Parents' attributions of the jigsaw-puzzle performance of their child with Prader-Willi syndrome (n = 20) or Williams syndrome (n = 21) were examined. Parents in both groups placed more importance on internal versus external attributions. Parents of children with Prader-Willi syndrome exhibited a hedonic bias by attributing their child's success to internal factors (particularly when the child was good at puzzles), downplayed the role of chance when their child had higher puzzle skills and higher cognitive functioning, and differentiated between internal and external attributions. In contrast, parents of children with Williams syndrome showed an inconsistent attributional pattern. These findings help identify how parents understand the intellectual performance of their child with disabilities and how interventions might be beneficial.
Parents are understandably interested when their children with disabilities perform high-level intellectual tasks. Presumably for this reason, parents of children with disabilities (compared to those without) generally adopt more didactic, controlling, and intrusive styles of interaction with their children. As one of the mothers in Jones's (1980) study noted concerning her child with Down syndrome: “It's sit him on your knee and talk to him, that's the main object. Play with him, speak to the child, teach him something” (p. 221).
Although such parental interactive behaviors have long been studied, much less is known about how such parents conceptualize their child's intellectual performance. One important theory of parental perceptions is attribution theory, with its three causal dimensions of locus, controllability, and stability (Weiner, 1979). Within achievement contexts, Weiner (1985) hypothesized that parents attribute their child's performance as mostly due to two factors internal to the child (ability and effort) versus two external factors (luck and task) characteristics.
In addition to making such internal and external attributions, parents of typically developing children also tend to be positively biased, to see their child's behaviors in light of the so-called “hedonic bias.” Thus, parents generally attribute their child's negative behavior or poor performance to external, unstable, or situational factors, whereas positive child behaviors are thought to involve internal, stable, and dispositional factors (Cohen, 1995). Similarly, parents are more likely to attribute their child's development of skills to talent as opposed to effort and teaching (Cashmore & Goodnow, 1986).
Parents of children with disabilities, however, may be less likely to show the hedonic bias. For example, in two studies they attributed their child's success to luck (external, unstable, and uncontrollable) and failure to a lack of ability (internal, stable, and uncontrollable) (Lavelle, 1978; Lavelle & Keogh, 1980). Similarly, Cohen (1995) found that parents of children with developmental disabilities attributed their child's positive behavior/performance to external and unstable factors and their child's negative behavior/performance to internal and stable factors.
Although such reversals of the usual hedonic bias occur in parents of children with disabilities in general, children with particular genetic syndromes also perform better or worse on specific cognitive tasks (Dykens, 1995; Hodapp & Dykens, 1994). For example, individuals with Prader-Willi syndrome have relative strengths in visual–spatial skills, particularly jigsaw puzzles. Dykens (2002) found that on a short jigsaw puzzle task, children with Prader-Willi syndrome outperformed their MA-matched peers with mental retardation and CA-matched typically developing counterparts. In contrast, children and adults with Williams syndrome generally exhibit profound weaknesses in visual–spatial skills, such as drawing or putting together a jigsaw puzzle (Mervis & Klein-Tasman, 2000). Parents of children with Prader-Willi syndrome, as opposed to those with Williams syndrome, may, therefore, make internal attributions of ability or effort if their child succeeds on jigsaw puzzles.
Given the strength of individuals with Prader-Willi syndrome and the weakness of those with Williams syndrome in visual–spatial skills, we examined parents' attributional patterns as children and their parents completed a jigsaw puzzle. We had four goals. First, we attempted to understand the nature of parents' attributional patterns. Similar to parents of typically developing children, parents in both groups were expected to make more internal than external attributions for their child's performance. Second, we examined whether parents in each group would show a hedonic bias towards their child's performance by attributing their child's success to internal attributions and failure to external attributions. Third, we considered whether parents in each group would similarly differentiate between internal attributions (ability, effort) and external attributions (task characteristics, chance). Fourth, we looked at within-group relations of parents' attributions to children's characteristics (e.g., child's puzzle skills, IQ). By investigating parents' attributions on a task in which one group usually succeeds and the other usually fails, our goal was to better understand parents' attributions toward children with disabilities.
