How do Toddler’s Shame and Guilt affect their Prosocial and
Externalizing Behaviour?
Student: Laura Goldberg ID: 12211664 Supervisor: Milica Nikolic
Date: 5.7.2019 Word Count: 5,540 University of Amsterdam
Abstract
Since externalizing behaviour that start early are the most costly and persistent, it is crucial to understand the underlying self-conscious emotions that can prevent it. This study examined the relation between guilt and shame in toddlers (N=93, Mage 4.21 years) and their early prosocial or
externalizing behaviour. Guilt was expected to relate to less externalizing problems and more prosocial behaviour, while shame was expected to relate to the opposite. Observational measures were used to assess prosocial behaviour, and a broken toy mishap task was used to measure guilt and shame. Externalizing problem behaviour was calculated from parents completing the Child-Behavioural Checklist. The toddlers who expressed more guilt tended to have less externalizing problems, and shame-related aversion was nearly significantly related to more externalizing problems. However, toddler’s self-conscious emotions did not relate to how prosocial they were. Altogether, guilt seems to be the more adaptive emotion for toddlers’ social behaviour.
Keywords: toddlers, self-conscious emotions, guilt, shame, externalizing, prosocial,
How do Toddler’s Shame and Guilt affect their Prosocial and Externalizing Behaviour?
In 2017 alone, the estimated global cost of violence was $14.7 trillion dollars, the
equivalent of $1,988 per person worldwide (Desjardins, 2018). Globally, children are growing up faced with terrorism, world hunger, wars, gangs, the media glamorizing aggression and many other examples of violence (Desjardins, 2018). As early as during toddlerhood, aggression is already a strong predictor for later violence and crime in adults (Thijssen, 2016). While
aggression is a part of normal child development that typically decreases with age, little is known about why aggression worsens for certain toddlers (Alink et al., 2006; Thijssen, 2016).
Understanding what worsens toddlerhood aggression is important because it can help reduce societally costly delinquency later in life at the onset of the problem (Thijssen, 2016; Tremblay, 2000). Additionally, investigating toddler’s aggression next to their prosocial behaviour is useful as prosocial behaviour may be a protective factor against aggression (Menesini & Camodeca, 2008).
Prosocial behaviour is defined as voluntary acts intended to benefit others by helping, sharing, volunteering or comforting (Eisenberg, Spinrad, & Knafo-Noam, 2015). As toddlers are more prosocial, which usually increases with age, they also tend to behave less aggressively (Barrett, 1979). A longitudinal parent and teacher-report study on children aged 6-8 found that more sympathetic children had lower levels of aggression and externalizing problems, because
sympathetic children may have more emotional regulation preventing them from behaving aggressively in general (Eisenberg et al., 1996). Indeed, being sympathetic towards others has been identified as a key component motivating prosocial behaviour, and preventing aggressive behaviour (Eisenberg et al., 1996). Hence, toddlers who are more prosocial are less likely to display externalizing behaviour. In order to more thoroughly understand both externalizing and prosocial behaviour, this study investigates their shared underlying self-conscious emotions.
Underlying both aggression and prosocial behaviour are self-conscious emotions, which toddlers usually develop around age two, as they become self-aware of how others perceive them (Lewis, 1971). Self-conscious emotions, such as shame and guilt, are those which require appraisal of a situation, the ability to understand how one is perceived in the mind of another, and a sense of responsibility towards the victim for harming them (Barrett, Zahn-Waxler & Cole, 1993; Stipek, Gralinski, & Kopp, 1990).
Specifically, toddler’s guilt and shame can help them later in life to more optimally navigate social interactions including aggression and being prosocial (Flouri & Sarmadi, 2016). While both occur after transgressing a social norm, shame is characterized by feelings of the whole self as a failure and the view that others disapprove of them, while guilt is characterized by feeling that one has done something wrong (Ferguson et al., 1991). Shame can thus only be ameliorated by withdrawing the whole self from the social situation, while guilt can be ameliorated by taking responsibility for the wrongful act and taking action to make amends (Tangney & Dearing, 2002). So, toddlers who feel more shame are predicted to be less prosocial,
and even to behave more physically or verbally aggressively, as they are more prone to blame others for their feelings (Lewis, 1971).
Indeed, studies find that toddlers with more shame tend to have more externalizing problems, in terms of being hostile and aggressive towards others, as shame involves more global humiliation of the self and blaming of others (de Hooge, Zeelenberg & Breugelmans, 2007) . This is because while experiencing shame people tend to be focused on their own emotions and want to avoid negative consequences, sometimes even reacting aggressively as a defence if it feels like the whole self is threatened (Thomaes, Stegge & Olthof, 2007; Steuwig, Tangney, Harty & McCloskey, 2010). Shame can contribute to aggression, particularly for narcissistic toddlers to maintain their grandiose sense of self and high rank that are perceived to be under threat (Muris & Meesters, 2013). So, shame can sometimes be related to more
externalizing problems.
