differences were obscured when we merely asked about the deservingness of “punishment”. PUNISHMENT OF CHILDREN AND ADULTS 48
5.3 Conclusions and future directions
Our studies focused on evaluations of hypothetical strangers, described in brief vignettes.
While this may be equivalent to many judgments made in the real world – such as when witnessing an unknown child commit a transgression in a public park or school – these responses may be further affected by the relationship between the transgressor and the punisher, or the punisher and the victim, such as if they are a parent and child rather than unrelated acquaintances. Other sorts of moral judgments (e.g., Weidman et al., 2019) and punishment behaviors (e.g., Lopez et al., 2019) are affected by familial bonds and close friendships, and reprimand and avoidance of children’s transgressions are also likely to be shaped by the punisher’s pre-existing relationships. While our research represents a step in the direction of specifying critical identity-relevant characteristics, given our focus on age, future research should examine how additional features of identity may also play a meaningful role in shaping punitive responses to norm violations (Hester & Gray, 2019; Uhlmann et al., 2015).
In contrast to the responses of our adult participants, 4- to 10-year-old child participants did not respond differently to adult and child transgressors, across most measures and most transgression types (Study 2). Previous studies have found that children hold different people to different moral standards, such as expecting a child to receive more help from their parent than from their child-aged peers (Mammen et al., 2021), or expressing more disdain toward non- normative members of their peer groups (Rottman et al., 2020). Adults are likely perceived as existing outside the scope of children’s social circles, and therefore are potentially subject to different moral standards. It is therefore surprising that child participants held adults and children equally accountable for their transgressions, neither holding fellow children to a higher standard than adults, nor treating children more leniently than adult transgressors. Future PUNISHMENT OF CHILDREN AND ADULTS 49 developmental research could investigate whether children differentiate between adult and child transgressors through other kinds of responses (e.g., tendencies to gossip or tattle).
This overall pattern of results demonstrates how punitive responses can take many different direct and indirect forms. The particular response that is chosen in a given situation depends on much more than just the nature of the transgressive action: Reprimand and avoidance afford unique risks and benefits, and they were deployed in different contexts depending on the age of the transgressor and on the demographic characteristics and personality traits of the punisher. Punishment therefore reflects the person-centered nature of moral cognition in ways that remain underexplored. PUNISHMENT OF CHILDREN AND ADULTS 50
References
Balafoutas, L., & Nikiforakis, N. (2012). Norm enforcement in the city: A natural field
experiment. European Economic Review, 56(8), 1773–1785. doi:
10.1016/j.euroecorev.2012.09.008
Balafoutas, L., Nikiforakis, N., & Rockenbach, B. (2014). Direct and indirect punishment among
strangers in the field. PNAS, 111(45), 15924–15927.
Boehm, C. (2012). Moral origins: The evolution of virtue, altruism, and shame. Basic Books.
Boehm, C. (2014). The moral consequences of social selection. Behaviour, 151(2/3), 167–183.
Boyd, R., Gintis, H., Bowles, S., & Richerson, P. J. (2003). The evolution of altruistic
punishment. PNAS, 100(6), 3531–3535. doi: 10.1073/pnas.0630443100
Burnette, J. L., McCullough, M. E., Van Tongeren, D. R., & Davis, D. E. (2012). Forgiveness:
Results from integrating information about relationship value and exploitation risk.
Personality and Social Psychology Bulletin, 38(3), 345–356. doi:
10.1177/0146167211424582
Cushman, F. (2015). Punishment in humans: From intuitions to institutions. Philosophy
Compass, 10(2), 117–133. doi: 10.1111/phc3.12192
Cushman, F. A., Sarin, A., & Ho, M. K. (2019). Punishment as communication [Preprint].
PsyArXiv. doi: 10.31234/osf.io/wf3tz
Delton, A. W., & Krasnow, M. M. (2017). The psychology of deterrence explains why group
membership matters for third-party punishment. Evolution and Human Behavior. doi:
10.1016/j.evolhumbehav.2017.07.003 PUNISHMENT OF CHILDREN AND ADULTS 51 dos Santos, M., & Wedekind, C. (2015). Reputation based on punishment rather than generosity
allows for evolution of cooperation in sizable groups. Evolution and Human Behavior,
36(1), 59–64. doi: 10.1016/j.evolhumbehav.2014.09.001
Duncan, L. A., Schaller, M., & Park, J. H. (2009). Perceived vulnerability to disease:
Development and validation of a 15-item self-report instrument. Personality and Individual
Differences, 47(6), 541–546. doi: 10.1016/j.paid.2009.05.001
Fu, G., Xiao, W. S., Killen, M., & Lee, K. (2014). Moral judgment and its relation to second-
order theory of mind. Developmental Psychology, 50(8), 2085–2092. doi:
10.1037/a0037077
Goodwin, G. P., & Landy, J. F. (2014). Valuing different human lives. Journal of Experimental
Psychology: General, 143(2), 778–803. doi: 10.1037/a0032796
Graham, J., Haidt, J., Koleva, S., Motyl, M., Iyer, R., Wojcik, S. P., & Ditto, P. H. (2013). Moral
Foundations Theory: The pragmatic validity of moral pluralism. In P. Devine & A. Plant
(Eds.), Advances in Experimental Social Psychology (Vol. 47, pp. 55–130). Academic
Press. doi: 10.1016/B978-0-12-407236-7.00002-4
Green, P., & MacLeod, C. J. (2016). SIMR: An R package for power analysis of generalized
linear mixed models by simulation. Methods in Ecology and Evolution, 7(4), 493–498. doi:
10.1111/2041-210X.12504
Helwig, C. C., Zelazo, P. D., & Wilson, M. (2001). Children’s judgments of psychological harm
in normal and noncanonical situations. Child Development, 72(1), 66–81. doi:
10.1111/1467-8624.00266
Hester, N., & Gray, K. (2019). The Moral Psychology of Raceless Genderless Strangers.