Participants were two groups of children ages 6 to 18 years and their parents: 21 children with Williams syndrome and 20 children with Prader-Willi syndrome. The mean age in both groups was 10 to 11 years, and all had genetic confirmation of their respective diagnoses. Within the group with Prader-Willi syndrome, 3 children were of the uniparental disomy subtype. Although children with this form of Prader-Willi syndrome generally do not show high levels of puzzle ability (Dykens, Cassidy, & King, 1999; Roof et al., 2000), these children nevertheless share the same syndromic diagnosis and, in many ways, show only subtle differences in behavior from the more common deletion form. We therefore included these 3 children in the Prader-Willi syndrome group.
Participants were recruited from a research program as well as the national parent conferences of both syndromes. Participants were primarily mothers, with the exception of 3 fathers in the Williams syndrome group and 4 in the Prader-Willi syndrome group. These fathers did not differ on relevant demographic variables, and their attributional ratings did not differ from those of mothers in their respective groups.
Participants were tested at the conference site or in their home. To establish an objective baseline of the child's puzzle skills, we gave children a puzzle to complete by themselves in 5 minutes. While their child was being tested in another room, parents completed demographic questionnaires about their child and the family (e.g., age, diagnosis, parent education). They then interacted with their child for 5 minutes using another jigsaw puzzle. All parent–child interactions were videotaped, and parents were told that the goal of the task was to have the child complete as many puzzle pieces as possible. After the interaction, parents assessed and made attributions about their child's performance.
Kaufman Brief Intelligence Test (K-BIT)
This test (Kaufman & Kaufman 1990) is a motor-free screening measure of intelligence that yields verbal, nonverbal, and composite IQ estimates. Standard scores were used to determine the children's cognitive functioning.
Parents' knowledge of their child's puzzle skills
Before the parent–child interaction, parents completed a questionnaire that assessed the degree to which parents were informed about their child's puzzle skills. Parents rated the degree to which their child is good at puzzles relative to the child's other abilities as well as relative to typical children of the same chronological age—CA (1 = much worse than, 7 = much better than). Parents' ratings in response to these two ratings were then averaged to be used as an indicator of parents' knowledge of their child's puzzle skills.
Two different jigsaw puzzles were used, one during child testing and the other during the interaction task with the parent. Both 30-piece puzzles depicted the cartoon character of Elliot Moose. The experimenter arranged the 30 pieces in random order, facing upward and kept the picture of the completed puzzle in front of participants at all times. Participants were told to put the pieces together as fast as they could within a 5-minute period. The total number of completed pieces in the presence of the experimenter (who gave no help) served as the measure of level of ability in jigsaw puzzles. The order of the jigsaw puzzles was counterbalanced between both tasks.
Parents' attributions of child's performance
After the parent–child interaction, parents rated their child's performance, as well as the importance of particular attributions on their child's performance, on a 7-point scale. Specifically, parents rated the degree to which they thought their child failed (1) or succeeded (7) on the puzzle task as well as their expectation for future success (1, very unlikely to 7, very likely). Parents also rated the degree to which four attributions (i.e., ability, effort, chance, and task characteristics) played an important role in their child's performance (1= not at all important to 7= very important). For example, parents were asked to rate “How important was ability to your child's performance?” They made ratings similar to those used for questions about effort, chance, and task characteristics.
As shown in Table 1, the two groups did not differ on most child, parent, or family variables. The educational level of mothers in the Williams syndrome group, however, was significantly higher than that of mothers in the Prader-Willi syndrome group. Because mothers' educational level did not correlate significantly with any of the outcome variables, it was not used as a covariate in subsequent analyses.
Etiological and puzzle-skill groupings
Children with Prader-Willi syndrome (M = 13.50, standard deviation [SD] = 11.20) completed over twice as many puzzle pieces as did children with Williams syndrome (Ms = 13.50, SD = 11.20, and 6.48, SD = 6.86, respectively), t(39) = 2.43, p < .05. A similar pattern was found for the number of puzzle pieces completed with the parent, t(39) = 2.28, p < .05 (Ms = 17.05, SD = 9.50, and 11.48, SD = 5.85, respectively).
