Do demand characteristics contribute to minimal ingroup preferences?

https://doi.org/10.1016/j.jecp.2020.105043Get rights and content

Highlights

  • Minimal group paradigms assess social preferences arising from group membership.

  • We asked whether demand characteristics contribute to children’s minimal group bias.

  • Previous findings were not replicated in two of three measures of social preference.

  • Results from one measure provide no evidence that demand characteristics interfere.

Abstract

“Minimal group” paradigms investigate social preferences arising from mere group membership. We asked whether demand characteristics contribute to children’s apparent minimal group bias in a preregistered experiment (N = 160). In a group condition, we attempted to replicate findings of bias following assignment to minimal groups. A second closely matched no-group condition retained potential demand characteristics while removing group assignment. Parallel bias in the no-group condition would suggest that demand characteristics contribute to findings of apparent ingroup bias. Three main findings emerged. First, in the group condition, ingroup preference emerged in one of three bias measures only. Second, this preference emerged even though participants evaluated ingroup/outgroup photos varying in race/ethnicity between trials. Third, the measure that yielded ingroup preferences in the group condition produced no parallel bias in the no-group condition, consistent with the view that mere membership in a group, not experimental demand, leads to minimal ingroup preferences.

Introduction

People often prefer their ingroups, both familiar real-world groups and arbitrary groups assigned by experimenters for the course of a study (Dunham, 2018, Mullen et al., 1992, Tajfel et al., 1971). The minimal group paradigm is commonly used to investigate downstream effects of group membership. In the barest forms of the minimal group paradigm, participants are randomly assigned to arbitrary novel social groups with essentially no meaning in hopes of revealing what follows from the mere fact of group membership (Dunham, 2018). Preferences for the ingroup following minimal group assignment are widely reported.

Minimal group studies of young children have become especially prominent, showing that we can possess minimal ingroup preferences before we have had extensive experience with real-world groups. Children show minimal ingroup bias in tasks involving liking/preferences (Baron and Dunham, 2015, Dunham et al., 2011, Elashi and Mills, 2014, Plötner et al., 2015, Richter et al., 2016), helping (Plötner et al., 2015), reputation management (Engelmann et al., 2018, Engelmann et al., 2013), loyalty (Misch et al., 2016, Misch et al., 2018), choosing who to trust (Elashi and Mills, 2014, MacDonald et al., 2013), and recalling/interpreting others’ actions (Dunham et al., 2011, Dunham and Emory, 2014, Schug et al., 2013; see Dunham, 2018, and Otten, 2016, for recent reviews). These biases are also persistent (Yang & Dunham, 2019).

Using the minimal group paradigm to explore the roots of group bias (Dunham, 2018) requires that the paradigm truly allows us to probe effects of group assignment. But most minimal group studies are fertile ground for interference from demand characteristics. For example, experimenters commonly assign children to a group based on color and then have them choose or rate other individuals of the same or a different color. This design could impel participants to choose, or answer favorably about, others wearing the same color they themselves were assigned based on implicit or explicit interpretations of the experimenter’s expectations (Durgin et al., 2009, Gerard and Hoyt, 1974, Orne, 1962).

Could demand characteristics contribute to findings of minimal group bias? Although a few studies have investigated this question, some directly and others indirectly, the relevant evidence is perhaps inconclusive and based mostly on very different methodologies from those typically used in today’s work with young children. St. Claire and Turner (1982) used a prediction condition that maintained the demand characteristics of the paradigm but removed the assignment of participants to groups, asking participants to predict how someone who was in a group would allocate resources to ingroup and outgroup members. They found no evidence of ingroup bias in this prediction condition but did in the control condition (a standard minimal group paradigm), suggesting that group membership, not experimental demand, causes bias. Berkowitz (1994) instead assigned all participants to minimal groups and, on different trial blocks, asked them to allocate resources to ingroup and outgroup members in a way that confirmed the experimenter’s expectations, in a way that went against these expectations, and in whatever way they wanted. Participants generally showed higher levels of ingroup bias when they were responding as they thought they were expected to do than when they were responding counter to these perceived expectations, and true ingroup bias scores in the free response trials correlated positively with bias scores in the confirm-expectations trials. This suggested that participants were able to determine that experimenters thought they would prefer their ingroup and were answering in ways that agreed with these expectations. Thus, these two investigations of demand characteristics in the minimal group paradigm arrived at different conclusions. Gerard and Hoyt (1974) attempted to reduce demand characteristics in their minimal group study, with findings suggesting that there is at least a tendency toward ingroup bias under some circumstances even when demand characteristics are reduced. A recent child minimal group study also arguably reduced the demand characteristics of the paradigm. Yang and Dunham (2019, Study 2) employed a minimal condition in which participants were randomly assigned to a minimal group but were then switched into the other group. This likely highlighted to participants that group membership was truly meaningless and that the researcher thought this as well, which might have made them feel less pressure to prefer their ingroup. Ingroup bias did not appear in the minimal condition on measures of preference and similarity, but it did in a resource allocation task. This result shows the formation of some kinds of minimal group bias in the minimal group paradigm even when demand characteristics are reduced, suggesting that demand characteristics are not the cause of at least certain forms of ingroup bias in this paradigm.