Perspectives on Psychological Science. PUNISHMENT OF CHILDREN AND ADULTS 52
Ho, M. K., Cushman, F., Littman, M. L., & Austerweil, J. L. (2019). People teach with rewards
and punishments as communication, not reinforcements. Journal of Experimental
Psychology: General, 148(3), 520–549. doi: 10.1037/xge0000569
Hofer, M. K., Buckels, E. E., White, C. J. M., Beall, A. T., & Schaller, M. (2018). Individual
differences in activation of the parental care motivational system: An empirical distinction
between protection and nurturance. Social Psychological and Personality Science, 9(8),
907–916. doi: 10.1177/1948550617728994
Jordan, J. J., Hoffman, M., Bloom, P., & Rand, D. G. (2016). Third-party punishment as a costly
signal of trustworthiness. Nature, 530(7591), 473-476D. doi: /10.1038/nature16981
Jordan, J. J., McAuliffe, K., & Warneken, F. (2014). Development of in-group favoritism in
children’s third-party punishment of selfishness. PNAS, 111(35), 12710–12715.
Jordan, J. J., & Rand, D. G. (2019). Signaling when no one is watching: A reputation heuristics
account of outrage and punishment in one-shot anonymous interactions. Journal of
Personality and Social Psychology. doi: 10.1037/pspi0000186
Kemper, N. S., & Newheiser, A.-K. (2018). To confront or to avoid: How do people respond to
violations of moral norms? Social Psychological and Personality Science, 9(6), 734–743.
doi: 10.1177/1948550617722831
Krasnow, M. M., Cosmides, L., Pedersen, E. J., & Tooby, J. (2012). What are punishment and
reputation for? PLoS ONE, 7(9). doi: 10.1371/journal.pone.0045662
Krasnow, M. M., Delton, A. W., Cosmides, L., & Tooby, J. (2016). Looking under the hood of
third-party punishment reveals design for personal benefit. Psychological Science, 27(3),
405–418. doi: 10.1177/0956797615624469 PUNISHMENT OF CHILDREN AND ADULTS 53
Kriss, P. H., Weber, R. A., & Xiao, E. (2016). Turning a blind eye, but not the other cheek: On
the robustness of costly punishment. Journal of Economic Behavior & Organization, 128,
159–177. doi: 10.1016/j.jebo.2016.05.017
Kupfer, T. R., & Giner-Sorolla, R. (2017). Communicating moral motives: The social signaling
function of disgust. Social Psychological and Personality Science, 8(6), 632–640. doi:
10.1177/1948550616679236
Li, M., Vietri, J., Galvani, A. P., & Chapman, G. B. (2010). How do people value life?
Psychological Science, 21(2), 163–167.
Lopez, L. D., Moorman, K., Schneider, S., Baker, M. N., & Holbrook, C. (2019). Morality is
relative: Anger, disgust, and aggression as contingent responses to sibling versus
acquaintance harm. Emotion. doi: 10.1037/emo0000707
Mammen, M., Köymen, B., & Tomasello, M. (2021). Young children’s moral judgments depend
on the social relationship between agents. Cognitive Development.
Marlowe, F. W., & Berbesque, J. C. (2008). More “altruistic” punishment in larger societies.
Proceedings: Biological Sciences, 275(1634), 587–590. JSTOR.
McAuliffe, K., Jordan, J. J., & Warneken, F. (2015). Costly third-party punishment in young
children. Cognition, 134, 1–10. doi: 10.1016/j.cognition.2014.08.013
McCullough, M. E., Kurzban, R., & Tabak, B. A. (2013). Cognitive systems for revenge and
forgiveness. Behavioral and Brain Sciences, 36(01), 1–15. doi:
10.1017/S0140525X11002160
Misch, A., Over, H., & Carpenter, M. (2018). The whistleblower’s dilemma in young children:
When loyalty trumps other moral concerns. Frontiers in Psychology, 9. doi:
10.3389/fpsyg.2018.00250 PUNISHMENT OF CHILDREN AND ADULTS 54
Molho, C., Tybur, J. M., Güler, E., Balliet, D., & Hofmann, W. (2017). Disgust and anger relate
to different aggressive responses to moral violations. Psychological Science, 28(5), 609–
619. doi: 10.1177/0956797617692000
Molho, C., Tybur, J. M., Van Lange, P. A. M., & Balliet, D. (2020). Direct and indirect
punishment of norm violations in daily life. Nature Communications, 11(1), 3432. doi:
10.1038/s41467-020-17286-2
Murray, D. R., Kerry, N., & Gervais, W. M. (2019). On disease and deontology: Multiple tests of
the influence of disease threat on moral vigilance. Social Psychological and Personality
Science, 10(1), 44–52. doi: 10.1177/1948550617733518
Pedersen, E. J., McAuliffe, W. H. B., & McCullough, M. E. (2018). The unresponsive avenger:
More evidence that disinterested third parties do not punish altruistically. Journal of
Experimental Psychology: General, 147(4), 514–544. doi: 10.1037/xge0000410
Robinson, J. S., Xu, X., & Plaks, J. E. (2019). Disgust and deontology: Trait sensitivity to
contamination promotes a preference for order, hierarchy, and rule-based moral judgment.
Social Psychological and Personality Science, 10(1), 3–14. doi:
10.1177/1948550617732609
Rottman, J., DeJesus, J. M., & Gerdin, E. (2018). The social origins of disgust. In N.
Strohminger & V. Kumar (Eds.), The moral psychology of disgust (pp. 27–52). Powman &
Littlefield.
Rottman, J., Johnston, A. M., Bierhoff, S., Pelletier, T., Grigoreva, A. D., & Benitez, J. (2020).
