Paper Review: Children’s and adolescents’ moral emotion attributions and judgements about exclusion of peers with hearing impairments

Chilver-Stainer, C., Gasser, L., & Perrig-Chiello, P. (2014). Children’s and adolescents’ moral emotion attributions and judgements about exclusion of peers with hearing impairments, Journal of Moral Education, 43, 3, 235-249, DOI:10.1080/03057240.2014.913515

Browsing Research Gate earlier this week, in the name of “constructive procrastination” (read: putting off re-analyzing some data for a little while longer), I discovered the above paper. One key criticism of my group-based emotion and social exclusion research is that it is very experimental (the groups in my studies are contrived by me, and don’t ever interact with one another for the sake of experimental control), and I am keen to look at what goes on when researchers test actual, real-life groups, so this one was worth following up.

So what happened here? Well, the researchers looked at how  215 Swiss 10-, 12- and 15-year-olds attending mainstream school felt about, and judged, the social exclusion of peers with hearing impairments.This sample is worthy of comment before I describe the study. It is said in the paper that there were a number of children with hearing impairment in the children’s class – yet (while these children’s parental approval of the study was sought) these children did not participate. This struck me, because I am asked by our ethics board to include all children in my research on disability and inclusion (and have tested those with disabilities) – and even if I hadn’t been, I wouldn’t want to exclude children with disabilities. Rather, even if numerically less significant, the qualitative responses of children with hearing impairment to these scenarios would have been interesting. This is another case where half the research design is apparently missing.

Moving on. Scenarios of social exclusion in different contexts were shown to the children.  There were four different scenarios, where a hearing child and a child with hearing impairment wanted to join an activity, but unfortunately, there was only room left for one more child. The protagonist decided to pick the hearing child in doing so excluded the child with hearing impairment.  (Note: the constant exclusion of the child with hearing impairment means that we cannot be sure that later measures relate to exclusion based on disability per se, as we have nothing to compare it to). The scenarios described either (a) a group of children preparing a presentation (b) a child doing oral homework, (c) a birthday party or (d) talking with a child about shared interests. The activities always involved oral communication. Nevertheless, participants were told that the hearing child and the child with hearing impairment were equally qualified for the activity. The contextual variables here were school versus home, and group versus dyadic interaction.

Next, participants were asked what emotions the protagonist might feel with respect to the exclusion, and how they judged their behaviour. So the two questions were: ‘How do you think Erwin [the excluder] will feel? Why?’. Participants could select one or two of the following emotions: proud, happy, sad, neutral, angry, fearful, guilty, ashamed or empathetic, and ‘What do you think? Is it good or bad that Oliver chose Rolf [the hearing child]? Why?’. There are several things to say about these two questions, even though they are derived from previous research (just because everyone else is doing it….).

Firstly, the first question is forced choice. Children were asked to select one or two emotions. This limits their responses: one could quite easily imagine a situation where a child might feel guilty and sad about having to exclude the child – as well as feeling fearful of the repercussions, and angry that they had to choose in the first place. Rather then, emotions could have been measured on Likert-type scales – ‘To what extent does X feel?’.. This would allow for neutral judgments of some emotions, whilst allowing children to indicate the ones they would expect the excluder to feel most strongly (leading to a more sensitive index). Furthermore, in later analysis, emotions were classified as moral (sad, guilty, ashamed, empathetic) or amoral (proud, happy, neutral, angry, fearful). In the adult emotion literature I am aware of pride and anger are very much moral emotions (cf. Saab, Tausch, Spears, & Cheung, 2014Tangney,  Stuewig & Mashek, 2007).

Secondly, the  question, about moral judgements. One could argue that this is a leading question – is it good or bad? Some actions might be morally neutral. Yet this is not, apparently, an option. One also wonders what children might be led to say, on the basis of having a teacher in the classroom (neutral territory?), supervising the research. This question is highly susceptible to the problem of social desirability bias (for this group of children at least: of course it’s wrong!) and also to audience effects. Who are the children responding to? The teacher? The school? The researcher? Other children? As research shows, who you tell participants will see your answers, makes a dramatic difference to what they report (e.g., Rutland, Cameron, Milne, & McGeorge, 2005). Moreover, classes differ in their inclusivity norms (cf. Glasser et al., 2013) and this point of reference would be worthy of future study.

