The expectation underlying the work on parenting styles has been that the authoritative style will be beneficial for theory-of-mind development and that the authoritarian style will be detrimental. This pattern is, of course, what has been found across a range of outcomes in the general parenting literature. Most of this research, it should be noted, has examined children’s compliance with parental directives and thus falls under the control domain of socialization (Table 3.5). It is not immediately clear that the same parental practices that result in compliance will also nurture understanding of mental states. The domain-specificity of parental practices is, as we have seen, a general theme in writings about domains of socialization (e.g., Grusec & Davidov, 2010). As noted in Chapter 3, however, the subset of parenting style studies with cognitive outcomes have generally reported authoritative-authoritarian differences that parallel those in the literature as a whole (Gauvain et al., 2013).
Why might authoritative parenting nurture theory of mind whereas authoritarian parenting fails do so? The most obvious explanation concerns the emphasis on reasoning and discussion that characterizes the former but not the latter approach. Ruffman et al. (1999) summarize the argument as follows: “Authoritarian parenting is characterized by strict punishment and poor communication, but discussion and communication must be the means by which mothers teach children about others’ mental states if mothers are to play any role in this regard” (p. 396). Vinden (2001) offers a similar argument while adding a point about the importance of the control dimension: “Authoritative parenting, by both providing structure yet also encouraging the child’s autonomy, simultaneously offers the child the parent’s perspective while acknowledging the child’s perspective. In this way, the authoritative parent is constantly providing the child with opportunity to reflect on dual, and possibly conflicting, perspectives on the world” (pp. 796-797). I will add that the warmth and responsiveness that characterize the authoritative approach may also be important, for these qualities may increase the probability that the child will attend to and be responsive to parental messages (Darling & Steinberg, 1993).
Research provides some support for the expected benefits of the authoritative approach (or at least aspects thereof) in comparison to an authoritarian style; this conclusion comes, however, with a number of qualifications and exceptions. Among the generally supportive sources are two of the reports cited in the preceding paragraph: Ruffman et al. (1999) and Vinden (2001). The main finding of the Ruffman et al. study (whose measurement of parental practices was summarized in Box 3.1) was that maternal disciplinary responses that stressed the effects of the child’s actions on the feelings of others (the How Feel category of discipline) related positively to children’s understanding of false belief. The efficacy of such responses fits with what we saw in Chapter 3 are general conclusions about the value of other-oriented reasoning. In contrast, the General Discussion category showed no relation to false belief success, and a relatively high use of Reprimand related negatively to false belief understanding (although not significantly so when other factors were controlled). In general, then, discussion and involvement (two characteristics of the authoritative approach) appeared beneficial, but only particular forms of discussion.
One further finding from the study is worth noting. Ruffman and colleagues reported that How Feel responses were directed most frequently to the youngest children within the study’s 3- to 5-year-old age span. Because the youngest children were least likely to have mastered false belief, this finding suggests that the How Feel responses were not elicited by the children’s relatively advanced theory of mind; rather the causal direction for the correlation was from parent to child.
Whereas the Ruffman et al. (1999) study reported positive effects of an aspect of authoritative parenting, the Vinden (2001) study reported negative effects of an aspect of the authoritarian style. Vinden’s measurement of parenting came from an instrument constructed for the study: the Parenting Attitudes Inventory, or PAI. The PAI assesses attitudes with respect to three dimensions relevant to the authoritative-authoritarian distinction: encouragement of autonomy (e.g., “I like to see a child have opinions and express them, even to an adult”), behavioral control (e.g., “Children should do as they are told without questioning their parents”), and freedom in learning (e.g., “It’s OK if my child tries to do things on his own”). For the Anglo American families in the study there was a negative relation between mothers’ endorsement of behavioral control and children’s theory-of- mind understanding. As expected, therefore, an authoritarian approach to parenting appeared to have a negative impact on theory-of-mind development. Results for the Korean American families, however, differed in two ways from those for the Anglo families. First, the Korean American mothers ranked higher in authoritarianism than did the Anglo mothers. And second, there was no relation between authoritarian parenting and theory of mind for the Korean American families.