In sickness and in filth: Developing a disdain for dirty people. Journal of Experimental
Child Psychology, 196, 104858. doi: 10.1016/j.jecp.2020.104858 PUNISHMENT OF CHILDREN AND ADULTS 55
Rottman, J., & Young, L. (2019). Specks of dirt and tons of pain: Dosage distinguishes impurity
from harm. Psychological Science, 30(8), 1151–1160. doi: 10.1177/0956797619855382
Smetana, J. G., Campione-Barr, N., & Yell, N. (2003). Children’s moral and affective judgments
regarding provocation and retaliation. Merrill-Palmer Quarterly, 49(2), 209–236. JSTOR.
Smith, A., McCauley, T. G., Yagi, A., Yamaura, K., Shimizu, H., McCullough, M. E., &
Ohtsubo, Y. (2019). Perceived goal instrumentality is associated with forgiveness: A test of
the valuable relationships hypothesis. Evolution and Human Behavior. doi:
10.1016/j.evolhumbehav.2019.09.003
Tisak, M. S. (1993). Preschool Children’s Judgments of Moral and Personal Events Involving
Physical Harm and Property Damage. Merrill-Palmer Quarterly, 39(3), 375–390. JSTOR.
Tybur, J. M., Molho, C., Dores Cruz, T. D., Cakmak, B., Singh, G. D., & Zwicker, M. V. (2019).
Tybur et al. Disgust Anger Aggression. PsyArXiv. doi: 10.31234/osf.io/wuspt
Uhlmann, E. L., Pizarro, D. A., & Diermeier, D. (2015). A person-centered approach to moral
judgment. Perspectives on Psychological Science, 10(1), 72–81. doi:
10.1177/1745691614556679
Viar-Paxton, M. A., Ebesutani, C., Kim, E. H., Ollendick, T., Young, J., & Olatunji, B. O.
(2015). Development and initial validation of the Child Disgust Scale. Psychological
Assessment, 27(3), 1082–1096. doi: 10.1037/a0038925
Wainryb, C., Brehl, B. A., Matwin, S., Sokol, B. W., & Hammond, S. (2005). Being hurt and
hurting others: Children’s narrative accounts and moral judgments of their own
interpersonal conflicts. Monographs of the Society for Research in Child Development,
70(3), i–122. PUNISHMENT OF CHILDREN AND ADULTS 56
Weidman, A. C., Sowden, W. J., Berg, M. K., & Kross, E. (2019). Punish or protect? How close
relationships shape responses to moral violations. Personality and Social Psychology
Bulletin, 0146167219873485. doi: 10.1177/0146167219873485
White, C. J. M., & Schaller, M. (2018). Are children perceived to be morally exceptional?
Different sets of psychological variables predict adults’ moral judgments about adults and
young children. Personality and Social Psychology Bulletin, 44(8), 1147–1162. doi:
10.1177/0146167218760800
White, C. J. M., Norenzayan, A., & Schaller, M. (2020). How strongly do moral character
inferences predict forecasts of the future? Testing the moderating roles of transgressor age,
implicit personality theories, and belief in karma. PLOS ONE, 15(12), e0244144. doi:
10.1371/journal.pone.0244144
Yoder, A. M., Widen, S. C., & Russell, J. A. (2016). The word disgust may refer to more than
one emotion. Emotion, 16(3), 301–308. doi: 10.1037/emo0000118
Young, L., & Saxe, R. (2011). When ignorance is no excuse: Different roles for intent across
moral domains. Cognition, 120(2), 202–214. doi: 10.1016/j.cognition.2011.04.005 PUNISHMENT OF CHILDREN AND ADULTS 57
Supplementary Materials for
A person-centered approach to punishment: Reprimand and avoidance are differentially directed
towards adult and child transgressors
PUNISHMENT OF CHILDREN AND ADULTS 58
6 Table of Contents
Demographic details of all samples ...... 59 Vignettes describing moral transgressions ...... 60 Study 1 ...... 60 Potential pathogens and no victims ...... 60 Potential pathogens and victims ...... 60 No potential pathogens or obvious victims ...... 60 Victims and no potential pathogens ...... 61 Studies 2 & 3 ...... 62 Purity transgressions ...... 62 Harm transgressions ...... 64 Factor analysis of punishment items ...... 66 Predicting Wrongness of Transgression ...... 68 Study 1 ...... 68 Study 2 ...... 70 Study 3 ...... 71 Full results of Study 1B ...... 72 Replication of Study 1A results ...... 72 Did responses change over time? ...... 72 Study 2: Does the age of child participants make a difference? ...... 76 Study 3 Full Models ...... 79
PUNISHMENT OF CHILDREN AND ADULTS 59
7 Demographic details of all samples
Table S1.
Study 1A Study 1B Study 2 Study 3 Sample age group Adults Adults Children Adults Adults Gender (% men) 38% 37% 0.57 0.36 0.57 % Parents 51% 49% 100% 63% Age M 38.22 42.89 6.85 38.67 39.21 Education (% Bachelor’s degree or 59% 66% higher) Household $50,000 - $50,000 - Income Mdn. $59,999 $59,999 Political conservatism M 3.60 3.76 (1 – 7 scale) Parental care and tenderness M 3.82 3.76 3.81 (1 – 5 scale)
Disgust sensitivity M 1.19 1.22 1.03 (0 – 2 scale) Ethnicity % White 77 77 % East Asian 4 5 % Black 7 5 % Hispanic or 7 8 Latino Other 5 5
COVID worry M 52.71 57.36 (0 – 100 scale)
Note. Age, gender, and parenthood were the only demographic variables assessed in Studies 2 and 3.
PUNISHMENT OF CHILDREN AND ADULTS 60
8 Vignettes describing moral transgressions
8.1 Study 1
8.1.1 Potential pathogens and no victims
1. James finds a piece of cheese that a rat is eating and decides to finish eating it after the rat leaves. 2. John pees into a cup and then drinks it 3. Robert sees a squirrel on the street that has been run over by a car. He picks it up with his bare hands and takes it home to keep in his closet. 4. Mary gets a nosebleed and decides to rub it all over her face and clothing 5. Daniel likes to eat earthworms that he finds on the sidewalk. 6. During a picnic lunch in the park, Maggie picks up a slimy slug from the ground, puts it in her mouth, chews on it for a few seconds, and then spits it out into a bowl of potato salad that she is eating. 7. While David is alone at home, he smears a handful of his own poop all over his bathroom wall.
