Journal of Autism and Developmental Disorders

, Volume 38, Issue 9, pp 1740–1750

Child Temperaments, Differential Parenting, and the Sibling Relationships of Children with Autism Spectrum Disorder

Authors

  • Jessica Wood Rivers
    • Institute on Human Development and Disability/UCEDDCollege of Family and Consumer Sciences, The University of Georgia
Original Paper

DOI: 10.1007/s10803-008-0560-z

Cite this article as:
Rivers, J.W. & Stoneman, Z. J Autism Dev Disord (2008) 38: 1740. doi:10.1007/s10803-008-0560-z

Abstract

This study examined associations between sibling temperaments, differential parenting, and the quality of the relationships between 50 children with autism spectrum disorder (ASD) and their typically developing siblings. The temperament dimension of persistence, but not activity level or emotional intensity, was found to relate to the quality of the sibling relationship. Effects were stronger for temperaments of the typically developing siblings, but persistence levels of both groups of children interacted to predict sibling relationship quality. Persistence also was the temperament dimension associated with differential parenting, with increased levels of differential parenting occurring when siblings, and to some degree the children with ASD, were low in persistence. When siblings were dissatisfied with differential parenting, quality of the sibling relationship was compromised.

Keywords

SiblingsTemperamentDifferential parentingFamiliesAutism

Numerous recent research studies have focused on siblings of children with autism spectrum disorders (ASD; i.e. Harris and Glasberg 2003; Hastings 2003a, b, 2007; Kaminsky and Dewey 2001, 2002; Macks and Reeve 2007; Pilowsky et al. 2004; Rivers and Stoneman 2003; Ross and Cuskelly 2006). The growing research interest in siblings reflects an increased understanding of the importance of the sibling relationship to families and to individual children, as well as an interest in understanding variability in the quality of sibling relationships across different pairs of children, including those affected by ASD (Stoneman 1993). Researchers are only beginning to understand the individual difference characteristics and family processes that support or compromise the development of positive sibling relationships. One factor identified in the developmental literature as contributing to the quality of sibling relationships is child temperament.

Temperament is a relatively stable individual difference characteristic, rooted in the child’s biology and influenced, as the child develops, by the environment and maturity (Bates 1989; Rothbart et al. 2000). Although researchers use different terms and measures, general agreement is emerging concerning the components of childhood temperament (Martin et al. 1994). A popular formulation is that of Rothbart (i.e. Rothbart et al. 2000) in which reactivity and self-regulation processes form the basis for temperament, mapped during childhood by three temperament dimensions: effortful control, extraverion/surgency, and negative affectivity. Closely aligned with Rothbart, the work of Martin and colleagues (i.e. Martin and Bridger 1999; Deal et al. 2005) employs a two-factor model: impulsivity (conceptualized as roughly parallel to Rothbart’s effortful control, Deal et al. 2005), and behavioral inhibition (parallel to Rothbart’s extraversion/surgency, Deal et al. 2005). Impulsivity is comprised of three components: negative emotionality (strong negative emotions in the face of frustration), persistence (sustained attention to tasks, even when challenging), and activity (intense, frequent physical movement). Behavioral inhibition (tendency to become upset or to withdraw in novel social situations) is a single factor. The Temperament Assessment Battery—Revised (Martin and Bridger 1999) reflects this model.

In one of the first studies of the temperaments of children with autism, Kasari and Sigman (1997) found that young children with autism had more difficult temperaments than comparison children. Bailey et al. (2000), using the Behavioral Style Questionnaire (BSQ; McDevitt and Carey 1978), identified temperament differences between 3- and 7-year-old boys with autism and the BSQ reference sample; boys with autism were less persistent, less emotionally intense, less distractible, slower to adapt to new situations, more withdrawing, and had a lower responsiveness threshold. Similarly, Hepburn and Stone (2006) compared 2–7 year old children with ASD to the BSQ reference sample and found children with ASD to be less persistent and less adaptable. Using the Child Behavior Questionnaire (Rothbart et al. 2001), which taps the three Rothbart temperament dimensions, Konstantareas and Stewart (2006) found that 3–10-year old children with ASD were rated as lower in effortful control than comparison children, but there were no differences between groups for negative affectivity or surgency/extroversion. In addition, their study revealed no significant differences between children with diagnoses of autism or PDD-NOS on any of the temperament dimensions. These studies suggest that temperament can be reliably and validly measured in children with ASD and that temperament dimensions differ for children with ASD as compared to either a normative sample or a comparison group of typically developing children.