On the other hand, toddlers who feel more guilt are predicted to show more prosocial repairing behaviour and less aggression (Roos, Salmivalli & Hodges, 2013). This is because toddlers can be motivated to help out of the empathy that comes with guilt, and because the reparative behaviour indicative of guilt are already similar to helping in general (de Hooge, Zeelenberg, & Breugelmans, 2007). Since guilt requires empathy while externalizing behaviour typically comes with a lack of empathy, guilt is also related to less externalizing problems (Muris & Meesters, 2013).Research shows that guiltier toddlers tend to have less externalizing problems as guilt involves regretting one’s actions rather than their whole self, and being empathetic
towards others (de Hooge, Zeelenberg & Breugelmans, 2007). This is because while
experiencing guilt people tend to focus on the other’s emotions and want to repair the harm they did (Steuwig, Tangney, Harty & McCloskey, 2010). In youth, aggression and conduct disorder are associated with a lack of guilt as it may hinder the development of conscience and empathy (Steuwig, Tangney, Harty & McCloskey, 2010). So, guilt is associated with fewer externalizing behaviour. Altogether, toddlers’ shame appears to relate to more externalizing behaviour and less prosocial ones, while guilt is associated with the opposite.
Well-balanced self-conscious emotions can contribute to more functional relationships in all areas of life. Specifically, high levels of guilt and low levels of shame may contribute to adaptive social behaviour. The experience of guilt contributes to more appropriate prosocial behaviour as well as less externalizing problem behaviour, and the absence of shame contributes to less externalizing (Muris & Meesters, 2013). For instance, children who bully are found to have significantly lower levels of guilt than prosocial children (Menesini & Camodeca, 2008; Tangney & Fischer, 1995). Additionally, using the mishap task on 32 two-year olds has also previously shown that shame-prone toddlers are less likely to empathetically (but not
instrumentally) help someone than guilt-prone toddlers (Drummond et al., 2016). Dysregulated self-conscious emotions like too much or too little guilt are also linked to increased odds of psychopathology like depression and anxiety (Tangney & Fischer, 1995). Generally, self-conscious emotions seem to be linked to the development of social skills, moral reasoning, conscience, prosocial behaviour, and avoiding moral transgressions (Menesini & Camodeca,
2008). Taken together, the development of more guilt and less shame may be considered a prime developmental goal and potential way to minimize aggression.
There are also potential covariates that could influence self-conscious emotions and social behaviour in toddlers. For instance, the same study showed that low-income families and low-performing schools had higher chances of more externalizing behaviour problems and less prosocial behaviour (Flouri, & Sarmadi, 2016). Cultural differences in self-conscious emotions and externalizing problems have also been found using scenario-based self-evaluations in children from grades 3-6 (Furukawa, Tangney & Higashibara, 2012). Guilt was related to less externalizing problems, while shame (which Japanese children were more prone to than Korean and American children) predicted more externalizing problem behaviour (Furukawa, Tangney & Higashibara, 2012). Small gender differences in a meta-analysis of infants to adults also indicate that boys had more anger (g = .09) than girls in toddlerhood-middle childhood (Chaplin & Aldao, 2013). So, gender, socio-economic status and culture may influence which children display which self-conscious emotions and social behaviour.
On the whole, within the methodology used to study toddlers’ self-conscious emotions and social behaviour, there is little observational research on how guilt and shame relate to different social behaviour in young toddlers aged 2-5. A meta-analysis with 166 studies showed that those that did look at younger ages from infancy to age 5 only used self-reports or parent-reports (Chaplin & Aldao, 2013). It is important to look at younger children because the strongest predictor for future aggression is early-onset childhood aggression (Jolliffe et al.,
2017). Using observation in addition to self-reports may further add to our knowledge because observations are considered more objective measures of self-conscious emotions (Hubbard et al., 2004). Hence, by using observations of shame, guilt, and prosocial behaviour, as well as parent reports of their toddler’s general externalizing behaviour, this study will contribute to the research on how younger toddlers’ self- conscious emotions relate to social behaviour. The research question posed is “what are the effects of self-conscious emotions of shame and guilt on the display of prosocial and externalizing behaviour in children aged 2-5 years?”. It is
hypothesized that more guilt is associated with more prosocial behaviour while more shame is associated with less prosocial behaviour. Inversely, it is also hypothesized that higher levels of shame and lower levels of guilt are related to more externalizing behaviour. More prosocial behaviour is also predicted to relate to fewer externalizing behaviour.
Methods Participants
Ninety-three typically developing, middle-upper class toddlers ages 2-5 (M = 4.21, SD = 1.83, 51% female, 49% male) participated, along with one of their parents or guardians who filled in questionnaires while the children engaged in tests with a trained experimenter. This age group was selected because this is when self-conscious emotions are theorized to develop (Muris & Meesters, 2013). Parents were on average 35.6 years old (SD = 7.10), as well as 81% Dutch, 7% European, 6% Central American, 2% South-East Asian, 2 % Middle Eastern, and 2% North
African. Mostly married (83%) mothers participated (82% female, 18% male), from households with above average incomes of €4000- €6000 monthly (I am Expat, 2019). Their average highest attained educational level was between higher vocational education and a bachelor's degree, most of whom worked part-time (60%) and full-time (31%). The participants were recruited from schools, day-cares, and parent Facebook pages. Data was collected between March 2018 and May 2019.
General Procedure
All procedures took place with a trained experimenter in a large playroom with four adjustable cameras on each wall, operated by two or three experimenters behind a one-way window. The room had two desks, one for the experimenter and child, and one in the corner for the parent to be present during all tests while filling out questionnaires. Each session began with a short introduction with the experimenter in a separate waiting room. Other tests as a part of a larger study on self-conscious emotions were administered which the room was set up for, but these are not discussed here. After some free play with the experimenter to help the child acclimatize, along with other tasks of the larger study, the broken toy task was used to measure self-conscious emotions, followed by two helping tasks to measure prosocial behaviour. All the while, the parents filled out the Child Behaviour Checklist (CBCL) checklist for externalizing problem behaviour on a laptop in the same room (Achenbach & Rescorla, 2000). These
observational methods were used as opposed to questionnaires in order to measure the behaviour and emotions more objectively. For externalizing problems, the externalizing problem subscale
of a valid questionnaire was used (Harris and Brown, 2010). All methods were approved by the ethics board of the University of Amsterdam.