8.1.2 Potential pathogens and victims
1. Mike is staying at a friend’s house when he realizes that he forgot his toothbrush. He uses his friend’s toothbrush without asking if that was okay. 2. Sarah is sick with a cold, but is eating lunch with friends anyways. She doesn’t have anything to drink, so when her friend isn’t looking, she takes a drink from her friend’s glass. 3. Jennifer passes some food to her friend with her bare hands, even though she hasn't washed her hands all week long. 4. Susan puts all of her fingernail clippings in a jar next to her sister's bed. 5. Mark pees into a public swimming pool while other people are also swimming. 6. Andrew has a runny nose, and he sneezes directly into a stranger’s face while riding the bus. 7. Ashley takes a bag of cockroaches with her to a restaurant.
8.1.3 No potential pathogens or obvious victims
1. Will is eating a snack while hiking through a forest, and he throws the package onto the ground instead of waiting to put it in the garbage 2. When camping, Emily draws a picture and writes her name on a big, pretty rock with permanent markers. 3. Kevin kisses his sister on the lips for fun, after she says it’s OK. 4. Melissa likes watching her best friend take baths. 5. Tom uses God’s name along with other really bad words to swear when he gets angry 6. Amanda rips a page out of the Bible to make a paper airplane 7. Chris finds an old gravestone and carries it home to use as a doorstop.
PUNISHMENT OF CHILDREN AND ADULTS 61
8.1.4 Victims and no potential pathogens
1. Stephanie won’t share a box of chocolates with her friend, even after being asked nicely 2. Tim yells at the family's dog when it gets in his way. 3. Brian says mean things about how an overweight person looks 4. Laura hears someone calling for help from another room, and does not do anything 5. While Amy is visiting another person’s house, she breaks an expensive lamp and laughs about it 6. Ryan takes money from his sister’s wallet without asking about it 7. Anne calls her friend “stupid” for making mistakes during a really hard game
PUNISHMENT OF CHILDREN AND ADULTS 62
8.2 Studies 2 & 3
Note: Images were not used in Study 3.
8.2.1 Purity transgressions
1. James finds a piece of pizza that a rat is eating and decides to finish eating it after the rat leaves.
2. Daniel likes to eat earthworms that he finds on the ground.
3. When camping, Jane draws pictures and carves her name onto a pretty tree.
PUNISHMENT OF CHILDREN AND ADULTS 63
4. Amanda rips a page out of the Bible to make a paper airplane
5. While David is at home, he smears a handful of his own poop all over his bathroom wall.
6. Kevin kisses his sister on the lips for fun, after she says it’s OK.
PUNISHMENT OF CHILDREN AND ADULTS 64
8.2.2 Harm transgressions
1. Ryan takes money from his sister without asking about it.
2. Robert says mean things about how an overweight person looks.
3. Anne calls her friend “stupid” for making mistakes during a really hard game.
PUNISHMENT OF CHILDREN AND ADULTS 65
4. Stephanie won’t share any chocolates with her friend, even after being asked nicely.
5. Tim yells at the family's dog when it gets in his way.
7. While Amy is visiting another person’s house, she breaks an expensive lamp and laughs about it.
PUNISHMENT OF CHILDREN AND ADULTS 66
9 Factor analysis of punishment items
9.1 Study 1
We conducted an exploratory factor analysis, with oblimin rotation, to determine appropriate subscales to create from the set of eight punishment-related items. A one- or two- factor solution did not provide a good fit for the data (RMSEA > .20, explained variance < 62%), whereas a three-factor solution provided a good fit to the data, RMSEA = .066 [.054, 0.079], proportion of variance explained = 74%, with no strong cross-loadings between factors. Factor loadings, displayed in Table 2, indicate one factor comprising reprimand of the transgressor, one factor comprising willingness to seek vengeance toward the transgressor, and one factor comprising the avoidance of the transgressor. The factors comprising willingness to avoid and seek vengeance were positively correlated (r = .49), and reprimand was only weakly associated with avoidance (r = .09) and vengeance (r = .11).
Table S2. Factor loadings of punishment items, Study 1. Factor loadings below .20 are not displayed.
Avoidance Reprimand Yell/Hit I would try to explain to her why it is 0.99 better to behave well I would try to explain to her why her 0.81 behavior was wrong I would yell, shout, or scream at her 0.80 I would want to slap or hit her 0.86 I would try to stay as far away from her 0.85 as possible I would not want to be her friend 0.89 I would feel embarrassed to have lunch 0.81 with her I wouldn’t want to be on her team when 0.84 playing a game
PUNISHMENT OF CHILDREN AND ADULTS 67
9.2 Study 3
An exploratory factor analysis, with oblimin rotation, was also used to confirm that the
12 punishment-related items in Study 3 could be broken down into sub-scales for analyses. A three-factor solution provided a good fit to the data, RMSEA = .042 [.034, .050], proportion of variance explained = 72%, with no strong cross-loadings between factors. Factor loadings, displayed in Table 3, indicate one factor comprising reprimand of the transgressor, one factor comprising the avoidance of the transgressor, and one factor comprising a general willingness to see the transgressor punished. The factors comprising willingness to avoid and punish the transgressor was strongly positively correlated (r = .68), and reprimand was more weakly associated with avoidance (r = .23) and punishment (r = .51).
Table S3. Factor loadings of punishment items, Study 3. Factor loadings below .20 are not displayed.