The impulsive temperament dimensions proposed by Martin (Martin and Bridger 1999; Deal et al. 2005), namely high emotional intensity, high activity, and low persistence, have repeatedly have been linked in the developmental literature with conflicted, negative sibling relationships (i.e. Brody et al. 1987a, b, 1996, 1994a, b; Lemery and Goldsmith 2001; Stocker et al. 1989; Stoneman and Brody 1993). Only a few sibling studies have followed a family systems perspective (i.e. Minuchin 1974) in which characteristics and behavior of both siblings are simultaneously studied. Munn and Dunn (1989), following this approach, found that siblings with more similar temperaments had interactions characterized by lower conflict. Stoneman and Brody (1993) found that the least conflicted dyads were those in which both siblings had low active temperaments; the most conflicted had two high active siblings. Sibling positivity was highest when siblings were similar in activity. When siblings had dissimilar temperaments, limited support was found for a buffering effect in which one child’s easy temperament protected the relationship from the deleterious effects of the other sibling’s more difficult temperament.

In addition to directly affecting the sibling relationship, children’s temperaments have been found to have an indirect effect on sibling relations through differential parenting. Parents do not treat siblings in the same family identically. Within-family differences in parenting experienced by siblings are referred to as differential parenting. In the general sibling literature, differential parenting repeatedly has been associated with negative sibling outcomes, contemporaneously and longitudinally, including less prosocial behavior and engagement (Brody et al. 1987a), greater competition (Stocker et al. 1989), and increased conflict (Brody et al. 1992; McHale et al. 1995; Stocker et al. 1989; Volling and Belsky 1992).

Beginning with the work of Farber (Farber and Jenne 1963), numerous researchers have documented increased differential parenting in families of children with disabilities, generally favoring the children with disabilities (i.e. Corter et al. 1992; Lobato et al. 1991; McHale and Pawletko 1992; Quittner and Opipari 1994; Brody et al. 1987a). Evidence suggests that it is not the absolute amount of differential parenting that is important for siblings, but rather children’s satisfaction with differential parenting. Children do not always object to being treated differently from their siblings, as long as they can find meaning in the difference and perceive the difference as being fair (i.e. Kowal et al. 2002; McHale et al. 2000).

In the typical sibling literature, there are demonstrated links between child temperament and differential parenting. Differential parenting has been found to be more pronounced when children have more difficult temperaments (Brody et al. 1992; Feinberg et al. 2000; Jenkins et al. 2003). To better understand the circumstances in which positive relationships develop between children with ASD and their siblings, it is important to study the associations among temperament, differential parenting, and sibling relationships.

The first aim of the current study was to examine associations between children’s temperaments (individually and interactively) and the quality of the relationship between children with ASD and their siblings. Based on the general sibling literature, it was expected that when typically developing siblings had temperaments characterized by lower activity, lower emotional intensity, and greater persistence, sibling relationships would be more positive and less negative than those involving more temperamentally difficult children. Since behavioral inhibition focuses on reactions toward unfamiliar social agents and since the research literature does not provide evidence of an association with sibling variables, it was not expected that this temperament dimension would be associated with the quality of the sibling relationship. No hypotheses were rendered for the associations between the temperaments of children with ASD and sibling relationship variables because of the lack of past research in this area. Similarly, no hypotheses were rendered for the interactive effects of the two siblings’ temperaments when simultaneously considered because of the lack of consistent findings for typically developing siblings.

The second aim of the study was to examine associations: (1) between child temperament and sibling satisfaction with differential parenting and (2) between sibling satisfaction with differential parenting and the quality of the sibling relationship. It was anticipated that more active, emotionally intense, less persistent temperaments in typically developing siblings would be associated with increased sibling dissatisfaction with differential parenting. No predictions were made for associations with differential parenting involving the temperaments of the children with ASD or for the interactive effects of the temperaments of the two children. Finally, it was predicted that when siblings were more satisfied with differential parenting, their relationships with their siblings with ASD would be more positive and less negative.

Method

Participants

Fifty families participated in the study. Families had a child with ASD between the ages of 4 and 12, as well as a typically developing sibling between the ages of 7 and 12. In addition to 48 families with a child reported by parents to have autism, one family of a child with Asperger’s disorder and one family of a child with PDD were also included in the sample. Due to the low incidence of the targeted population, this study employed non-random, purposive sampling. Families were recruited through five university-based centers focusing on children with ASD. Additional families were recruited through Autism Society parent groups and newsletters.