Importantly, potential effects of demand characteristics have not been directly tested in the context of the modern child minimal group paradigm. Even widely supported findings in other areas of experimental psychology have fallen prey to similar complications during recent years (e.g. Durgin et al., 2009, Firestone and Scholl, 2016). Although concerns about demand characteristics may be partially mitigated by recent findings of minimal group bias in implicit measures (e.g., Dunham et al., 2011, Dunham and Emory, 2014), these are not necessarily immune. Moreover, most minimal group studies use explicit measures.

Here, we aimed to determine whether demand characteristics might contribute to the appearance of bias in the child minimal group paradigm. One approach to exploring the potential role of demand characteristics is to remove them from the paradigm and see whether an effect lessens or disappears, thereby suggesting that experimental demand created or contributed to the original effect, as in Durgin and colleagues’ (2009) classic work on behavioral potential influencing perception. Because removing potential demand cues is very difficult in the minimal group paradigm, we used a different approach—setting up a scenario that should not elicit the effect of interest (Firestone & Scholl, 2016). We aimed to preserve potential demand characteristics while removing the typically accepted cause of ingroup bias (minimal group membership). To achieve this, we created a novel condition in which experimenters introduced a salient color to participants but did not assign them to groups or mention groups at all. If group membership is the source of minimal ingroup bias, then bias should not emerge in this condition. If, on the other hand, demand characteristics contribute to apparent ingroup preferences in the minimal group paradigm, then bias might emerge in this novel condition.

Our preregistered study included two main between-participants conditions. The group condition used a typical minimal group paradigm, based on Dunham et al. (2011), in which participants were assigned to a green or orange minimal group and wore a green or orange smock. In the novel no-group condition, participants were assigned a salient colored object (a green or orange box). In both conditions, participants completed three measures to assess their preferences for photos of children wearing green or orange. These measures determined whether participants liked color-match individuals more than nonmatch individuals, whether they attributed more positive behaviors and fewer negative behaviors to color-match individuals, and whether they expected color-match individuals to exhibit more sharing/reciprocity toward them. If participants were biased toward favoring color-match individuals only in the group condition, then this would suggest that group membership, not demand characteristics, causes true ingroup bias in this paradigm. However, if color-match bias emerged in both conditions, then this would suggest that demand characteristics might contribute to the appearance of bias in minimal group studies.

Section snippets

Participants

A total of 165 children participated (80 female and 85 male; Mage = 5.4 years, SD = 0.57, range = 4.5–6.7). Participants were excluded for failure to complete at least two measures (preregistered grounds for exclusion; n = 1), experimenter error (n = 3), or parental interference (n = 1). The final sample consisted of 160 children (78 female and 82 male; Mage = 5.4 years, SD = 0.58, range = 4.5–6.7; n = 47 4-year-olds, n = 85 5-year-olds, and n = 28 6-year-olds), our preregistered sample size,

Results

Deidentified data are available at https://osf.io/925u7/. Our-preregistered analyses had three main goals (https://aspredicted.org/4hd8m.pdf). First, we aimed to determine whether participants in the group condition demonstrated color-match bias for each of the three measures (bias favoring photos of individuals in the participant’s ingroup wearing the same color as the participant). Second, we asked whether participants in the no-group condition displayed color-match bias for each of the three

Discussion

There were three main findings; we focus on the child sample, which was the main emphasis of this study, throughout the Discussion. First, in the group condition, ingroup preference emerged in one of the three bias measures only. Second, this preference emerged even though participants evaluated ingroup/outgroup photos that varied in race/ethnicity between trials. We return to these findings below, focusing first on our main question of interest: Could demand characteristics contribute to the

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

We thank the participants and their families; the Connecticut Science Center; lab coordinators Katherine Williams and Alexandra Zax; research assistants Ilana Ladis, Sarah Corner, Sofia Kinney, Amanda Fiorentino, Maya Layne, Katie Vasquez, Jenny Chelmow, Esha Bhandari, Sarah Aduke Ohiomah, Lauren Barragan, and Sophie Charles; and the Department of Psychology at Wesleyan University. We also thank Yarrow Dunham, Chenmu (Julia) Xing, Royette Dubar, Andrea Patalano, and Valerie Nazzaro and the

References (31)

  • Y. Dunham et al.

    Consequences of “minimal” group affiliations in children

    Child Development

    (2011)
  • Y. Dunham et al.

    Of affect and ambiguity: The emergence of preference for arbitrary ingroups

    Journal of Social Issues

    (2014)
  • F.H. Durgin et al.

    Who is being deceived? The experimental demands of wearing a backpack

    Psychonomic Bulletin & Review

    (2009)
  • J.M. Engelmann et al.

    Concern for group reputation increases prosociality in young children

    Psychological Science

    (2018)
  • J.M. Engelmann et al.

    Young children care more about their reputation with ingroup members and potential reciprocators

    Developmental Science

    (2013)
  • Cited by (4)

    View full text