Avoidance Reprimand Punishment
I would try to explain to him why it is better 0.90 to behave well I would try to explain to him why his 0.92 behavior was wrong I would confront him and tell him that he did 0.68 something bad I would express my disappointment in his 0.68 behavior I would try to stay as far away from him as 0.89 possible I would feel embarrassed to sit at the same 0.82 table with him for lunch I would avoid sitting next to him on the bus, 0.86 even if it was the only seat left I wouldn’t want to be on his team when 0.78 playing a game He should be punished for his behavior 0.90 There should be severe consequences for 0.74 what he did He should get in trouble for what he did 0.83 Someone should make him pay for what he 0.78 did PUNISHMENT OF CHILDREN AND ADULTS 68
10 Predicting Wrongness of Transgression
Although it was not our main focus, in each study we also measured participants’ evaluations of the wrongness of each transgression. Children’s actions were perceived as significantly less wrong than identical actions performed by adults, across all types of moral transgressions in Studies 1 and 3 (although not in Study 2).
10.1 Study 1
Figure S1. Mean [95% CI] evaluation of the wrongness of each type of transgression. Wrongness was created from a composite of how much each action was perceived as wrong, anger-provoking, gross, and indicative of being a bad person. Transgression type: P = pathogens, no victim; P&V = pathogens and victim; V = victim, no pathogens; NP/NV = neither pathogens nor victim.
PUNISHMENT OF CHILDREN AND ADULTS 69
Table S3. Regression model predicting wrongness evaluations from experimental conditions, individual differences, mind attribution and perceived strangeness of the transgression. Wrongness b [95% CI] p b [95% CI] p Intercept 2.64 <.001 2.87 <.001 [2.51, 2.78] [2.75, 2.98] Transgressor Age 0.50 <.001 0.06 .055 [0.43, 0.56] [-0.00, 0.11] Pathogens 0.56 <.001 0.25 .017 [0.30, 0.82] [0.05, 0.46] Victims -0.01 .97 0.19 .069 [-0.27, 0.25] [-0.01, 0.40] Disgust Sensitivity 0.16 <.001 0.15 <.001 [0.10, 0.21] [0.10, 0.20] PCAT 0.01 .69 0.02 .45 [-0.05, 0.07] [-0.04, 0.08] Transgressor Age * Pathogens 0.05 .41 -0.12 .018 [-0.07, 0.17] [-0.21, -0.02] Transgressor Age * Victims 0.20 .001 0.16 .001 [0.08, 0.32] [0.06, 0.26] Pathogens * Victims -0.27 .32 0.09 .67 [-0.78, -0.25] [-0.32, 0.51] Transgressor Age * PCAT 0.06 .050 0.04 .090 [-0.00, 0.12] [-0.01, 0.37] Transgressor Age*Pathogen*Victims 0.32 .007 0.18 .064 [0.09, 0.56] [-0.01, -0.11] Mental capabilities -0.14 <.001 [-0.17, -0.11] Strangeness 0.51 <.001 [0.48, 0.54] Marginal R2 / 0.174 / 0.439 0.371 / 0.625 Conditional R2
PUNISHMENT OF CHILDREN AND ADULTS 70
10.2 Study 2
Figure S2. Mean [95% CI] evaluation of the wrongness of each type of transgression.
Table S4. Regression model predicting wrongness evaluations from experimental conditions, individual differences, mind attribution and perceived strangeness of the transgression. Wrongness b [95% CI] p b [95% CI] p Intercept 3.28 <.001 3.28 <.001 [2.97, 3.60] [3.01, 3.55] Transgressor Age 0.00 .99 0.00 .97 [-0.17, 0.17] [-0.16, 0.16] Participant Age -0.4 <.001 -0.32 .001 [-0.60, -0.21] [-0.51, -0.13] Transgression Type 0.73 .02 0.79 .003 [0.11, 1.35] [0.27, 1.30] Transgressor Age 0.11 .39 -0.05 .67 *Participant Age [-0.13, 0.35] [-0.28, 0.18] Transgressor Age * 0.03 .87 -0.1 .54 Transgression Type [-0.31, 0.36] [-0.42, 0.22] Participant Age * -0.39 .025 -0.36 .029 Transgression Type [-0.72, -0.05] [-0.67, -0.04] Transgressor Age * 0.07 .76 0.22 .34 Participant Age * [-0.40, 0.55] [-0.23, 0.67] Transgression Type Strangeness 0.27 <.001 [0.20, 0.33] Marginal R2 / 0.107 / 0.426 0.175 / 0.451 Conditional R2 PUNISHMENT OF CHILDREN AND ADULTS 71
10.3 Study 3
Figure S3. Mean [95% CI] evaluation of the wrongness of each type of transgression. Wrongness was created from a composite of how “wrong” the action was, how “angry,” and “grossed out” it made them feel, how “harmful” the action was, and how “strange” it would be for other adults/children to do what the transgressor did.
Table S5. Regression model predicting wrongness evaluations from experimental conditions, individual differences, mind attribution and perceived strangeness of the transgression.
Wrongness b [95% CI] p Intercept 2.82 <.001 [2.50, 3.13] Transgressor Age 0.33 <.001 [0.23, 0.43] Transgression Type -0.26 <.001 [-0.40, -0.12] Transgressor Age * 0.07 .46 Transgression Type [-0.12, 0.27] Marginal R2 / 0.031 / 0.339 Conditional R2
PUNISHMENT OF CHILDREN AND ADULTS 72
11 Full results of Study 1B
Participants (n = 157) from Study 1 returned at a second timepoint, one year later and after the onset of the global coronavirus pandemic, to complete the survey again. The overall survey was identical to that completed at Time 1, although at both timepoints participants were randomly assigned to 4 out of the possible 28 scenarios and therefore may not have been responding to exactly the same vignettes at both timepoints.
11.1 Replication of Study 1A results
We first tested whether the main pattern of results found in Study 1 could be replicated at
Timepoint B, using the same preregistered analytic strategy. Results are depicted in Figure 4 and
Table 6. Participants were substantially more likely to avoid adult transgressors than child transgressors, more likely to seek vengeance toward adult transgressors than child transgressors, and substantially less likely to reprimand adults compared to child transgressors. Controlling for strangeness and mind attribution somewhat reduced this tendency to avoid adults and seek vengeance toward adults, while not reducing the tendency to reprimand children more than adults. This pattern replicates the main findings from Time 1.
