Sibling composition, executive function, and children’s thinking about mental diversity

Abstract

Prior investigations of relations between sibling composition and theory of mind have focused almost exclusively on false belief understanding in children 6 years of age and younger. The current work expands previous research by examining whether sibling composition predicts 4- to 11-year-olds’ (N = 192) more advanced mental state reasoning on interpretive theory of mind tasks. Even when controlling for age and executive function, children with a greater number of older siblings or with more same-sex siblings demonstrated stronger knowledge in both their predictions and explanations that people with different past experiences can have diverse interpretations of ambiguous stimuli. These data provide some of the first documentation of sibling constellations that predict individual differences in theory of mind during middle childhood.

Introduction

Over the past 20 years, there has been growing empirical interest in sources of individual differences in the development of theory of mind (ToM); that is, understanding of people in relation to their internal mental lives such as desires, emotions, beliefs, and thoughts (Carpendale and Lewis, 2004, Dunn, 2002, Wellman, 2011a). Researchers have identified several cognitive, social, and family factors that explain variability in children’s ToM performance: executive function (Carlson and Moses, 2001, Sabbagh et al., 2006), language (Astington & Jenkins, 1999), cultural background (Shahaeian, Peterson, Slaughter, & Wellman, 2011), maternal education level (Pears & Moses, 2003), family socioeconomic status (Shatz, Diesendruck, Martinez-Beck, & Akar, 2003), parent–child discourse (Hughes et al., 2014, Ontai and Thompson, 2008), parenting behaviors (Hughes & Ensor, 2006), attachment security (Symons & Clark, 2000), and sibling composition (Perner, Ruffman, & Leekam, 1994). Because this research has focused on children 6 years of age and younger, however, it is unclear whether these relations are specific to early childhood. Developmental changes and individual differences in ToM during middle childhood remain greatly understudied (see Lagattuta et al., 2015, Miller, 2012, Pillow, 2012).

The current research expands inquiries of individual differences in social–cognitive development to middle childhood by investigating relations between sibling composition and advanced ToM in 4- to 11-year-olds. Sibling composition provides an interesting test case because several studies have shown that 3- to 6-year-olds with more siblings, especially older siblings, demonstrate superior ToM performance compared with children with fewer or no siblings (e.g., Farhadian et al., 2011, Lewis et al., 1996, McAlister and Peterson, 2006, McAlister and Peterson, 2007, McAlister and Peterson, 2013, Perner et al., 1994, Ruffman et al., 1998, Wolfe et al., 2003; but see Carlson and Moses, 2001, Cole and Mitchell, 2000, Cutting and Dunn, 1999, and Hughes & Ensor, 2005, for examples of null effects). Researchers have interpreted these data to indicate that siblings provide social opportunities that improve young children’s knowledge about mental states, with older siblings being particularly advantageous because they serve as social mentors to the younger “apprentices” (Hughes et al., 2014, Perner et al., 1994, Ruffman et al., 1998). Potentially, however, sibling benefits may appear only during early childhood when mental state concepts are first emerging and children spend most of their waking hours just with family. That is, this well-accepted apprenticeship model may only characterize early development. Moreover, across age, older versus younger status might not matter as much as having siblings above a certain chronological age (age threshold model). By targeting a sample of children across a wide age range, we can test these different accounts as well as determine more generally whether sibling–ToM relations even exist during middle childhood.

There are several reasons to suspect that associations between siblings and ToM may differ in older children as opposed to younger children (Miller, 2012, Recchia and Howe, 2009). First, when children learn new concepts, there can be higher variability in performance both within the same child and within groups of children of the same age group compared with when children have mastered those concepts (Siegler, 2007). Given that most children pass the benchmark false belief task (understand that people can believe things that are not true) by 4 or 5 years of age (Wellman, Cross, & Watson, 2001), it is possible that variables that contributed to individual differences during the initial emergence of ToM no longer hold. Second, when children enter school at 5 or 6 years of age, socialization practices shift such that children begin to spend significantly more time outside of the home with nonfamilial peers, with the number of peer contact hours increasing over childhood into adolescence (Larson & Verma, 1999). The frequency of mental state talk with peers—talk about emotions, desires, beliefs, thoughts, and intentions—predicts individual differences in ToM (see Symons, 2004), with studies indicating that school-age children engage in more mental state talk with friends than with siblings (e.g., Brown, Donelan-McCall, & Dunn, 1996).