Measures
Mishap. To measure the independent variables shame and guilt, the well-validated
experimental mishap paradigm introduced by Barrett, Zahn-Waxler and Cole (1993) was used. The interrater reliability based on 20 % of the double-coded videos was excellent, with a kappa max of κ = .91 for shame-related gaze, body and head aversions, as well as good at κ = .72 for shame-related latency to show/talk about Teddy. Guilt-related repairing was moderate at κ = .64, and guilt-related talking/showing Teddy to parents or the test leader was also κ = .72. Parents were first given an extra consent form and were told that they can skip this task if they are uncomfortable with it as it involved deception. If they approved, toddlers were presented with a teddy bear that they were told was special to the experimenter, but that was also rigged to break once they started to play with it and the experimenter left the room (Drummond et al., 2016). The experimenter delivered six scripted cues at approximately 15 second time intervals using a neutral tone of voice, and all toddlers received each cue. When the child stopped playing with the toy or one minute elapses, the experimenter returned and gave the first five cues, which were: 1) standing outside the door loudly saying ‘ok, thank you very much, see you soon’ to announce their return, 2) looking at the toy with neutral affect, 3) asking the child what happened to teddy, 4) asking more specifically what happened to make teddy’s arm/leg fall off, 5) saying teddy was their favourite toy while looking at it. The final cue debriefed the child that the experimenter just
forgot that the teddy was already broken. If the child showed strong signs of distress or started crying, the experimenter skipped straight to the final cue. The child’s reaction was filmed and coded by trained observers (Barrett et al., 1993; Cole et al., 1992; Kochanska et al., 1995).
Specifically, the toddler’s facial expressions were filmed, as well as behaviour like shame-related gaze aversion, bodily avoidance, and guilt-related confessing to the parent or experimenter, and repairing behaviour were coded. The guilt variable was created by averaging the proportion of time toddlers spent repairing and talking about Teddy, minus the proportion of time toddler’s gaze was not visible from these respective tasks. This amount was then
standardized, as each toddler spent a different amount of time in total on the mishap task. The total guilt variable was then computed by averaging the two proportions of guilt-indicative behaviour. Next, two variables (total aversion and latency to talk) for shame were created as they were significantly negatively related to each other and could thereby not be combined. Guilt was measured as of the moment Teddy broke, and both guilt and shame were measured when the test leader re-entered and toddler’s self-conscious emotions would ensue. Both were also corrected for the proportion of time the toddler’s body language was not visible. Latency to talk was calculated by combining and averaging percentages of latency to talk/show Teddy to the parent or test leader. Next, aversion was calculated by combining and averaging the percentages of toddler’s gaze, head and body aversions from the test leader. See Table 1 for a detailed description of how behaviour were coded using the ‘Observer 13 XT Noldus program’ (Zimmerman et al., 2009). This scheme was adapted from Drummond et. al (2016).
Table 1. Self-Conscious Emotions Coding Schemes
Table 1
Shame and Guilt Relevant Behavioural Codes
Self-conscious Emotion Behaviour Coding in Observer
Shame Gaze aversion away from the
experimenter Child averts gaze from experimenter upon their return, looks away towards non-meaningful objects like the floor, wall or ceiling Bodily avoidance of the
experimenter The child backs away from the experimenter upon their return, or moves away from them towards a non-meaningful object or the parent after focusing on the experimenter
Head aversion Child turns head away from experimenter upon their return, towards a non-meaningful object or their parent after focusing on experimenter
Guilt Child tells parent or
experimenter about the mishap
Child shows the parent or experimenter the broken toy, or brings the broken bear or leg/arm to them, or verbalizes Child tries to repair Teddy Child tries to stick the bears
arm or leg back in, or asks parent or experimenter to help them
Prosocial behaviour. To measure prosocial behaviour, two well-validated measures
were used based on the Svetlova, Nichols and Brownwell’s empathetic and instrumental helping tasks (2010). The inter-rater reliability in this study was excellent at κ=1.00 for instrumental helping and κ=0.99 for empathetic helping, which is compatible to that of Drummond et. al with
an excellent interrater reliability score (κ = .97) (2016). The first measure that was used was an instrumental helping task involving helping the experimenter pick up a dropped item.
Specifically, the test leader announced that they are hanging up a poster, drop the tape, and deliver six cues, each 6-7 seconds apart. The cues were: 1) the experimenter saying “oops”, 2) “my tape fell on the floor” 3) “my tape, I need it back” 4) “oh my tape” while reaching for it with one hand 5) “[child's name], can you help me?” 6) “[child name], can you help me get my tape back?”. The experimenter delivered all the cues or until the child helped.