11.2 Did responses change over time?
We next tested whether responses to adults and children’s responses changed over time, through models that predicted responses to transgressions (avoidance, vengeance, reprimand) from transgressor age (0 = child, 1 = adult), timepoint (0 = 2019, 1 = 2020), and their interaction, with random intercepts per participants and per vignette to account for the repeated-measures nature of the data. This was followed by models that added individual differences in worry about contracting or passing on COVID-19, as a predictor interacting with each of these variables, to assess whether subjective perceptions of the pandemic affected responses. As depicted in Table PUNISHMENT OF CHILDREN AND ADULTS 73
7, responses were largely consistent across time, and not meaningfully moderated by worry about
COVID-19.
Figure S4. Adult and child participants’ mean tendency to avoid, reprimand, and punish adult and child transgressors, for each type of transgression. Means, and 95% CIs, are predicted from the preregistered multi-level models, which include random intercepts nested within participants (all studies) and control for individual differences in PCAT and disgust sensitivity.
PUNISHMENT OF CHILDREN AND ADULTS 74
Table S6. Study 1 Timepoint B: Regression model predicting responses to transgressors from experimental conditions, individual differences, mind attribution and perceived strangeness of the transgression. Avoidance Vengeance Reprimand b [95% CI] p b [95% CI] p b [95% CI] p b [95% CI] p b [95% CI] p b [95% CI] p Intercept 2.26 2.48 1.64 1.75 4.04 4.06 <.001 <.001 <.001 <.001 <.001 <.001 [2.10, 2.42] [2.33, 2.64] [1.50, 1.78] [1.61, 1.89] [3.88, 4.19] [3.91, 4.22] Transgressor Age 0.81 0.37 0.28 0.07 -0.63 -0.68 <.001 <.001 <.001 <.001 .18 <.001 [0.71, 0.91] [0.27, 0.47] [0.20, 0.37] [-0.03, 0.16] [-0.73, -0.52] [-0.80, -0.55] Pathogens 0.41 0.02 0.06 -0.13 0.09 -0.03 .82 .002 .86 .53 .18 .42 [0.15, 0.67] [-0.20, 0.25] [-0.13, 0.25] [-0.31, 0.06] [-0.13, 0.31] [-0.24, 0.19] Victims 0.05 0.30 0.00 0.12 0.11 0.18 .10 .69 .009 .96 .19 .33 [-0.20, 0.31] [0.08, 0.52] [-0.19, 0.20] [-0.06, 0.31] [-0.11, 0.32] [-0.04, 0.39] Disgust 0.16 0.15 -0.01 -0.02 0.13 0.13 .049 .005 .008 .83 .78 .044 Sensitivity [0.05, 0.28] [0.04, 0.27] [-0.13, 0.10] [-0.13, 0.10] [0.00, 0.25] [0.00, 0.25] PCAT -0.22 -0.17 -0.08 -0.06 0.13 0.14 .041 <.001 .006 .18 .34 .070 [-0.34, -0.10] [-0.29, -0.05] [-0.21, 0.04] [-0.18, 0.06] [-0.01, 0.26] [0.01, 0.28] Transgressor 0.31 0.08 0.18 0.08 0.13 0.06 .54 .002 .32 .033 .35 .22 Age*Pathogens [0.11, 0.50] [-0.08, 0.25] [0.01, 0.35] [-0.08, 0.24] [-0.08, 0.34] [-0.14, 0.27] Transgressor Age 0.05 -0.01 0.19 0.16 0.1 0.08 .45 .65 .86 .028 .050 .33 * Victims [-0.15, 0.24] [-0.18, 0.15] [0.02, 0.36] [-0.00, 0.32] [-0.11, 0.31] [-0.13, 0.29] Pathogens * -0.32 -0.08 -0.16 -0.05 -0.44 -0.33 .13 .22 .73 .42 .81 .047 Victims [-0.84, 0.19] [-0.52, 0.37] [-0.54, 0.22] [-0.41, 0.32] [-0.87, -0.01] [-0.75, 0.09] Transgressor Age 0.22 0.13 0.12 0.08 0.11 0.08 .14 <.001 .003 .005 .058 .044 * PCAT [0.12, 0.32] [0.04, 0.21] [0.04, 0.21] [-0.00, 0.16] [0.00, 0.22] [-0.03, 0.18] Transgressor -0.27 -0.27 0.16 0.16 0.53 0.53 .012 Age *Pathogens* .18 .11 .35 .33 .014 [-0.65, 0.12] [-0.59, 0.06] [-0.18, 0.49] [-0.16, 0.48] [0.11, 0.95] [0.11, 0.94] Victims Mental -0.08 -0.03 -0.13 <.001 <.001 .13 capabilities [-0.13, -0.04] [-0.08, 0.01] [-0.19, -0.08] Strangeness 0.58 0.28 0.13 <.001 <.001 <.001 [0.52, 0.63] [0.22, 0.33] [0.06, 0.20] Marginal R2 / .199 / .512 .334 / .655 .035 / .480 .080 / .527 .113 / .446 .131 / .466 Conditional R2 Note. Transgressor age was dummy coded (0 = child transgressor, 1 = adult transgressor, such that estimates reflect the values for child transgressors, with transgressor age interactions indicating the change in these values when the transgressor was an adult) and the presence of a victim or potential pathogens was contrast coded (-0.5 = no victim/pathogens, 0.5 = victim pathogens, such that estimates indicate the average effect across both conditions). All models also include random intercepts nested within participants.