This constellation of changes during middle childhood—improved ToM, reduced time with siblings and family, and growing prominence of peers—may reduce or extinguish family-based factors that predicted ToM earlier in development. Specific to our focus on sibling composition, Miller (2013) recently examined whether number of siblings predicted 5- to 8-year-olds’ reasoning about second-order false belief tasks (children’s ability to reason about a person’s belief about another person’s belief or emotion). Results showed no relation between sibling composition and ToM, although this null finding may have been due to near-ceiling performance and limited sibling variability (see also Calero, Salles, Semelman, & Sigman, 2013). Studies with older children from atypical populations provide contrasting results. For example, 4- to 12-year-olds with autism spectrum disorder (ASD) who have older siblings exhibit lower ToM compared with those with no older siblings, potentially caused by being routinely “overhelped” by older siblings in social situations (O’Brien, Slaughter, & Peterson, 2011). Clearly, further investigation is needed to build a stronger research base on individual differences in ToM during middle childhood, especially connections to sibling composition.

Although several studies have documented significant relations between siblings and ToM during early childhood, research remains mixed regarding which combination is most ideal—having more siblings (Jenkins and Astington, 1996, McAlister and Peterson, 2007, McAlister and Peterson, 2013, Perner et al., 1994) or having more siblings who are older in age (Farhadian et al., 2011, Lewis et al., 1996, Ruffman et al., 1998). Furthermore, a subset of studies have found that preschoolers with siblings who are diverse in gender or age demonstrate the most advanced ToM; however, researchers disagree whether it is most beneficial to have siblings who are both older and younger than oneself (Lewis et al., 1996, Peterson, 2000) or to have siblings who are of the opposite sex (Carlson and Moses, 2001, Cassidy et al., 2005, Ruffman et al., 1998). Exposure to a greater variety of sibling perspectives may benefit children’s social–cognitive development by providing more frequent opportunities to interact with children whose views differ from their own (Ross, Recchia, & Carpendale, 2005) such as exposure to differences in preferences and play styles that often follow gender lines (Martin & Ruble, 2010). Because these studies focused on children 6 years of age and younger, it is unknown whether the same sibling patterns hold in older age groups, particularly given that sibling relationships also undergo significant changes as children grow older (Cicirelli, 1995, Kim et al., 2006).

As described previously, a further issue common to prior investigations of sibling-ToM relations in preschool samples is that sibling chronological age has been confounded with relative age (older vs. younger status). That is, in most studies “older siblings” were by default at least 4 or 5 years of age—the age when most typically developing Western children demonstrate solid understanding about several mental state concepts, including false belief (Wellman et al., 2001). Perhaps it is not having a greater number of older siblings that helps ToM (apprenticeship model) but rather having more siblings who have reached a certain age threshold and, thus, have a certain level of social–cognitive skills (age threshold model). For example, having a 6-year-old sibling may benefit ToM development regardless of whether that 6-year-old is older or younger than the target child. By using a large sample of typically developing 4- to 11-year-olds, the current study can pit these models against each other. We can also test whether additional sibling composition variables, such as gender, predict variability in ToM during middle childhood.

More specifically, the current research tested relations between 4- to 11-year-olds’ ToM and (a) number of siblings, (b) sibling chronological ages, and (c) sibling gender. Within each of these categories, we also separated number of older versus younger siblings. We further examined diversity in sibling age (both older and younger siblings) or gender (both male and female siblings). We assessed advanced ToM knowledge using interpretive theory of mind (IToM) tasks. Chandler and Helm (1984) defined IToM as the capacity to recognize that the same situation can be open to multiple legitimate interpretations (see also Lalonde & Chandler, 2002). The standard task involves showing children a full drawing and then covering it up except for one small ambiguous part (e.g., a curve). Participants predict and explain the interpretations of people who did and who did not have prior access to this specific full picture. Children start to exhibit knowledge that different people can interpret the same situation in different ways, especially ambiguous situations, between 5 and 7 years of age (Carpendale, 1995, Carpendale and Chandler, 1996), with further development between 7 and 10 years of age and into adulthood (Lagattuta et al., 2010, Lagattuta et al., 2014, Ross et al., 2005). This contrasts with false belief understanding (knowledge that a person’s thoughts can be wrong), which children typically master around their fourth birthday (Wellman et al., 2001). Awareness that there can be many ways to interpret the same situation and that prior experiences shape these psychological reactions (IToM) continues to improve through middle childhood and beyond (see also Barquero et al., 2003, Lagattuta et al., 2015, Pillow, 2012, Pillow and Henrichon, 1996, Pillow and Mash, 1999, Ruffman et al., 1991).