Both findings, it should be clear, mirror conclusions from the general parenting literature: Asian parents often score higher on measures of authoritarianism than do Western parents, and authoritarianism is less likely to have detrimental effects in Asian samples than in Western samples.
Follow-ups of the Ruffman et al. (1999) and Vinden (2001) studies provide both some support for their general conclusions and some amendments. Using a modified version of Vinden’s PAI, O’Reilly and Peterson (2014) reported effects of parenting in both directions: a positive relation between authoritative parenting and theory of mind, and a negative relation between authoritarian parenting and theory of mind. Using a modified version of the Ruffman et al. questionnaire and coding system, Shahaeian, Nielsen, Peterson, & Slaughter (2014) reported that the Discuss category related positively to theory of mind and that a category labeled Silence related negatively. They also found that the How Feel category, although not often used by their Iranian sample, related positively to children’s understanding of false belief. Finally, Lewis and colleagues (Lewis, Huang, & Rooksby, 2006), working with a Chinese sample, reported no relation between response to a slightly modified version of the Ruffman et al. (1999) questionnaire and children’s false belief performance (performance that lagged behind that typical in Western samples). This, then, may be another case in which a pattern found in Western samples does not generalize to Asian cultures. On the other hand, the variability in both the parents’ responses and the children’s false belief performance was low, and restriction of range is therefore another possible explanation for the negative results.
The restriction-of-range argument may apply more generally to the studies reviewed in this section. Ruffman et al. (2006) offer restriction of range as a possible explanation for their failure to find any relation between their measure of parenting practices (with a special emphasis on warmth) and theory-of-mind development. As they note, their sample consisted primarily of middle- and upper-income families, and parental practices were skewed toward the optimal end of the dimensions measured. A somewhat different restriction-of-range problem may account for the absence of significant results in Holmes-Lonergan (2003), another study to report no relations between parental practices and child performance. In Holmes-Lonergan’s sample of Head Start children and mothers, approximately 75% of the maternal disciplinary responses (as reported by the children) fell into the Power Assertion category.
That a relatively high use of power assertion can often be detrimental is clear from other studies. We saw in Chapter 3 that this is a general conclusion from the parenting literature, a conclusion that comes primarily from studies of behavioral compliance but that also extends to work with cognitive outcomes (Pears & Moses, 2003). In research with theory of mind as the outcome a number of studies have reported negative effects of a relatively high use of power assertive or harsh forms of discipline (Hughes et al., 1999; Hughes & Ensor, 2006; Olson,
Lopez-Duran, Lunkenheimer, Chang, & Sameroff, 2011; Pears & Moses, 2003; Rohrer et al., 2011). This finding holds across several different outcome measures and several different ways of measuring parental behavior. Conversely, relatively high levels of parental warmth or parental empathy have emerged as a facilita- tive factor in a number of reports (Cahill, Deater-Deckard, Pike, & Hughes, 2007; Farrant, Devine, Maybery, & Fletcher, 2012; Olson et al., 2011). The Farrant et al. (2012) study also found that mothers’ encouragement of perspective taking related positively to children’s theory-of-mind understanding, a finding compatible with the value of How Feel responses reported by Ruffman et al. (1999).
On the other hand, and as is often true in parenting research, these conclusions are accompanied by various exceptions and complications. In the Hughes et al. (1999) study, the effects of parental practices varied with the sex of the child. For girls, relatively high parental warmth was associated with good performance on a battery of theory-of-mind tasks. For boys, there was no effect of parental warmth; rather, severity of discipline (high use of physical control and criticism) related positively to theory-of-mind performance. The finding for girls is, of course, expectable, but that for boys is not. The authors make the intriguing, although admittedly speculative, suggestion that the sex difference may be in part child-driven: that girls may use their budding theory-of-mind skills to build affectionate relationships with their parents, whereas boys use their skills in a testing- the-limits fashion that often results in parental discipline. This suggestion clearly requires further testing.