PUNISHMENT OF CHILDREN AND ADULTS 75
Table S7. Change over time and association with worry about coronavirus
Avoidance Vengeance Reprimand b [95% CI] p b [95% CI] p b [95% CI] p b [95% CI] p b [95% CI] p b [95% CI] p Intercept 2.33 2.34 1.64 1.65 4.09 4.08 <.001 <.001 <.001 <.001 <.001 <.001 [2.15, 2.51] [2.16, 2.51] [1.50, 1.78] [1.51, 1.79] [3.93, 4.25] [3.92, 4.24] Transgressor Age 0.98 0.99 0.25 0.25 -0.58 -0.58 <.001 <.001 <.001 <.001 <.001 <.001 [0.88, 1.09] [0.88, 1.09] [0.16, 0.34] [0.16, 0.34] [-0.69, -0.47] [-0.69, -0.47] Timepoint -0.07 -0.08 0.00 -0.01 -0.05 -0.04 .43 .16 .15 .92 .89 .41 [-0.17, 0.03] [-0.18, 0.03] [-0.09, 0.08] [-0.09, 0.08] [-0.15, 0.06] [-0.15, 0.07] Transgressor Age -0.17 -0.17 0.03 0.03 -0.05 -0.05 .54 .017 .018 .67 .61 .54 *Timepoint [-0.32, -0.03] [-0.32, -0.03] [-0.10, 0.15] [-0.09, 0.16] [-0.20, 0.11] [-0.20, 0.11] COVID Worry -0.01 0.09 -0.01 .94 .80 .13 [-0.13, 0.10] [-0.03, 0.21] [-0.13, 0.12] Transgressor Age * 0.02 -0.02 0.07 .22 .71 .73 Worry [-0.08, 0.12] [-0.10, 0.07] [-0.04, 0.18] Timepoint * Worry 0.12 0.04 0.05 .33 .025 .39 [0.01, 0.22] [-0.05, 0.13] [-0.06, 0.16] Transgressor Age -0.06 -0.03 0.04 .65 .46 .62 *Timepoint * Worry [-0.20, 0.09] [-0.16, 0.09] [-0.12, 0.19] Marginal R2 / .138 / .445 .140 / .447 .016 / .440 .024 / .441 .060 / .375 .065 / .377 Conditional R2
PUNISHMENT OF CHILDREN AND ADULTS 76
12 Study 2: Does the age of child participants make a difference?
Additional analyses explored whether child participants’ age moderated their responses.
Using only the data from child participants, we tested multilevel models that predicted responses to each transgression from the transgressor’s age (0 = child, 1 = adult), transgression type (0 =
Purity, 1 = Harm), and participants’ age (in months, centered), with random intercepts by participant and by vignette scenario. As depicted in Table 8 and Figures 5 – 7, participants’ age did not moderate evaluations of transgression wrongness or avoidance of transgressors.
Participants’ age also did not moderate the tendency to reprimand purity-related transgressions, but did trend towards a moderation of reprimanding harm-related transgressions, such that older children were more likely to reprimand children than adults, whereas younger children responded more similarly to adults and children. Overall, 4–9-year-old child participants tended to evaluate children’s transgressions similarly to adults’ transgressions, in contrast to adult’s participants’ consistent tendency to avoid adults more and reprimand adults less than child transgressors.
PUNISHMENT OF CHILDREN AND ADULTS 77
Table S8. Study 2: Regression model predicting child participants’ responses to transgressors from experimental conditions and participants’ age (in months).
Wrongness Avoidance Reprimand b [95% CI] p b [95% CI] p b [95% CI] p Intercept 3.29 <.001 2.98 <.001 3.07 <.001 [2.92, 3.65] [2.66, 3.31] [2.84, 3.30] Participant Age 0.00 .55 -0.01 .22 0.01 .004 [-0.00, 0.01] [-0.02, 0.00] [0.00, 0.02] Transgressor Age 0.03 .66 -0.07 .47 -0.09 .41 [-0.12, 0.19] [-0.24, 0.11] [-0.30, 0.12] Transgression Type 0.66 .073 0.22 .45 0.39 .045 [-0.06, 1.37] [-0.36, 0.80] [0.01, 0.77] Participant Age 0.00 .58 0.00 .28 -0.01 .063 *Transgressor Age [-0.01, 0.01] [-0.00, 0.01] [-0.02, 0.00] Participant Age * 0.01 .32 0.01 .40 0.01 .19 Transgression Type [-0.01, 0.02] [-0.01, 0.02] [-0.00, 0.02] Transgressor Age * 0.07 .67 0.17 .36 -0.29 .18 Transgression Type [-0.24, 0.37] [-0.19, 0.52] [-0.71, 0.13] Participant Age -0.01 .17 -0.01 .11 -0.02 .071 *Transgressor [-0.03, 0.00] [-0.03, 0.00] [-0.04, 0.00] Age* Transgression Type Marginal R2 / 0.107 / 0.542 0.027 / 0.502 0.050 / 0.332 Conditional R2
Figure S5. Study 2: Transgression wrongness predicted from participant age (in months), transgressor age, and transgression type.
PUNISHMENT OF CHILDREN AND ADULTS 78
Figure S6. Study 2: Avoidance predicted from participant age (in months), transgressor age, and transgression type.
Figure S7. Study 2: Reprimand predicted from participant age (in months), transgressor age, and transgression type.
PUNISHMENT OF CHILDREN AND ADULTS 79
13 Study 3 Full Models
Tables 3 – 5 in the main text present an abbreviated version of the preregistered models, highlighting the key theoretically-interesting predictors of punitive responses. Below are the full results of all predictors in those models.
Table S9. Study 3: Regression model predicting responses to transgressors from evaluations of transgressions.
