Promoting social reasoning about intergroup conflicts through small-group discussions of literary fiction: A two-year intervention study

Video-recorded small-group discussions involving participating fourth- and fifth-grade children and the teacher facilitating the discussion. Observational coding focused on teacher-child and peer interactions as well as children’s individual contributions to the discussions.

Children were recruited through participating school classes and assigned to small discussion groups within their classes. Three to four small-group discussions per classroom were recorded at each of three measurement points (T0, T1, T2). Participation in the study and video recording was subject to informed consent.

The small-group discussions lasted approximately 15 minutes and were video-recorded. The recordings were subsequently coded using the CLASS Upper Elementary framework to assess interaction quality and an adapted Quality Talk coding scheme to assess features of children’s discourse, including questions, reasoning, challenges, build-ons, and unreasoned claims. Coding generated quantitative variables that are included in the public student-level dataset. The video recordings themselves are not included in the public dataset.

Written justifications provided by participating fourth- and fifth-grade children in response to moral dilemma stories administered at up to three measurement points (T0, T1, T2). The unit of analysis was the individual child’s written response.

The written responses were obtained from children participating in the SKiLL study who completed the respective assessment at a given measurement point. Children were recruited through participating school classes. Participation required informed consent; availability of written responses additionally depended on attendance and completion of the task.

Children’s written justifications were coded for individual moral reasoning. Justification units were classified as moral or group-based based on the Social Reasoning Development framework, with additional subcategories refined during coding. Moral justifications referred to concerns such as fairness, harm, welfare, and perspective-taking, whereas group-based justifications referred to loyalty demands and anticipated social consequences for the peer group or close friends. All written responses were double-coded. Interrater reliability was ICC = .92 for moral justifications and ICC = .90 for group-based justifications. The public dataset contains the derived coded variables, including counts of moral and group-based justifications, word counts, and an ordinal indicator of whether responses contained no justification, a single-domain justification, or a dual-domain justification. The original written responses are not publicly shared.

Participating fourth- and fifth-grade children. The unit of observation was the individual child. Psychological assessment included reading comprehension and performance-based social perspective taking.

Children were recruited through participating primary school classes in German-speaking Switzerland. Participation required informed consent. The assessments were administered to participating children who were present at the respective measurement occasion.

Two psychological performance measures were administered. Reading comprehension was assessed at T0 and T2 using a research beta version of ELFE III, a revision of the standardized ELFE II reading comprehension test. The test was administered on tablets in a multiple-choice format and included separate timed subtasks assessing word-, sentence-, and text-level reading comprehension. Scores represent the number of correctly solved tasks within each subtest. Social perspective-taking performance was assessed at T0 and T2 using a Swiss adaptation of the Assessment of Social Perspective Taking Performance (ASPP; Kim et al., 2018). Children responded in writing to hypothetical social scenarios involving actors with different social roles and perspectives. Responses were coded for articulation, defined as the number of different actors’ perspectives articulated, and positioning, defined as the use of social roles, circumstances, or prior experiences to contextualize these perspectives.

Participating fourth- and fifth-grade children. The unit of observation was the individual child. Data comprised self-reported questionnaire responses and peer nominations.

Children were recruited through participating primary school classes in German-speaking Switzerland. Participation required informed consent. Questionnaires and peer nominations were administered to participating children who were present at the respective assessment occasion.

Standardized self-report questionnaires and peer nominations were administered individually on tablets using SurveyCTO. The questionnaires assessed need for cognition, cooperative and competitive learning attitudes, conflict-resolution strategies, intergroup attitudes and empathy, inclusive classroom norms, intergroup friendships, and responses to hypothetical intergroup situations, including evaluations of exclusion and bystander behavior. Peer nominations assessed friendship, peer acceptance and rejection, popularity, relational and physical aggression, and prosocial behavior. The public dataset contains item-level variables as well as derived scale and peer-nomination scores. Detailed information on items, response coding, measurement points, and score construction is provided in the accompanying codebook.