Most of the children with ASD were reported by their parents as having either mild (n = 17) or moderate (n = 25) impairments; eight children were reported to have severe impairments. Fourteen of the children lacked toileting skills; 13 were reported to be nonverbal. The children with ASD were an average of 7.6 years of age; their typically developing siblings averaged 9.7 years. If families had more than one typically developing sibling in the targeted age range, the parent reported on the sibling who was the closest in age to the child with ASD, selecting a child who would be able to read and complete the questions. This was usually the next oldest sibling.

Forty of the typically developing children were older siblings. The ages of the older and younger siblings differed by 11–90 months (mean spacing = 26 months). As would be expected, the children with ASD were primarily boys (42 males, 8 females). Fifty percent of the typically developing siblings were male. Of the 23 older sisters, 19 had younger brothers with ASD and 4 had younger sisters. Of the 20 older brothers, 16 had younger brothers with ASD and 4 had younger sisters. Of the younger siblings, 5 were boys and 2 were girls; all 7 had older brothers with ASD.

One parent per family (49 mothers, 1 father) participated in the study. The average age of participating parents was 37 years (range = 30–49 years). Forty-six families were Euro-American, three were African-American, and one was Hispanic-American. Most parents were currently married (n = 45); three were divorced, and two were separated. An average of 2.9 children lived in the home (range = 2–6 children). Four parents were high school graduates, 18 had some college or technical school, 19 had a college degree, and 9 had graduate degrees. Most of the participating families were middle to upper-middle class (incomes ranged from under $10,000 per year to over $80,000; the average family income was between $50,000 and $59,999).

Families mailed completed study packets directly back to the researchers. Due to the wide distribution of research packets, it is not possible to estimate how many packets were actually given to families by the referral sources and what percentage of families who received packets fit the study criteria and returned the packets. A $15 honorarium was paid to each family upon receipt of a completed packet (in a protocol approved by the university IRB, identifying information was kept separate from the research packets and was destroyed as soon as honorarium payments were processed so that all responses were anonymous).

Procedure

Each research packet contained two sets of questionnaires (one for the parent and one for a sibling), instructions for completion, and a postage-paid return envelope. Instructions stressed that the questionnaires were to be completed independently by each participant.

Measurement

Sibling Relationship

The quality of the sibling relationship, as perceived by the parent and the typically developing sibling, was assessed by two instruments that have previously been used in studies of siblings with and without disabilities. The 28-item Sibling Inventory of Behavior (SIB, Schaeffer and Edgerton 1981; modified by McHale and Gamble 1987) asks respondents to rate the frequency of a set of positive and negative behaviors of the typically developing child toward the child with ASD during the past two months, using a 5-point Likert scale. The SIB assesses two positive (involvement/leadership, empathy/concern) and two negative (unkindness, avoidance/embarrassment) relationship dimensions. For the regression analyses, the two positive indicators for each rater were summed to create a Positive Relationship Quality variable, as were the two Negative Relationship Quality items. This strategy was confirmed by strong correlations between each pair of indicators for both parents and siblings (r’s ranged from .74 to .84, p’s < .0001). In the current sample, alphas for Positive Relationship Quality were .91 for sibling respondents and .93 for parent respondents. Alphas for Negative Relationship Quality were .88 for siblings and .93 for parents. Positive and Negative Relationship Quality were negatively correlated: r sibling respondents = −.72, r parent respondents = −.80, p’s < .001.

The second sibling measure, the Satisfaction with the Sibling Relationship Scale (McHale and Gamble 1989), has four items focusing on the typically developing sibling’s overall happiness with the sibling relationship, satisfaction with time spent with sibling, caretaking, and how the siblings get along. Two modifications were made: (1) a 5-point response scale was used instead of the original 9-point scale, and (2) respondents reported on behavior in the last 3 months, instead of 2 months, so that both sibling measures would refer to the same time-frame. In the current study, coefficient α was .82 for the modified 4-item scale for parents and .74 for siblings.

Temperament

Two temperament measures were used in the study, depending on the age of the child. Parents rated the temperaments of both their typically developing children and their children with autism. The order of the instruments in the research packets was counter-balanced, such that approximately half the parents rated the typically developing child first. The Temperament Assessment Battery for Children—Revised (TABC-R, Martin and Bridger 1999) was used with children aged 4–7 years. The School-Age Temperament Inventory (SATI, McClowry 1995) was used for children ages 8–12. McClowry (1995) reported strong convergent validity between parallel scales of the TAB-R and the SATI.