The large variances in the child's puzzle skills indicated that some children in each group did not conform to their etiology-related strengths and weaknesses. Two groups of puzzle skills level were, therefore, created: low-puzzle skills and high-puzzle skills. Children who belonged in the high-skill group independently completed more than 10 puzzle pieces (i.e., above the combined group mean of 9.90), and consisted of 11 children with Prader-Willi syndrome and 7 with Williams syndrome. The low-skill group completed less than 10 pieces and included 9 children with Prader-Willi syndrome and 14 with Williams syndrome. The high-skill group did not differ from the low-skill group on any child, parent, or family variable except for the child's age. Children in the high-skill group were older (M = 12.25, SD = 3.99) than children in the low-skill group (Ms = 12.25, SD = 3.99, and 9.97, SD = 2.74, respectively), t(37) = −2.16, p < .05. Furthermore, within the Williams syndrome group, children in the low-skill group were younger (M = 9.18, SD = 2.15) than children in the high-skill group (Ms = 9.18, SD = 2.15, and 12.94, SD = 3.47), t(19) = −3.08, p < .01. Nevertheless, the interaction between etiological group and puzzle-skill group was not significantly different. Child age, however, was used as a covariate in subsequent analyses.
Parents' knowledge of child's puzzle skills
Parents were indeed knowledgeable about their child's jigsaw-puzzle skills, even before they interacted with their child. When asked to compare their child's puzzle skills to those of typically developing children, parents of children with Prader-Willi syndrome gave higher average ratings than did parents of children with Williams syndrome (Ms = 3.38, SD = 1.71, and 2.45, SD = 1.19, respectively), t(38) = 1.99, p < .05. In addition, parents' ratings moderately correlated with their child's actual puzzle skills, rs = .58 for Prader-Willi syndrome and .47 for Williams syndrome, ps < .05. Similarly, parents of children in the high-skill group (regardless of the child's etiology) rated their child's puzzle skills (M = 3.39, SD = 1.91) more similarly to that of typically developing children than did parents of children in the low-skill group (M = 1.68, SD = 1.04), t(38) = −3.59, p < .001.
Across and within the etiological groups (Prader-Willi syndrome vs. Williams syndrome) or puzzle-skill groups (high vs. low), gender of child did not affect parents' ratings of the child's performance and attributions.
Group Differences in Parents' Ratings
Like parents of typically developing children, parents of children with Prader-Willi or Williams syndromes placed more importance on internal than external attributions (Ms = 5.63, SD = 1.39, and 3.78, SD = 1.20, respectively), t(40) = 7.70, p < .001. Parents in the Prader-Willi syndrome and Williams syndrome groups did not significantly differ in their ratings of their child's performance or expectations of their child's future success. However, when looking at puzzle-skill levels, parents of children with high (vs. low) puzzle skills rated their child as being more successful (Ms = 5.61, SD = 1.54, and 3.00, SD = 1.68, respectively) and more likely to succeed on future puzzle tasks (Ms = 6.11 and 3.30, SDs = 1.18 and 1.74, respectively), Fs(1, 39) = 26.25 and 34.23, ps < .001.
To examine parents' attributional patterns, we conducted a 2 (etiology) × 2 (puzzle-skill level) MANCOVA, with child's age as the covariate and attributions of ability, effort, chance, and task characteristics as the dependent variables. No main effects occurred for etiology for any of the attributions. There was a main effect, however, of puzzle-skill level on chance attributions. Compared to parents of children with high-puzzle skills, parents of children with low-puzzle skills placed more importance on the role of chance (Ms = 2.11 and 3.30, SDs = 1.02 and 1.77, respectively), F(1, 36) = 6.13, p < .05.