Avoidance Reprimand Punishment b [95% CI] p b [95% CI] p b [95% CI] p Intercept 2.50 <.001 3.68 <.001 2.72 <.001 [2.37, 2.63] [3.54, 3.81] [2.55, 2.90] Transgressor Age 0.43 <.001 -0.56 <.001 -0.06 .18 [0.34, 0.52] [-0.66, -0.46] [-0.15, 0.03] Anger 0.16 <.001 0.24 <.001 0.34 <.001 [0.10, 0.21] [0.18, 0.31] [0.29, 0.40] Transgression Type 0.22 <.001 0.21 <.001 0.19 <.001 [0.12, 0.32] [0.10, 0.32] [0.10, 0.29] Grossness 0.13 <.001 0.04 .30 -0.02 .43 [0.07, 0.19] [-0.03, 0.10] [-0.09, 0.04] Harmfulness 0.14 <.001 0.32 <.001 0.21 <.001 [0.08, 0.20] [0.25, 0.38] [0.16, 0.27] Strangeness 0.24 <.001 0.00 .90 0.09 .006 [0.18, 0.31] [-0.07, 0.07] [0.03, 0.15] Anger * -0.16 .016 -0.10 .14 -0.3 <.001 Transgression Type [-0.29, -0.03] [-0.24, 0.03] [-0.43, -0.17] Transgression Type* 0.03 .71 -0.03 .67 0.18 .02 Grossness [-0.12, 0.18] [-0.19, 0.12] [0.03, 0.32] Transgression Type 0.12 .089 -0.11 .11 0.01 .92 * Harmfulness [-0.02, 0.25] [-0.25, 0.02] [-0.12, 0.13] Transgression Type 0.05 .50 -0.05 .46 0.19 .004 * Strangeness [-0.09, 0.18] [-0.20, 0.09] [0.06, 0.33] Marginal R2 / 0.233 / 0.610 0.248 / 0.511 0.227 / 0.607 Conditional R2
PUNISHMENT OF CHILDREN AND ADULTS 80
Table S10. Study 3: Regression model predicting responses to transgressors from expected consequences of confrontation
Avoidance Reprimand Punishment b [95% CI] p b [95% CI] p b [95% CI] p Intercept 2.55 <.001 3.69 <.001 2.63 <.001 [2.43, 2.67] [3.59, 3.79] [2.48, 2.79] Transgressor Age 0.24 <.001 -0.48 <.001 -0.05 .24 [0.15, 0.32] [-0.57, -0.38] [-0.14, 0.04] Judgment 0.37 <.001 0.4 <.001 0.35 <.001 composite [0.32, 0.41] [0.35, 0.45] [0.30, 0.40] Confrontation 0.20 <.001 -0.12 <.001 0.01 .83 anxiety [0.15, 0.25] [-0.17, -0.07] [-0.04, 0.05] Consequence 0.20 <.001 0.33 <.001 0.30 <.001 [0.16, 0.24] [0.29, 0.38] [0.25, 0.34] Judgment * 0.02 .29 -0.03 .17 -0.03 .22 Confrontation [-0.02, 0.06] [-0.07, 0.01] [-0.07, 0.02] anxiety Judgment * 0.03 .25 -0.09 <.001 0.03 .18 Consequence [-0.02, 0.07] [-0.14, -0.05] [-0.01, 0.07]
Marginal R2 / 0.276 / 0.677 0.295 / 0.567 0.205 / 0.608 Conditional R2
PUNISHMENT OF CHILDREN AND ADULTS 81
Table S11. Study 3: Regression model predicting responses to transgressors from individual difference variables
Avoidance Reprimand Punishment b [95% CI] p b [95% CI] p b [95% CI] p Intercept 1.51 <.001 3.69 <.001 1.87 <.001 [1.11, 1.91] [3.32, 4.06] [1.46, 2.28] Transgressor Age 0.78 <.001 -0.43 .021 0.37 .036 [0.44, 1.12] [-0.80, -0.06] [0.02, 0.72] PCAT -0.07 .19 0.22 <.001 0.05 .36 [-0.17, 0.03] [0.12, 0.31] [-0.06, 0.15] Interaction Anxiousness 0.28 <.001 -0.03 .62 0.23 <.001 [0.18, 0.39] [-0.13, 0.07] [0.12, 0.34] Moral Disgust 0.10 .045 -0.04 .40 0.08 .14 Sensitivity [0.00, 0.21] [-0.14, 0.05] [-0.02, 0.18] Transgression Type 0.03 .67 0.27 <.001 0.25 <.001 [-0.11, 0.17] [0.12, 0.42] [0.11, 0.40] Pathogen Disgust 0.18 .001 0.06 .25 0.14 .010 Sensitivity [0.07, 0.29] [-0.04, 0.16] [0.03, 0.25] Transgressor Age * 0.06 .26 -0.02 .76 0.01 .88 PCAT [-0.05, 0.17] [-0.13, 0.10] [-0.10, 0.12] Transgressor Age -0.04 .47 -0.01 .90 -0.09 .12 *Interaction [-0.15, 0.07] [-0.13, 0.11] [-0.20, 0.02] Anxiousness Transgressor Age -0.08 .16 0.15 .008 0.07 .22 *Moral Disgust [-0.18, 0.03] [0.04, 0.27] [-0.04, 0.17] Transgressor Age * 0.00 .98 0.03 .77 0.08 .45 Transgression Type [-0.20, 0.20] [-0.18, 0.24] [-0.12, 0.28] Moral Disgust * 0.12 .10 0.13 .12 0.09 .26 Transgression Type [-0.03, 0.27] [-0.03, 0.29] [-0.06, 0.24] Transgressor Age 0.06 .27 0.15 .012 0.02 .71 *Pathogen Disgust [-0.05, 0.17] [0.03, 0.27] [-0.09, 0.13] Transgression Type -0.07 .33 -0.18 .028 -0.13 .081 *Pathogen Disgust [-0.22, 0.07] [-0.34, -0.02] [-0.28, 0.02] Transgressor 0.02 .84 -0.08 .51 0.02 .86 Age *Moral Disgust * [-0.19, 0.23] [-0.30, 0.15] [-0.19, 0.23] Transgression Type Transgressor Age * -0.07 .49 -0.09 .42 -0.1 .37 Transgression Type [-0.28, 0.14] [-0.32, 0.13] [-0.31, 0.11] *Pathogen Disgust Marginal R2 / 0.147 / 0.442 0.129 / 0.310 0.081 / 0.404 Conditional R2