In addition to testing children’s understanding of mental diversity in interpretation, we also assessed their executive function (EF). EF refers to a range of cognitive processes, including inhibitory control, working memory, attentional control, task switching, and planning skills (see Zelazo & Müller, 2011). We selected EF because it reliably predicts ToM in 2- to 6-year-olds even when controlling for age and language skills (e.g., Carlson et al., 2002, Hughes, 1998, Sabbagh et al., 2006), with recent work also demonstrating significant EF–ToM relations in older children and adults (Lagattuta et al., 2010, Lagattuta et al., 2014). Despite this strong research base on EF–ToM relations, only a few studies have controlled for EF when examining connections between sibling composition and ToM in 2- to 6-year-olds (Hughes and Ensor, 2005, McAlister and Peterson, 2006, McAlister and Peterson, 2013) or in older children with ASD (O’Brien et al., 2011). In the current work, we extended this research to test whether sibling composition predicts IToM during middle childhood independent of individual differences in EF. We included two types of EF—verbal working memory and inhibitory control—because performance on these measures significantly correlates with ToM in younger and older children (Carlson and Moses, 2001, Lagattuta et al., 2010, Lagattuta et al., 2014).

Finally, it is notable that researchers have primarily measured relations between EF and ToM and between siblings and ToM using children’s judgments or predictions about people’s mental states, not their explanations for why people have particular thoughts or beliefs (e.g., Carlson and Moses, 2001, Cassidy et al., 2005, Cole and Mitchell, 2000, Farhadian et al., 2011, Hughes and Ensor, 2005, Jenkins and Astington, 1996, Lewis et al., 1996, McAlister and Peterson, 2006, McAlister and Peterson, 2007, McAlister and Peterson, 2013, Peterson, 2000). When studies have asked children to explain their responses, explanations are typically combined with predictions to create a single score rather than treated as separate dependent variables (e.g., Brown et al., 1996, Miller, 2013; for an exception with preschoolers, see Hughes, 1998). Prior research indicates that explanations can be an especially powerful method for capturing children’s psychological reasoning because they provide a more in-depth assessment of thought processes (Bartsch and Wellman, 1989, Lagattuta and Wellman, 2002, Sayfan and Lagattuta, 2009, Wellman and Lagattuta, 2004). That is, explanations have “special potential for revealing, and aiding, children’s thinking, reasoning, and development” (Wellman, 2011b, p. 34). Thus, in the current study, we not only investigated age, EF, and sibling composition as predictors of 4- to 11-year-olds’ judgments about how different people will interpret ambiguous situations, but we also did so for their explanations for why people will have the same or different thoughts. In particular, we were interested in children’s ability to explain interindividual mental diversity as caused by differences in prior experience.

In summary, the current research addressed a significant gap in the literature on individual differences in ToM during middle childhood by assessing relations among sibling composition, EF, and an advanced measure of mental state understanding (IToM) that included both predictions and explanations as separate dependent variables. By studying 4- to 11-year-olds, we could determine whether sibling–ToM connections still exist during middle childhood independent of EF–ToM relations. The inclusion of school-age children also provided a stringent test of the apprenticeship model (older siblings matter for ToM because they are more developmentally advanced than the target child) by pitting it against an age threshold model (having siblings of a certain age matters for ToM regardless of whether they are older or younger than the target child). Our large sample size also allowed for a more detailed analysis of relations between different types of sibling composition and advanced ToM such as sibling gender and diversity. Based on previous literature, we hypothesized that (a) children with higher EF and with a greater number of siblings (especially older siblings or siblings diverse in gender or age) would demonstrate better understanding that different people can have different interpretations of the same situation, (b) this knowledge would be evident in both their predictions and explanations, and (c) both EF and sibling composition would independently contribute to IToM performance.

Data for this research were taken from Lagattuta and colleagues (2010), who documented significant relations between EF (inhibitory control and working memory) and IToM in 4- to 11-year-olds and adults. Unique to this current article are (a) analyses of relations among sibling composition, EF, and children’s judgments about people’s interpretations (sibling composition was not considered in that prior work) and (b) analyses of relations among sibling composition, EF, and children’s ability to explain people’s interpretations based on their past experiences (only age was examined as a predictor of explanations previously).