The second measure was an empathetic helping task, involving the experimenter saying that they are cold and pointing to a blanket with various cues. Earlier, the experimenter
announced that they were cold and covered themselves in a blanket to indicate to the child what to do when someone is cold. During the task, the experimenter delivered six cues, 6-7 seconds apart, after pretending to read on the couch for 10 seconds. The cues were: 1) saying “brrrr”, 2) “I am cold”, 3) “I need something so I can feel warm,'' 4) “Oh, a blanket!”, 5) “[child name], please will you help?”, 6) “[child name’, please, will you help me get my blanket?”. Again, all cues were delivered or until the child helps. If the toddler did not help after all the cues, the experimenter reached for the blanket themselves and concluded by announcing that they were warm again. Trained observers filmed and coded at which cue the child helps. For both tasks, the toddler’s prosocial behaviour was scored in reverse, 6 points for responding on the first cue, five for the second, etc. So, the fewer cues the child needs before they help the experimenter, the more points for prosocial behaviour the child scored.
Externalizing problem behaviour. Externalizing behaviour were measured by parents
completing the valid externalizing problems subscale of the 100-item CBCL for parents of children aged 1.5 to 5 (Achenbach & Rescorla, 2000). The internal consistency of this measure was excellent at ⍺= .95, which is compatible with similar studies finding a range of ⍺= .86 to ⍺= .89 for toddlers ages 3 to 5 (de la Osa et al., 2015). The externalizing scale consists of 24 of statements like the child having “temper tantrums or hot temper” that are rated on a 3-point scale (1 = not true, 2 = somewhat/sometimes true, and 3= very/often true). This tool was chosen because of its high reliability and feasibility in the lab as opposed to a psychological interview (Achenbach & Rescorla, 2000). Each child and parent were thoroughly debriefed before going home, and the child was given a choice of a male, female, or gender-neutral toy as a reward for participating.
Data Analysis
Standardized mean scores for the externalizing behaviour subscale of the CBCL checklist were computed using SPSS. For the self-conscious emotions percentages of time each child spent on each behaviour were calculated, and also standardized to compute the averages of guilt and shame. The prosocial behaviour scores were also standardized. The data was checked for outliers using the rule +/- 3 SD and skewness and kurtosis were checked for normal distribution. To answer the question of whether self-conscious emotions in toddlers relate to their social behaviour, a Pearson’s correlations and partial correlations tests accounting for toddler’s age as a covariate were first ran. Multiple regression analyses were used to assess the relations between
shame and guilt as independent variables, and externalizing and prosocial behaviour as dependent variables in two separate models.
Results Preliminary Analyses
Preliminary results revealed that only the shame-related variable ‘latency to talk and show parent or test leader’ was found to have two outliers and was therefore positively skewed. These outliers were winsorized and included in the final analyses (Wilcox, 2005). All data was then normally distributed, and further assumptions for later multiple regression analyses were also met, as data were independent, normally distributed and variables had linear relationships. Moreover, shame-related latency to talk was related to less guilt, r(57)= -.49, p <.001, and guilt was also negatively related to externalizing, r(57) = -.33, p = .012.
Next, sex did not influence guilt or shame-related behaviour, externalizing behaviour and prosocial helping. Thus, we did not include it in the further analyses as a covariate. However, with increasing age, toddlers were significantly more prosocial, r(44) = .28, p = .022, and expressed more guilt too, r(57) = .36, p< .001. Thus, age was included as a covariate in further analyses.
Perhaps not identifying more expected findings could be explained by the small sample size of only 59 diminishing the power of this study, after accounting for incomplete data that could not be imputed due to the fact that the missing data was not at random. It is possible that a confounding variable accounts for the missing data, like age. As in, maybe the mishap task was
mostly missing for younger children who were not attentive enough to complete it. See Table 2 for further details on data distribution.
Table 2. Descriptive Statistics for All Variables
Table 2
Means and Distributions of Guilt, Shame, Externalizing and Prosocial Behaviour in Toddlers
Variable nb Min Max M SD Skewness
Statistic Skewness SE Kurtosis Statistic Kurtosis SE
Average Guilt 59 1.34 44.2 19.6 11.4 .53 .31 -.77 .61 Latency to talk/ show (shame) 59 .16 2.01 .86 .46 .71 .31 -.18 .61 Total percentage Aversion (shame) 59 3.91 97.4 47.0 21.0 .28 .31 -1.53 .61 Average Prosocial 69 0.00 6.00 3.85 1.65 -.82 .29 -.010 .57 Externalizing behaviour scorea 89 1.00 2.54 1.53 .36 0.87 .26 0.59 .51
Note. Prosocial behaviour were on a scale of 0-6.
a Externalizing behaviour were on a scale of 1-3.
b Data was missing for toddlers or parents who did not complete the tasks. Early Self-Conscious Emotions and Prosocial Behaviour
Hierarchical multiple regression analysis with age as a covariate in the first step was used to test if (1) guilt and (2) shame-related behaviour were significantly related to how prosocial toddlers were. The results of the regression indicated that the three predictors explained 16.5% of the variance in prosocial behaviour, R2= .165, F(5, 40)= 1.58, p =.187. In the first step, age was
added as a predictor of prosocial behaviours. This model was significant r(42) = .33, p = .025. In the second step, shame and guilt were added as predictors. In the second step, these predictors
did not significantly add to the prediction of the outcome r(39) = .37, F(3,41) = .47, p = .702. It was found that neither guilt or shame in toddlers significantly related to how prosocial they are.
Upon further investigation, instrumental helping was more influenced by shame and guilt than empathetic helping. Shame and guilt significantly explained 36.8% of the variance in instrumental helping, R2= .368, F(5, 38)= 4.42, p =.003, as opposed to insignificantly explaining
7.6% of empathetic helping, R2= .076, F(4, 40)= .83, p =.516. See Table 3 for the beta
coefficients for prosocial behaviour.