The Hughes et al. (1999) study is not the only report of an unexpectedly beneficial effect of power-assertive discipline. Pears and Moses (2003) examined relations between power assertion (as determined from mothers’ reports of their typical methods of discipline) and various measures of their children’s theory- of-mind understanding. When false belief was the outcome, power assertion related negatively to children’s understanding, a conclusion that emerged from both correlational and regression analyses. When understanding of emotions was the outcome, however, the direction of the relation reversed, and power assertion related positively to children’s understanding. Why might this be? Pears and Moses suggest that a clear expression of emotions, especially anger, is likely to be a frequent accompaniment of power-assertive discipline, and that children may learn from their parents’ emotional displays. They also suggest that the parent’s feelings about the disciplinary situation in question are likely to differ from the child’s feelings, and that such differences in perspective may also help children to learn about emotions. They also add, however, that the negative effects of power assertion are likely to outweigh any limited and possibly transitory benefits of the approach.
In some studies the interest has been not simply in the effects of parenting on theory of mind but also in the interplay of parenting and theory of mind with respect to some other developmental outcome. In a study by Hughes and
Ensor (2006) the outcome was behavioral problems in 2-year-olds, as measured from both maternal reports and observations of mother-child interaction. Both harsh parenting and deficits in theory of mind proved to be significant predictors of behavioral problems. Although both findings are familiar ones from the general literature, the demonstration of effects as early as age 2 is a novel contribution of the study. A further contribution is the demonstration of an interaction between the two predictors: Harsh parenting was most predictive of behavioral problems when the child’s theory-of-mind skills were low. The authors suggest that theory-of-mind skills may serve as a buffer against harsh parenting—that the negative effects of such parenting are reduced “when children have sufficient ToM [theory of mind] skills to recognise and anticipate others’ feelings and intentions” (Hughes & Ensor, 2006, p. 493). As their study also shows, however, harsh parenting makes it less likely that children will develop the needed buffering skills.
The further outcome of interest in a study by Cahill and colleagues (Cahill et al., 2007) was children’s feelings of self-worth. The child participants were 3-year-old twins (the same sample that had served as the basis for the Hughes et al., 1999, report), and the measure of self-worth was the Eder (1990) puppet task, a measure designed to assess aspects of self-concept in preschool children. Maternal warmth, as scored from observations in the home, showed a modest positive relation to the children’s theory-of-mind understanding. When considered separately, both theory of mind and maternal warmth related positively, although again modestly, to the self-worth scores. When the two predictors were considered together, an interaction emerged: When warmth was high, theory of mind related positively to self-worth; when warmth was low, the relation between theory of mind and self-worth was negative (although significantly so in only some analyses). The general message is the same as that in the Hughes and Ensor (2006) study: Parenting and theory of mind may be important not only in themselves but also as each moderates the effects of the other.
As the discussion to this point suggests, research on styles (or on style-related dimensions) has been skewed toward the authoritative and authoritarian styles and the contrasts (differences in warmth and reasoning) that define these styles. A report by Guajardo and colleagues (Guajardo et al., 2009) provides a rare focus on the involved-uninvolved dimension. These authors used maternal selfreports to classify mothers as either overreactive or lax in their response to disciplinary situations. An example of a response choice indicative of the former is “When I am upset or under stress I am picky on my child’s back”; an example of a choice indicative of the latter is “I threaten to do things that I know I won’t actually do.” The assessment for the child component of the study was unusually broad, including both the Wellman and Liu (2004) seven-task battery and a separate measure of emotion understanding. Overreactive parenting did not relate to either the theory-of-mind battery or the test of emotion understanding;
lax parenting, however, related negatively to theory-of-mind performance. This finding is an expectable one—that lax parenting is not conducive to theory-of- mind development can be predicted from both the general parenting literature and the research reviewed earlier in this section. Because there was no measure of maternal warmth, however, the study does not tell us the extent to which the lax parenting was part of an overall permissive or uninvolved parenting style.