The TAB-R (Martin and Bridger 1999) is based on the work of Thomas et al. (Thomas and Chess 1977) and draws upon an earlier temperament instrument (The Behavior Style Questionnaire, McDevitt and Carey 1978) for approximately 85% of its items. The TAB-R consists of 37 behaviors, which are rated on a 7-point Likert scale and create four scales: Negative Emotionality, Activity Level, Persistence, and Behavioral Inhibition. Reliability on these scales, as reported by alpha coefficients, ranges from .70 to .90 (Ball et al. 2001; Martin and Bridger 1999). In the current study, alpha coefficients for the four factors for the children with autism were .82, .73, .75, and .81, respectively. Because almost all typically developing siblings were age eight and older, psychometric properties of the TAB-R for this group were not calculated.

The 38-item SATI, conceptually similar to the TAB-R (McClowry 1995), uses a 6-point response scale. Three SATI temperament dimensions, Negative Emotionality, Activity Level, Persistence, and Behavior Inhibition, are comparable to the parallel scales of the TABC-R. In the current study, Cronbach alphas for the SATI dimensions for typically developing siblings were .92, .80, .92, and .81, respectively. For children with autism, alphas were .81, .87, .82, and .42, respectively. As recommended by Martin (R. P. Martin, who developed and validated the TAB-R, personal communication, June 23, 1996), scores from the subscales of the TABC-R and the SATI were separately converted to standardized z scores for each of the temperament dimensions and merged across subjects, forming three continuous temperament variables: Negative Emotionality (also referred to as Emotional Intensity), Activity, and Persistence. Because of the low alpha for children with ASD on the SATI, behavioral inhibition was not analyzed further.

Satisfaction with Differential Parenting

Two items from the Satisfaction with the Sibling Relationship Scale (McHale and Gamble 1989) were used to assess sibling satisfaction with parental differential treatment. The first item asked about the relative attention parents paid to the child with ASD and his or her sibling; the second item asked about general satisfaction with the overall level of differential parenting. For both questions, siblings rated their satisfaction with their parent’s differential treatment in the past three months on a 5-point scale ranging from Very Happy to Very Unhappy. Parents rated how they believed that the typically developing siblings felt about the parents’ treatment of the two children, using two parallel items and a similar rating scale. The two items were correlated for siblings, r = .76, p < .001, and for parents, r = .84, p < .001; thus, the two items were summed for each respondent to yield a single measure.

Results

Descriptive Analyses

As study data are presented, the terms “sibling(s)” and “child/children” will be used to refer to the typically developing siblings. The siblings with ASD are referred to by mentioning their disability. All sibling measures evidenced positive associations between parents’ and children’s ratings; Satisfaction: r = .65; Positive Relationship Quality: r = .59; Involvement: r = .53; Empathy/concern: r = .63; Negative Relationship Quality: r = .80; Avoidance/embarrassment: r = .75; Unkindness: r = .74, p’s < .001. Parents’ ratings of the temperaments of the children with ASD and their siblings were not significantly correlated, r’s = .05 to −.13. Neither parent nor child reports of the quality of the sibling relationship differed by sibling gender, as examined by t tests. Gender of the children with ASD was not analyzed for these variables since almost 85% of the sample was male, compromising the group sample size.

No significant associations were found between the sibling relationship indicators and family income, parent education, age of the sibling, the age of the child with ASD, or sibling spacing (the difference in ages between the two children). There were not sufficient numbers of younger siblings to analyze for birth order, or for birth order by gender interactions. Parents were asked if the child with ASD had verbal speech, used speech as their primary form of communication, and whether or not they had toileting skills. The sibling variables did not differ depending on the answers to these questions.