Finally, participants inclusive behaviour was assessed through  peer nomination of those in the class who “let other children participate”. This was a good idea. Having peer nominations as well as self-report reduces shared method variance, and gives a different perspective from the child’s own on how inclusive they are. The inclusive behaviour measure also brings these researchers closer to measuring actual behaviour than most of the scenario-based literature in this field. Unfortunately, participants were then split into not-inclusive or inclusive, based on the mean scores given from this task (rather than absolute scores, as used in a lot of the bullying literature) to ensure that some children fell into each category. I would argue that retaining the continuous nature of the initial variable would have led to a richer pattern of results.

Quentin Blake argues for greater inclusion of disabled characters in children's books.

Quentin Blake argues for greater inclusion of disabled characters in children’s books. Image from:

Another potential strength of this study is its mixed method approach. As you will have noted from above, the authors asked ‘why’ a character might feel that way, and ‘why’ the children felt that way about him or her. Unfortunately, as the authors note, the increasingly elaborate answers with children’s age likely reflects their writing, rather than their reasoning ability. Nevertheless, it was found, in line with prior research that reasoning correlated with emotion, was particularly astute among more inclusive children, and fell into one of two categories; moral reasons for feeling bad following exclusion (e.g., equal rights for all), social-functioning reasons for exclusion (e.g., he would slow the group down). Examples of negative moral reasoning (it was the most efficient thing to do) and positive social-function reasoning (he would have added something new to the group) were not given in the paper. If this dichotomy exists (if children did justify exclusion on moral grounds, or encourage inclusion on social-functioning grounds) a four-way split of the data would have been interesting to examine.

It is perhaps the latter of these categories (social-function reasoning for exclusion) that demands further exploration, since it reflects children’s beliefs about when exclusion is OK. As noted above, the scenarios cited the oral nature of the interaction – which children with a hearing impairment might find challenging. Given that children were told they could equally cope with the task, if one took the stance that it is never OK to exclude someone with a hearing impairment, then modifying these beliefs associated with this kind of interaction would be a key focus for intervention. Again, however, another element of the research design is lacking. Research shows that children misperceive disabilities such that impairment “spreads” (eg., Abrams et al., 1990). So, for example, in my research, children have said that a child in a wheelchair because she is unable to walk wouldn’t be able to play a musical instrument. Thus, measuring children’s exclusion justifications where a child with a hearing impairment was excluded in a non-oral interaction would be worth looking at.

It would also be interesting to take further advantage of the open-ended questioning to look at ways in which children think about overcoming the exclusion (e.g., how could you include X in the group?). I present children with a “winning ticket” scenario (the child has four tickets, and four friends) similar to the group-leisure scenario above. We discuss in groups how the friend left without a ticket might feel, and what can be done to mend those feelings. And the children come up with very creative ways of sorting things out, even when I emphasise the fact that there are only four tickets to be had. Examining children’s strategies for dealing with unfair exclusion, in order to enhance these would also be worthy of future research.

Bringing this together then, this study looked at hearing children’s responses to the reported forced-choice exclusion of a child with a hearing impairment from either a leisure or school, group or dyadic interaction, as given in a scenario. Additionally, and unlike previous research, children’s actual inclusion behaviour was also examined. And the evidence showed that moral emotions were linked to moral justifications in a similar way to prior research, with inclusive children being more in tune with this. However, this is a piece of a much larger jigsaw. We must now ask many questions: how do reasons differ for positive versus negative emotions? What audience are the children responding to? What do child with a hearing impairment think? Does the nature of the interaction matter? And what happens afterwards? Exclusion isn’t the end of the story.


Paper Review: Cooley and Killen (in press): Children’s Evaluations of Resource Allocation in the Context of Group Norms

Cooley, S., & Killen, M. (in press). Children’s evaluations of resource allocation in the context of group norms. Developmental Psychology.