There was an interaction between etiological group and high–low puzzle skills for attributions of ability, F(1, 36) = 4.12, p < .05, and a trend in the expected direction for attributions of effort, F(1, 36) = 3.79, p = .059. As shown in Figure 1, parents of children with Prader-Willi syndrome placed less importance on ability when their child had low-puzzle skills, but more importance when their child had high-puzzle skills. Parents of children with Williams syndrome showed the opposite pattern: Higher ability attributions when the child had low-puzzle abilities, lower when the child performed better at the jigsaw puzzle task.
To determine whether parents exhibit a hedonic bias, we examined the relations between parents' judgments of child success and their ratings of ability and effort separately by etiological group. Like parents of typically developing children, parents of children with Prader-Willi syndrome showed the hedonic bias. Parents who viewed their children as increasingly successful at the puzzle task increasingly attributed the child's success to both the child's ability, r = .56, p < .01, and effort, r = .60, p < .01. In contrast, within Williams syndrome, no such hedonic bias occurred between relations of parents' perceptions of their child's performance and ratings of either ability, r = −.10, or effort, r = .07, nor did parents of children with Williams syndrome show a negative bias by attributing their child's success to chance or task characteristics, rs = −.22.
Links of Internal vs. External Attributions
Parents of children with Prader-Willi and Williams syndrome also showed differences in their respective associations between internal and external attributions. Overall, parents of children with Williams syndrome did not differentiate internal from external attributions, providing correlated ratings of the two types of attributions, r = .51, p < .05. In contrast, the internal and external attributions of parents of children with Prader-Willi syndrome did not correlate, r = .21.
Examining individual internal and external causes, we found that the two internal attributions (ability and effort) were exceptionally highly related in parents of children with Prader-Willi syndrome but were not highly related to each other in parents of children with Williams syndrome (see Table 2). Instead, within the Williams syndrome group, the two internal attributions correlated with task characteristics. Furthermore, the correlations between ability and effort for the two groups (rs = .93 and .42, respectively) were significantly different, z = 3.46, p < .001. The two external attributions (chance and task characteristics) also showed a similar (albeit not significantly different) pattern. Specifically, within the Prader-Willi syndrome group, chance and task characteristics were significantly and moderately correlated, whereas within the Williams syndrome group, no correlations occurred (see Table 2). Parents of children with Prader-Willi syndrome, then, grouped together both individual internal attributions and individual external attributions and distinguished the two general types of attribution from one another. In contrast, parents of children with Williams syndrome did not differentiate between internal and external attributions for their child's performance.
Links to Relevant Child Characteristics
Overall, the child's actual puzzle skills showed similar relations to many aspects of parental attribution in both etiological groups. In both groups, parents of children with higher puzzle skills rated their child as being more successful and more likely to succeed on future puzzle tasks. None of the children's actual skills correlated with their parents' attributions of ability, effort, and task characteristics.
Attributions of chance, however, showed contrasting connections for the Prader-Willi syndrome and Williams syndrome groups. As Table 3 shows, parents of children with Prader-Willi placed lesser importance on chance because their children had increasingly higher puzzle skills and increasingly higher cognitive functioning. In Williams syndrome, no relations occurred between child characteristics and any parental attributions, including chance.
Going beyond parental attributions of children with disabilities in general, in this study we used etiology-related strengths and weaknesses to examine parental attributions of children thought to perform well (Prader-Willi syndrome) versus poorly (Williams syndrome) on a particular cognitive task. As such, this study goes beyond prior findings concerning parental attributions with either typical children or children with disabilities.
This study has three main findings. First, parents in our two groups were both similar and different from parents of typically developing children in their attributions of their children's performance. In both groups, parents relied more on internal attributions—the child's ability and effort—as opposed to such external attributions as chance or task characteristics. Parents were also able to judge their child's puzzle skills reasonably accurately, and parents of children who were higher versus lower in puzzle abilities rated their children as more successful and as more likely to succeed in future puzzle tasks.