Table 3. Coefficients between Age, Guilt & Shame, and Prosocial Behaviour
Table 3
Toddler’s Age, Guilt, Shame, and their relations to Prosocial Behaviour
Variable Unstandardized
𝜷𝜷 Coefficient SE Standardized 𝜷𝜷 t Significance
Age .43 .19 .33 2.32 .008 Latency to Talk/Show (shame) -.38 .67 -.11 -.57 .57 Aversion (shame) -.003 .010 -.042 -.28 .78 Guilt .17 .42 .083 .39 .70
Early Self-Conscious Emotions and Externalizing Behaviour
Next, hierarchical multiple regression analysis with age as a covariate showed that the three measures of self-conscious emotions in toddlers trended towards significantly predicting
Comment [1]: can you report
which predictors exactly were significant with the beta coefficients because this is not clear now
16% of the variance in their externalizing behaviour, R2= .16, F(4, 52)= 2.48, p =.055. Toddler’s
guilt also trended towards significantly relating to less externalizing behaviours, β = -.33, t(4,52),
p =.077. Additionally, toddlers’ shame related behaviours were not significantly related to
externalizing behaviour. See Table 4 for the beta coefficients for externalizing behaviour.
Table 4. Coefficients between Age, Guilt & Shame, and Externalizing Behaviour
Table 4
Toddler’s Age, Guilt, Shame, and their relations to Externalizing Behaviour
Variable Unstandardized
𝜷𝜷 Coefficient SE Standardized 𝜷𝜷 t Significance
Age -.069 .041 -.22 -1.70 .094 Latency to Talk/Show (shame) -.026 .13 -.034 -.20 .84 Aversion (shame) .003 .002 .19 1.45 .15 Guilt -.15 .085 -.33 -1.80 .077
Lastly, as more prosocial toddlers were predicted to display less externalizing behaviours, linear regression analysis were run with instrumental, and empathetic helping as independent variables and externalizing behaviours as the dependent variable. A discrepancy was found between the types of prosocial behaviour; instrumental helping, R2=.043, F(1,63) = 2.80, p =
being significantly related to less externalizing problem behaviour. See Table 5 for beta coefficients between the different types of helping and externalizing behaviour.
Table 5. Coefficients between Prosocial and Externalizing Behaviour
Table 4
Toddler’s Empathetic and Instrumental Helping and their relations to Externalizing Behaviour
Prosocial Behaviour Type
Unstandardized
𝜷𝜷 Coefficient SE Standardized 𝜷𝜷 t Significance Empathetic
Helping -.003 .022 -.018 -.143 .89
Instrumental
Helping -.037 .022 -.21 -.1.68 .099
Discussion
This study investigated the relation between self-conscious emotions in toddlers and social behaviour using observational methods for the first time. The main findings were that toddlers who felt more guilt also had a tendency towards significantly less externalizing problem behaviour. In addition, shame-related latency to talk also negatively corresponded to guilt-related behaviour, indicating like past research that shame and guilt are two distinct self-conscious emotions (Lewis, 1971). Lastly, prosocial behaviour did not seem to be influenced by self-conscious emotions, but as toddlers became older they did become more prosocial. This is interesting because it was hypothesized that more prosocial toddlers would display fewer
externalizing problems, as being helpful requires empathy which are typically lacking when individuals have externalizing problems.
First of all, guilt and shame do appear to develop throughout toddlerhood and increase with age. Toddlers also start to be increasingly prosocial during this time. However, further findings suggest that guilt and shame from a young age do not relate to how helpfully toddlers behave. This contradicts the first hypothesis as well as previous research which finds that at least guilt should be related to more helping behaviour, and particularly more empathetic helping behaviour (Drummond et al., 2016). Whereas the current study found that instrumental helping was significantly related to the predictors, shame and guilt. Methodological differences could account for the difference between this study and that of Drummond and colleagues (2016). As Drummond and colleagues divided their sample into shame and guilt-prone toddlers and observed effects in groups, whereas this study observed results in proportions of shame and guilt-related behaviour which co-occurred (Drummond et al., 2016). So, the results of this study may be more nuanced in terms of which combinations of emotion effect which behaviour.
Additionally, it is also plausible that the effects were greater for instrumental helping than for empathetic helping for methodological reasons within this study. For the empathetic task, the child was supposed to give the test leader a blanket because they were cold. However, during the summer months this premise of being cold may not have made sense to the toddlers with the heat in the unairconditioned space. Meanwhile, helping with a fallen roll of tape while having to hold up a poster is a more acceptable premise for helping all-year round. Hence, there may have been
a discrepancy between how believable the instrumental helping task seemed compared to the empathetic one, contributing to how willing the toddlers were to help the test leader during the empathetic task.
A final explanation for toddler’s guilt and shame not influencing how empathetically prosocial they are could be theoretical. Instrumental helping behaviour have been found to simply develop earlier than empathetic helping, as instrumental behaviour have been explained as stemming from a larger array of cognitions, emotions and desires than helping that stems solely from empathy (Svetlova, Nichols & Brownell, 2010). For instance, instrumental helping in toddlers can come from just wanting to participate in a social activity with an adult, mimicking adult helping behaviour, seeking adult approval, and finding the task fun rather than being concerned for the adult’s wellbeing (Svetlova, Nichols & Brownell, 2010). So, toddler’s helping in instrumental ways may in general already be more common than empathetic helping, making it easier to find an effect in a smaller sample.