Although in a different way, a study by Humfress and colleagues (Humfress, O’Connor, Slaughter, Target, & Fonagy, 2002) is also a bit of an outlier among the studies reviewed here. As Table 4.1 indicates, this study is the only entry under the styles/dimensions heading to focus on children beyond the preschool period. As a consequence of this focus, it is also the only study to use a measure of advanced theory of mind (Happe’s, 1994, Strange Stories), and it is one of only two studies (Holmes-Lonergan, 2003, being the other) to use child reports as the measure of parenting.
One emphasis of the study was on the relation between attachment and theory of mind, and I defer this component of the research until Chapter 5. As Table 4.1 shows, the study also included a number of measures of parenting quality, with a focus on dimensions (warmth, support, control) central to the Baumrind typology. This aspect of the study can be easily summarized: There was no relation between the parenting measures and children’s theory-of-mind performance.
Earlier I noted several other instances of a failure to find expected relations between parenting style and theory of mind. I will add two further instances here. Murray et al. (1999) measured two aspects of parenting: sensitivity of maternal communication and parental conflict. Neither measure showed any relation to children’s false belief understanding. Cole and Mitchell (1998) administered the Child Rearing Practices Report (Block, 1981), an instrument that includes both items reflective of an authoritative parenting style (e.g., inductive reasoning, expression of positive affect) and items reflective of an authoritarian style (e.g., physical punishment, repression of affect). The parenting scores showed no relation to children’s theory-of-mind performance.
On the other hand, one of the parental measures in the Cole and Mitchell study (1998) did prove predictive of children’s performance, and that was parental stress. The direction of the relation varied, however, across the two child outcome measures included in the study. One measure consisted of standard false belief and appearance/reality tasks, and in this case parental stress related negatively to children’s performance. The other measure was a test of expressive deception: children’s ability, as determined from both verbal and facial cues, to tell a convincing lie. In this case parental stress was a positive predictor: the greater the stress, the better the performance. Stress associated with single parenthood appeared especially important. We have, then, another instance in which a generally negative aspect of parenting is associated with relatively advanced theory-of-mind development. Presumably (although the study does not attempt to determine the basis), being reared by a stressed single parent provides opportunities to experience and practice deception beyond those available in a typical home situation.
The Cole and Mitchell (1998) study is not the only demonstration that parental stress can affect theory-of-mind development, although the direction of effect is not consistent across studies. Guajardo et al. (2009) included a broad-based measure of stress: the Parenting Stress Index (Abidin, 1995). Relatively high scores on the index related negatively to performance on the Wellman and Liu theory- of-mind battery. Symons and Clark (2000) also administered the Parent Stress Index; in their analyses, however, responses to the index were combined with responses to several other measures (e.g., the State-Trait Anxiety Inventory, the Centre for Epidemiological Studies Depression Scale) to form a composite measure of maternal distress. Distress assessed when the child was 2 years old related positively to false belief performance when the child was 5 years old. Interestingly, however, the relation held only for a specially developed form of the unexpected location task in which a mother was the object that changed position. The authors suggest that distressed mothers might talk more than most mothers about both their own and their child’s mental states, thus providing clues that aid the child’s understanding, especially when a mother was the target of the child’s theory-of- mind efforts. As with other post hoc explanations for unexpected results that we have seen, this suggestion clearly requires further study.
The variations in stress or distress examined in the studies just discussed fall, at least for the most part, into the typical, nonclinical range of individual differences. We will see in the concluding section of this chapter that more severe forms of emotional distress in a parent can have a definite effect on the development of theory of mind.