What do you do when your group has a norm for fairness and equality, and someone in your group advocates for inequality that would benefit the group? Extant research shows that, for very young children, equality trumps all other values, including group loyalty and ingroup bias but that, as we get older, peer relations, existing resources, and a whole host of other factors will come into play (e.g., Blake & McAuliffe, 2011). In their paper, Cooley and Killen deal with the tension between moral values and ingroup bias, in a kindergarten setting. Children, aged three and a half to six years were asked to give their opinion of a classmate [deviant] who either (a) advocated for unequal shares of building blocks when the class had a norm for equal sharing, or (b) advocated for equal sharing of building blocks when the class had a norm for sharing the blocks in its favour (i.e., giving itself more blocks). What happened?

Cooley and Killen report that most children had a dim view of children who went against the class norm by dividing things unequally, even when it benefitted their class; they explained their opinion in terms of the importance of fairness. However, participants who appreciated those classmates who shared resources in the class’s favour reasoned about group functioning and the group benefits . With age, children displayed increasing social group nous by differentiating their own opinion of the classmate who runs counter to class norms from their expectations of the group’s likely opinion of this classmate.

This research finding is notable for several reasons. Previous research looking at fairness and peer relations in this age group had done so  in terms of dyads. For example, Olson and Spelke (2008) found that found that three and a half year-olds allocated equal resources to puppet friend dyads more often than to non-friend dyads. And, Paulus and Moore (2014) demonstrated that preschool aged children expected a person to share more with a friend than a disliked peer. More often, however, children’s experiences of peer relations are packaged in terms of groups. Children learn about group identities, like gender and ethnicity from the preschool years, and a plethora of research has shown how these categories shape their opinions of others, as well as their choices of playmate and classroom behaviour (for a review see Bennett & Sani, 2004). This research paper speaks to group processes.

And this research is more “real” than its predecessors. Often such research makes use of  a minimal group paradigm (where judgements of others and resource allocation are based solely on group membership) where the minimal groups are contrived at random,  at the start of the study. Here, children made judgements about their membership class (be it for example the red class), and an outgroup [orange] class. These were actual classes in their school and judgements were made on the basis of a photo’ board of the real class. However, judgements seem to have been made about a fictional member of the class. So, there was no prior knowledge of that individual characteristics about them, upon which judgements could have been based. The study also measured children’s identification with their class (as opposed to simple liking for it). Unfortunately, identification ratings were not then factored into any of the analyses  – and like me, they didn’t measure outgroup identification (oh, to have Dr. Hindsight as a colleague) –  but identification was measured – and found to be high – which is good.


The red room and the orange room from Cooley and Killen (in press) Developmental Psychology


It was interesting to note that the authors found no difference between opinions of the counter-normative classmate, depending on whether he was a member of the child’s class, or another class (e.g., the red class or the orange class). This is interesting to note because it is at odds with literature on the black sheep effect. (e.g., Abrams, Rutland, & Cameron, 2003, Abrams et al., 2013). Such research  tells us that ingroup members think ingroup deviants, like the classmate in these scenarios, should be treated more severely than outgroup deviants, for the same norm-transgression: it is possible that this is because behaving in line with, or in favour of, outgroup members represents more of a threat, than simply not following the group rules. Yet, the authors state,

 ” participants viewed it just as wrong for an ingroup member to deviate from an equal group norm as it was for an outgroup member to do so” (p. 9).

This is worthy of further investigation. It is possible that this null finding is due to  the classes’ actual existence, as opposed to research where this finding is detected in minimal group settings. There again, maybe the black sheep effect itself, in children at least, is an artefact of the minimal group paradigm….It would be worth controlling for this – and also for the “reality” of the colour-coded groupings, which may carry unknown histories or status differentials. As a case in point, when I was in Reception class (UK aged 4-5 years) I was in the red group. There was also a yellow group , a green group, and a blue group – in other words, a rainbow. And the red group were the best at reading, then yellow, and so on. And as 4 year-olds, we all knew this to be the case, even though groupings weren’t used by the teachers to make this clear to us…

A further potential avenue for future exploration surrounds the initial creation, and ingroup knowledge of the ingroup norm – which is unsourced, so far as I can tell. Who is the classmate who deviates transgressing against? The class as a whole, or their teacher, or an aggregate of past exchanges (as seems to be inferred “your group like to….”?) How is this class norm knowledge obtained? Is it displayed prominently in the classroom as a prescriptive norm, or is it a descriptive norm for class action (cf. Hitti et al., 2013). One might also ask at what level the norm of  “equality” is important to the children. If the red class already had ten building blocks, and the orange class none, would they still want to divide new blocks equally? What reasons beyond group functioning could be given a priori to the children for unequal allocation of resources? What if the classmate deviant wanted to address an existing inequality in resource allocation – giving less to the ingroup? Is equality-in-the-moment, or absolute equality what matters?