Interestingly, shame was also not related to less prosocial behaviour as was predicted by the first hypothesis. So, it seems that bodily aversions and waiting to talk about the mishap after a moral transgression are not related to how much toddlers help in a later task. Since there are only a few studies that separate the various shame-related behaviour in toddlers, and none to date on how these relate to prosocial behaviour, this finding can only be explained through
speculation (Walter & LaFreniere, 2007). One possibility is that in predominantly Dutch, middle-upper class toddlers with highly engaged parents have more protective factors in their
environment that prevent shame from contributing to how prosocial they are. Altogether, it is unclear why shame in toddlers did not influence how prosocial they were.
On the whole, as predicted by the last hypothesis, more prosocial toddlers were also less inclined towards externalizing behaviour. Theories explain that this occurs because helping requires empathy and conscience, which are typically lacking in people with a tendency to externalize (Menesini & Camodeca, 2008). So, being prosocial and externalizing behaviour are unlikely to co-occur. Interestingly, if it is indeed the case that instrumental helping is more linked to less externalizing than empathetic helping as was found in this study, more research is needed to explain this because empathetic helping by definition requires empathy and
conscience, while instrumental helping does not necessarily. Perhaps toddlers simply help in instrumental ways more often than in empathetic ways, as instrumental helping is easier and develops sooner. So, toddlers who help more instrumentally may have the self-concept that they are helpful in general, and thus experience more cognitive dissonance if they were to display externalizing behaviour, which discourages them from behaving as such as it does not fit with their newly developing self-image. In short, toddlers who were more helpful also tend to be less aggressive, but interestingly enough not for the type of help that requires empathy. However, self-conscious emotions like shame and guilt do not seem to influence how prosocial toddlers were.
On the other hand, self-conscious emotions were more related to externalizing behaviour. Specifically, toddlers who felt more guilt also behaved had a tendency towards behaving
significantly less aggressively in general. This aligns with the second hypothesis that more toddler’s guilt should at least not relate to more externalizing behaviour, and perhaps even relate to more prosocial behaviour. This is because toddlers who feel more guilt feel more empathy towards the one they transgressed, they feel like they can repair their wrongdoing, and do not feel as though their entire identity is at fault for the mishap which could contribute to humiliated fury (Lewis, Alessandri & Sullivan, 1992). As toddlers who feel guilty inherently do so out of empathy, they also display less externalizing problem behaviour since this usually requires a lack of empathy (Tangney et al., 1996). Some studies even find that toddlers showing more guilt are more likely to be prosocial in general even beyond guilt-related repairing behaviour after a moral transgression (Drummond et al., 2016). Perhaps in a study with more power, this finding would have reached significance.
Furthermore, toddler’s shame-related behaviours were unexpectedly insignificantly related to more externalizing problems. Previous well-known research has commonly observed how shame-related behaviours typically relate to more externalizing problems (Tangney et al., 1996). Perhaps this discrepancy could be explained by a difference in samples used. In this study, upper-middle class, predominantly European toddlers ages 2-5 were observed, while Tangney and colleagues had a sample of American children ranging from grade 4 to adults (1996). The more diverse sample of American participants from suburban public schools (47% Caucasian, 41% African-American, 8% other) than this study’s sample of predominantly Caucasian, upper-middle class toddlers with highly-involved parents, may have contributed to
higher levels of externalizing problem behaviours in the study by Tangney and colleagues (1996). Indeed, a study on American 1st-graders using the CBCL, showed that
African-American toddlers were exposed to more racial socialization practices (like parents’ willingness to intervene on delinquent behaviour, and mistrust of races outside their own) which related to more externalizing behaviour problems in toddlers (Caughy, Nettles, O'Campo & Lohrfink, 2006). Hence, higher levels of externalizing behaviours in a more diverse sample may have facilitated the finding of a significant relation between shame and externalizing behaviours. Perhaps with a more diverse and larger sample, this study may also have found significant relations between toddler’s self-conscious emotions and externalizing problems.
One strong point of this study was that observational methods were used on a young sample of participants together with parent reports. This makes the findings arguably more objective than those solely relying on parent-reports for this age-group. Another strength was the combination of observations with standardized parent-reports on externalizing behaviour in general, as this made it possible to compare parents reports to each other. That said, a notable weakness was that the sample size was limited, and the missing data was not random enough to be imputed. Additionally, the reliance on parental reports of children’s externalizing behaviour may not be representative of the full-picture, as toddlers may behave differently in the context of day-cares and schools with their peers. Another limitation was that parents were from higher socio-economic backgrounds, perhaps because they were recruited from day-cares, Facebook groups, and schools. Parents with higher economic status who take the initiative to involve their
children in these studies may be particularly highly involved with their children’s upbringing (Rigby, 1993). This may skew the results as parental involvement may enhance family functioning, which when high has been linked to less externalizing behaviour like bullying and more helping behaviour in children (Rigby, 1993). Toddlers in this sample may thus be higher than the general population in helping and lower in externalizing behaviour, making the results only generalizable to other toddlers of highly-involved parents (Rigby, 1993).