Beyond this more general finding, however, the two groups showed intriguing differences in their attributional patterns. Although parents in both etiological groups placed more importance on internal than external attributions, patterns of parents' attributions were different based on their child's general level of puzzle skills. As Figure 1 shows, an interaction occurred between the child's level of puzzle ability and parental attributions of ability. The parents of children with Prader-Willi syndrome provided higher ability ratings as the child moved from less to more skill in puzzles. Parents of children with Williams syndrome showed the opposite pattern.
Our second main finding concerned the presence of hedonic bias in one—but not the other— of our two etiology groups. Parents of children with Prader-Willi syndrome attributed their child's success to both the child's ability and to the child's effort. As these children were seen as increasingly successful, parents in the Prader-Willi group more and more highly rated the internal attributions of ability and effort. In contrast, parents of children with Williams syndrome showed attributional patterns that reflected neither a hedonic nor a negative bias. Rather, they seemed inconsistent in their attributional patterns.
The third finding concerned the ways in which this pattern of consistency in Prader-Willi syndrome and inconsistency in Williams syndrome again arose when we examined the structure of parental attributions. Specifically, parents of children with Prader-Willi syndrome rated “in synch” the two internal attributions of ability and effort; the two external attributions of chance and task characteristics were also moderately correlated. In contrast, parents of children with Williams syndrome were inconsistent both within and across internal and external attributions. Parental ratings of ability and effort (the two internal attributions) were not significantly related, nor were ratings of the two external attributions (chance and task characteristics). Instead, parents in the Williams syndrome group linked task characteristics (an external attribution) to both ability and effort (the two internal attributions).
When examining parental attributions to their children's performance, then, it is important to differentiate by both etiology and by task. Parents of children with Prader-Willi syndrome seem similar to parents of typically developing children when it comes to tasks at which their child is competent. That is, they show a hedonic bias by increasingly endorsing the importance of ability and effort when their child succeeds and downplaying the role of chance on their child's performance when their child is less cognitively impaired and has higher puzzle skills. In contrast, parents of children with Williams syndrome do not attribute their child's success to internal factors or failure to external factors. At least for parents of children with Prader-Willi syndrome, both etiology and task seem important.
Why did parents of children in the two groups differ? One possibility might be that parents of children with Williams syndrome perceive their child to have lower ability and, therefore, hold lower expectations for the child to succeed. This, however, did not appear to be the case. Recall that parents of children with Williams syndrome, as a group, rated their children as equally successful and likely to succeed in the future as did parents of children with Prader-Willi syndrome.
A second possibility might be that parental attributional patterns relate to parental knowledge of etiology-related profiles. Again, however, this explanation seems unlikely, as parents of children with Prader-Willi syndrome or Williams syndrome generally know little about each syndrome's etiology-based cognitive profiles (Fidler, Hodapp, & Dykens, 2002). Instead, they were more attuned to the distinctive features associated with the syndrome, such as overeating and obsessive-compulsive behaviors in Prader-Willi syndrome and musical ability in Williams syndrome. However, because parents in our study were recruited from syndrome-specific conferences and research studies, and the proficiency with jigsaw puzzles in Prader-Willi syndrome (Dykens, 2002) and visual–spatial deficits in Williams syndrome (Farran & Jarrold, 2003) are well-documented, these parents may be quite knowledgeable about etiology-related profiles. Thus, the degree to which parents are knowledgeable of their child's profile may affect the type and accuracy of attributions.
Another possibility might be developmental factors in addition to etiological factors affecting puzzle ability in children with Williams syndrome. Recall that within the group of parents of children who had Williams syndrome, the children with low-puzzle skills were younger than those with high-puzzle skills. In addition, within the Williams syndrome group, the child's age correlated positively with puzzle ability, r = .51, p < .05. Nevertheless, the child's age did not correlate with parents' ratings of the child's performance or with parents' attributions. At least for the Williams syndrome group, we do not know the extent to which developmental factors also play a role.