Further interesting questions remain, now that Cooley and Killen have brought this finding into the group domain. Children were asked whether the class/they as participants would like / want to be friends with the deviant classmate. This leads me to wonder what the effect of the class actively friending or de-friending versus individually friending or de-friending the deviant might be? How would this manifest itself in the classroom(s)?  What does to like/dislike or friend/not-friend a child look like to the children? How would they expect the deviant child to respond? Does the outgroup class know of this like/dislike ? What do they make of it? And, in the case of ingroup deviant dislike, are the boundaries between the [sic] red and orange classrooms permeable? Can children move classes? Would the outgroup class be more likely to accept an ingroup deviant into their fold?  Would the children, given the option, suggest that this should be the case, rather than dislike or de-friending (reported based on the forced choice they were given)?

This research suggests that children aged three and a half to six years privilege equality over group functioning – recognizing differences between their opinions and the likely consensus of the group, as they age. The best kind of research asks more questions than it answers, and this is undoubtedly the case here.There is so much we still have to find out about children’s moral reasoning as group members.

Memory Monday: How are you today?

In the spirit of blogging culture, this morning, as Time to Talk Day 2015 approaches, I’d like to look back on a post I wrote this time last year, and ask what, if anything, has changed.

The original post may be found here.


Well, I still have an open-door policy, and I still see a lot of students in my office with mental health related concerns. And students are still very welcome to come and raise concerns with me; nothing has changed there. Indeed, from where I am sitting, mental health concerns at university are still normative.

But the plural of anecdote….

….is not evidence. So what’s changed, evidence-wise in the past year? Time to check the oracle (read: internet) .

First thing I realize is that since February last year, there has been a huge upsurge of student voices talking about mental health at university. There are many pieces on the taboo that surrounds it, noting, as I did last year that according to the latest NUS survey (2013) that one in five students say that they have a mental health problem, but most stay silent about it. I can’t find evidence (but am happy to stand corrected) of more recent large-scale surveys of UK student mental health. But this year, there are  more stories about mental health at university out there, with The Guardian having an overwhelming response to a request for them – gathering over 200 pieces. True, that the plural of anecdote is not evidence….but maybe the time is ripe for a qualitative study of student experiences…..

It was also interesting to note, on two counts for me, that the conversation has expanded. It’s not just about student mental health anymore, but also about mental health in academia. There is evidence that academics, from PhD students to professors are struggling in high-stressage environments. Alongside this, is the hypothesis that there is a culture of acceptance around mental health problems in the academy: in other words, social psychology is at work – stressage is part of the job.

And recently published research by Ken Mavor and his colleagues (2014) supports this contention. That is, a strong social identity as a medical student is associated with high levels of social support and improved well-being  (strong social identity = good) , but this comes with a set of unhealthy group norms (for overwork et al.) that may have a greater influence on students with a strong social identity, encouraging them to do things that put their well-being at risk (strong social identity = risk for poor mental health). Maybe the same is true of PhD students, top professors, early career researchers…If we cast the latter as peripheral group members  to use Jolanda Jetten’s term (that is, those who are on the edge, and want to be in the group of “established academics”) there would be even more reason to suppose that ECRs would be at risk….there are another two hypotheses to test.

So, what has changed? It seems that people are more vocal these days – and that there are a lot of stories out there about mental health in academia. But, beyond small-scale experimental work, there is not much hard-core evidence on the nature of the problem. Now it has been driven out from underground, and now that the hypotheses are being put forward, the time for up-to-date full-scale research seems to have arrived.

How are you today? The Department of Psychology, Social Work and Public Health will be marking Time to Talk Day at 11am, this Thursday, 5th February. If you would like to join us, drop me an email.

Remember my open door policy, if you have been affected by any of the issues mentioned in this piece – and the variety of support offered by Brookes Well-Being. And you can always contact Samaritans or Nightline for help and support, too.