For the future, studying larger samples recruited from a greater variety of socio-economic backgrounds and schools would benefit how much is understood about how the typical toddler’s self-conscious emotions develop alongside their social behaviour. Also, using teacher reports on externalizing behaviour, or even observations from home and school contexts, alongside parent-reports may also contribute to a fuller picture of the toddler’s externalizing behaviour. Lastly, perhaps using a different empathetic helping task that is not influenced by external factors like room temperature is a more valid way to discern differences between instrumental and empathetic helping behaviour. This study has shown that observational methods are certainly promising in observing younger samples and observing this development from the start. As for parents and teachers, it may be beneficial to inform them on how to recognize shame and guilt-related behaviour in order to help them address these emotions. In the case of shame-guilt-related behaviour like latency to talk and especially avoidance after transgressing a norm, parents and teachers could encourage reparative behaviour, and thereby hamper the development of further externalizing behaviour. Perhaps fostering guilt rather than shame in schooling and upbringing in
general should be considered, if reducing externalizing problems that could lead to costly lifelong delinquency is considered a goal. Nonetheless, more research is still needed on how to best intervene without causing more harm than good.
In conclusion, it is important to investigate from a young age the underlying self-conscious behaviour that contribute to moral behaviour that can have a lifelong impact on societal violence, as aggression is most pervasive when it starts young (Thijssen, 2016). Using observational methods combined with parental reports on larger and more diverse samples, looking at the specific sub-behaviour that make up shame and guilt may be a useful way forward in the research on toddler’s self-conscious emotions and social behaviour. This study found that neither toddler’s guilt nor shame were implicated in how prosocial they were. However, more guilt in toddlers was related to fewer externalizing problems, and shame-related aversion behaviour were nearly related to more externalizing problems. So, it seems that guilt may be the more adaptive self-conscious emotion in toddlers as it is linked to lower levels of externalizing problems than shame. In light of these findings, perhaps parents and teachers should focus on fostering guilt in from a young age, as it seems to contribute to less undesirable social behaviour in toddlers that could escalate over time.
References
Achenbach, T., & Rescorla, L. (2000). Manual for the ASEBA preschool forms & profiles (pp. 74-100). Burlington, Vt.: ASEBA.
Alink, L., Mesman, J., van Zeijl, J., Stolk, M., Juffer, F., & Koot, H. et al. (2006). The early childhood aggression curve: Development of physical aggression in 10- to
50-month-old children. Child Development, 77(4), 954-966. doi: 10.1111/j.1467-8624.2006.00912.x
Barrett, D. (1979). Relations between aggressive and prosocial behaviour in children. The
Barrett, K., Zahn-Waxler, C., & Cole, P. (1993). Avoiders vs. Amenders: Implications for the investigation of guilt and shame during toddlerhood?. Cognition & Emotion, 7(6), 481-505. doi: 10.1080/02699939308409201
Chaplin, T. and Aldao, A. (2013). Gender differences in emotion expression in children: A meta-analytic review. Psychological Bulletin, 139(4), pp.735-765.
Caughy, M., Nettles, S., O'Campo, P., & Lohrfink, K. (2006). Neighborhood matters: Racial socialization of African American children. Child Development, 77(5), 1220-1236. doi: 10.1111/j.1467-8624.2006.00930.x
de Hooge, I., Zeelenberg, M., & Breugelmans, S. (2007). Moral sentiments and cooperation: Differential influences of shame and guilt. Cognition & Emotion, 21(5), 1025-1042. doi: 10.1080/02699930600980874
de la Osa, N., Granero, R., Trepat, E., Domenech, J. and Ezpeleta, L. (2015). The discriminative capacity of CBCL/1½-5-DSM5 scales to identify disruptive and internalizing disorders in preschool children. European Child & Adolescent Psychiatry, 25(1), pp.17-23.
Desjardins, J. (2018). Here's how much violence costs economies around the world. Retrieved from
https://www.businessinsider.com/heres-how-much-violence-costs-economies-around-the- world-2018-6
Drummond, J., Hammond, S., Satlof-Bedrick, E., Waugh, W., & Brownell, C. (2016). Helping the one you hurt: Toddlers’ rudimentary guilt, shame, and prosocial behaviour after
harming another. Child Development, 88(4), 1382-1397. doi: 10.1111/cdev.12653 Eisenberg, N., Fabes, R., Karbon, M., Murphy, B., Wosinski, M., & Polazzi, L. (1996). The
relations of children's dispositional prosocial behaviour to emotionality, regulation, and social functioning. Child Development, 67(3), 974-992. doi:
10.1111/j.1467-8624.1996.tb01777.x
Eisenberg, N., Spinrad, T., & Knafo-Noam, A. (2015). Prosocial development. Handbook Of
Child Psychology And Developmental Science, 1-47. doi:
10.1002/9781118963418.childpsy315
Ferguson, T., Stegge, H., & Damhuis, I. (1991). Children's understanding of guilt and shame.
Child Development, 62(4), 827-839. doi: 10.1111/j.1467-8624.1991.tb01572.x
Flouri, E., & Sarmadi, Z. (2016). Prosocial behaviour and childhood trajectories of internalizing and externalizing problems: The role of neighbourhood and school contexts.
Developmental Psychology, 52(2), 253-258. doi: 10.1037/dev0000076
Frick, P., & White, S. (2008). Research review: The importance of callous-unemotional traits for developmental models of aggressive and antisocial behaviour. The Journal of Child
Psychology and Psychiatry, 49(4), 359-375. doi: 10.1111/j.14697610.2007.01862.x
Furukawa, E., Tangney, J., & Higashibara, F. (2012). Cross-cultural continuities and
discontinuities in shame, guilt, and pride: A study of children residing in Japan, Korea and the USA. Self And Identity, 11(1), 90-113. doi: 10.1080/15298868.2010.512748 Harris, L. and Brown, G. (2010). Mixing interview and questionnaire methods: Practical
problems in aligning data. Practical Assessment, Research & Evaluation, 15(1). Hoffman, M. (2000). Empathy and moral development (pp. 128-134). Cambridge: Cambridge
University Press.