Despite these possibilities, different parental attributions most likely are associated with the sheer amount of time that children in the two groups perform jigsaw puzzles every day. Rosner, Hodapp, Fidler, Sagun, and Dykens (2004) recently examined the everyday leisure-time activities of children with Prader-Willi or Williams syndromes. Playing with puzzles was among the most common of all activities for 50% of children with Prader-Willi syndrome but for only 2% of children with Williams syndrome. If children with Williams syndrome shy away from puzzles, then parents have few opportunities to interact with their child around such tasks. As a result, they may not know how to interpret the performances of those children who are indeed better at puzzles than would be expected by their etiology-related cognitive profile.
Parents of children with Prader-Willi syndrome had more differentiated and complex attributional patterns than did parents of children with Williams syndrome. Nevertheless, parents in the Prader-Willi group closely linked ability and effort, although distinguishing between the two has meaningful practical implications in explaining a child's success and failure (Graham & Weiner, 1991). Specifically, ability and effort are uniquely related to a distinct set of affective reactions and behaviors, which in turn can guide parental behaviors (Graham, 1984). For example, when a child's failure is perceived to be caused by low ability (something that is uncontrollable and stable), others are more likely to feel pity or sympathy and, thus, more likely to offer help (Graham, Doubleday, & Guarino, 1984). In contrast, when a child's failure is due to low effort (which is controllable and stable), others feel anger and are less likely to offer help. Although such links among parental attributions, emotions, and behaviors were not examined here, it may be important for parents of children with Prader-Willi syndrome to differentiate these two internal attributions.
Another practical implication involves attribution retraining. According to Graham (1991), making attributions of low effort in response to a child's failure is more adaptive because the child has control over how much effort to exert. In contrast, making attributions of high ability in response to a child's success may be more productive, because praising the child's ability for his or her success may result in greater intrinsic motivation (Koestner, Zuckerman, & Olsson, 1990) and higher expectations for success (Schunk, 1983; see Henderlong & Lepper, 2002, for a review). To what extent (and how) parental attributions are internalized by children with different syndromes remains almost totally unexamined.
It is also important to point out the limitations of this study. First, in addition to the relatively small sample size, not all children conformed to their etiology-based strength or weakness on jigsaw puzzles. Second, success versus failure may be subjective, and parents of children with Prader-Willi syndrome may have different standards of success than do parents of children with Williams syndrome. Finally, because some children with Williams syndrome rarely played with puzzles, parents may have limited experiences or opportunities from which to draw accurate attributions.
Still, as the first examination of parents' attributions that relates to children's etiology-based strengths and weaknesses, several important results were found. Unlike parents of children with Williams syndrome, parents of those with Prader-Willi syndrome who were good at puzzles highlighted the importance of ability and effort and downplayed the role of chance. Like parents of typically developing children, parents in the Prader-Willi syndrome group held the hedonic bias for their child's success and connected their attributions to their child's characteristics. In contrast, parents of children with Williams syndrome were much less consistent in their attributional patterns, showed neither hedonic nor negative biases, and held attributions that were generally unconnected to their children's perceived or actual abilities. As we begin the process of understanding parental attributions, we will clearly need to pay attention both to etiology and to task characteristics. Only by attending to both etiology and task characteristics will we be able to understand how, exactly, parents' perceptions and attributions relate to their children's performance of high-level intellectual tasks.
This research was based on the doctoral dissertation of the first author. The authors thank the families and persons with Prader-Willi and Williams syndromes who participated in this research. We are also grateful to Elisabeth Dykens and Sandra Graham for their comments on an earlier draft of this manuscript. The support of National Science Foundation postdoctoral fellowship No. 0310013 to the first author is also gratefully acknowledged. Portions of this paper were presented at the Gatlinburg Conference on Mental Retardation, 2004, San Diego.
Authors: Tran M. Ly, PhD (email@example.com), National Science Foundation Postdoctoral Fellow, Graduate School of Education & Information Studies, 2005D Moore Hall, Box 951521, University of California, Los Angeles, Los Angeles, CA 90095. Robert M. Hodapp, PhD, Professor, Vanderbilt University Kennedy Center for Research on Human Development, Box 328, 230 Appleton Place, Nashville, TN 37203