Hubbard, J., Parker, E., Ramsden, S., Flanagan, K., Relyea, N., Dearing, K., Smithmyer, C., Simons, R. and Hyde, C. (2004). The relations among observational, physiological, and self-report measures of children's anger. Social Development, 13(1), pp.14-39.
I am Expat (2019). Dutch statistics & Facts about the Netherlands. [online] Iamexpat.nl. Available at: https://www.iamexpat.nl/expat-info/the-netherlands/dutch-statistics [Accessed 7 Jun. 2019].
Jolliffe, D., Farrington, D., Piquero, A., Loeber, R. and Hill, K. (2017). Systematic review of early risk factors for life-course-persistent, adolescence-limited, and late-onset offenders in prospective longitudinal studies. Aggression and Violent Behavior, 33, pp.15-23. Kochanska, G., Gross, J., Lin, M., & Nichols, K. (2002). Guilt in young children: Development,
determinants, and relations with a broader system of standards. Child Development,
73(2), 461-482. doi: 10.1111/1467-8624.00418
Lewis, H. (1971). Shame and guilt in neurosis. Psychoanalytic Review, 58(3), 419-433. Lewis, M., Alessandri, S., & Sullivan, M. (1992). Differences in shame and pride as a function
of children's gender and task difficulty. Child Development, 63(3), 630. doi: 10.2307/1131351
with bullying, victimization and prosocial behaviour. British Journal Of Developmental
Psychology, 26(2), 183-196. doi: 10.1348/026151007x205281
Muris, P., & Meesters, C. (2013). Small or big in the eyes of the other: On the developmental psychopathology of self-conscious emotions as shame, guilt, and pride. Clinical Child
And Family Psychology Review, 17(1), 19-40. doi: 10.1007/s10567-013-0137-z
Rigby, K. (1993). School children's perceptions of their families and parents as a function of peer relations. The Journal Of Genetic Psychology, 154(4), 501-513. doi:
10.1080/00221325.1993.9914748
Roos, S., Salmivalli, C., & Hodges, E. (2013). Do guilt and shame-proneness differentially predict prosocial, aggressive, and withdrawn behaviour during early adolescence?.
Developmental Psychology, 50(3), 941-946. doi: 10.1037/a0033904.supp
Stipek, D., Gralinski, J., & Kopp, C. (1990). Self-concept development in the toddler years.
Developmental Psychology, 26(6), 972-977. doi: 10.1037//0012-1649.26.6.972
Stuewig, J., Tangney, J., Heigel, C., Harty, L., & McCloskey, L. (2010). Shaming, blaming, and maiming: Functional links among the moral emotions, externalization of blame, and aggression. Journal Of Research In Personality, 44(1), 91-102. doi:
10.1016/j.jrp.2009.12.005
Spruit, A., Schalkwijk, F., van Vugt, E. and Stams, G. (2016). The relation between self-conscious emotions and delinquency: A meta-analysis. Aggression and Violent
Svetlova, M., Nichols, S., & Brownell, C. (2010). Toddlers’ prosocial behaviour: From instrumental to empathic to altruistic helping. Child Development, 81(6), 1814-1827. doi: 10.1111/j.1467-8624.2010.01512.x
Tangney, J, & Dearing, R. (2002). Shame and guilt. Encyclopedia of Crime and Punishment. doi: 10.4135/9781412950664.n388
Tangney, J., & Fischer, K. (1995). Self-Conscious Emotions. New York: Guilford Press.
Tangney, J., Wagner, P., Hill-Barlow, D., Marschall, D., & Gramzow, R. (1996). Relation of shame and guilt to constructive versus destructive responses to anger across the lifespan.
Journal Of Personality And Social Psychology, 70(4), 797-809. doi:
10.1037//0022-3514.70.4.797
Thijssen, J. (2016). Children with externalizing behaviour problems: Risk factors and preventive
efforts (pp. 105-129). Maastricht: Ridderprint BV.
Thomaes, S., Stegge, H. and Olthof, T. (2007). Externalizing shame responses in children: The role of fragile-positive self-esteem. British Journal of Developmental Psychology, 25(4), pp.559-577.
Tremblay, R. (2000). The development of aggressive behaviour during childhood: What have we learned in the past century?. International Journal Of Behavioral Development, 24(2), 129-141. doi: 10.1080/016502500383232
for emotional adjustment, guilt, and shame. North American Journal Of Psychology, 9(1), 111-130.
Wilcox, R. (2005). Trimming and Winsorization. Encyclopedia Of Biostatistics. doi: 10.1002/0470011815.b2a15165
Zahn-Waxler, C., Cole, P., Welsh, J., & Fox, N. (1995). Psychophysiological correlates of empathy and prosocial behaviour in preschool children with behaviour problems.
Development And Psychopathology, 7(1), 27. doi: 10.1017/s0954579400006325
Zimmerman, P., Bolhuis, J., Willemsen, A., Meyer, E., & Noldus, L. (2009). The Observer XT: A tool for the integration and synchronization of multimodal signals. Behavior Research