Individual Children’s Temperaments and Sibling Relationship

Table 1 presents the correlations between parents’ temperament ratings and the sibling relationship measures. When the siblings had temperaments higher in persistence, both the siblings and their parents rated the relationship as being less negative, with less unkindness and less avoidance/embarrassment. In addition, parents rated sibling satisfaction and empathy/concern as being higher when the siblings were higher in persistence. No significant associations were found between temperaments of the children with ASD and sibling relationship measures.
Table 1

Associations between typical siblings’ temperaments and child and parent perceptions of the quality of the sibling relationship

Sibling relationship

Temperament of typically developing siblings

Activity

Persistence

Emotionality

Child reports of sibling relationship quality

    Satisfaction

−.15

.20

−.16

    Positive quality

−.16

.19

−.16

    Involvement/leadership

−.14

.10

−.09

    Empathy/concern

−.18

.18

−.15

    Negative quality

.09

−.33*

.16

    Unkindness

.14

−.34*

.17

    Avoidance/embarrassment

.09

−.29*

.24

Parent reports of sibling relationship quality

    Satisfaction

−.14

.50****

−.27

    Positive quality

−.13

.28

−.14

    Involvement/leadership

−.18

.24

−.16

    Empathy/concern

−.17

.29*

−.17

    Negative quality

.07

−.36**

.19

    Unkindness

.17

−.38**

.18

    Avoidance/embarrassment

.10

−.31*

.16

Note Con’t

p < .05; ** p < .01; **** p < .001;  p < .06

Predicting Sibling Relationship Quality from the Temperaments of Both Siblings

A series of hierarchical multiple regressions were used to examine the interactive effects of the temperaments of the two siblings on the quality of sibling relationships. Persistence was the temperament dimension associated with sibling relationship quality when zero-order correlations were executed; therefore, persistence was the only temperament dimension included in the regression analyses. Six regressions were executed, predicting three child-report and three parent-report measures of the sibling relationship (Satisfaction, SIB-Positive Relationship Quality, SIB-Negative Relationship Quality). Three predictor variables were entered into each regression in the following order: persistence of the child with ASD, persistence of the sibling, and the interaction between these two variables. Four regression models were significant (see Table 2, indices in the table include change in model R2 when each predictor variable entered the model, total model R2 and adjusted model R2 at each step, and standardized Beta and F change values at each step for the last variable entered).
Table 2

Significant hierarchical multiple regressions using persistence to predict sibling relationship quality and sibling satisfaction with differential parenting

Predictor variable

Beta

ΔF

ΔR2

Model R2

Adjusted R2

Indicators of sibling relationship quality

Child ratings of positive sibling relationship quality

    Persistence—C with autism

.26

2.72

.07

.07

.04

    Persistence—sibling

.19

1.49

.03

.10

.05

    Persistence interaction

−.40

7.75**

.16

.26

.20

Child ratings of negative sibling relationship quality

    Persistence—C with autism

−.23

2.14

.05

.05

.03

    Persistence—sibling

−.31

4.19*

.10

.15

.10

    Persistence interaction

.25

2.81

.06

.21

.15

Parent ratings of child satisfaction with the sibling relationship

    Persistence—C with autism

.09

.33

.01

.01

.00

    Persistence—sibling

.46

10.04***

.21

.22

.18

    Persistence interaction

−.32

5.24*

.10

.32

.26

Parent ratings of negative sibling relationship quality

    Persistence—C with autism

−.21

2.76

.06

.06

.04

    Persistence—sibling

−.38

3.56*

.09

.15

.10

    Persistence interaction

.15

.92

.02

.17

.10

Predictor variable

Beta

Partial F

R2

Model R2

Adjusted R2

Indicators of differential parenting

Child reports of satisfaction with differential parenting

    Persistence—C with autism

.36

5.65*

.13

.13

.11

    Persistence—sibling

.29

3.79*

.08

.21

.17

    Persistence interaction

−.14

.97

.02

.23

.17

Parent reports of sibling satisfaction with differential parenting

    Persistence—C with autism

−.05

.08

.00

.00

.00

    Persistence—sibling

.53

14.09****

.28

.28

.24

    Persistence interaction

−.02

.03

.00

.28

.22

Note Con’t

p < .05; ** p < .01; *** p < .005; **** p < .001

Persistence ratings for the siblings contributed significant variance (ranging from 9 to 21%) over and above the variance contributed by the persistence ratings of the children with ASD for parent and child ratings of negative sibling relationship quality and parent ratings of children’s satisfaction with the sibling relationship. For all three outcome measures, higher persistence in the sibling was associated with more positive ratings of the sibling relationship. The interaction between the persistence ratings of the two siblings contributed significant variance for the models predicting child ratings of positive sibling relationship quality and parent ratings of satisfaction. Following the recommendations of Cohen and Cohen (1983), significant interaction terms were examined by plotting two representative slope lines, one for a low value of persistence for the child with autism (1 sd below the mean) and one for a higher value (1 sd above the mean).

The sibling relationship was rated as being less positive by the children and by their parents (less satisfying) when both siblings had temperaments characterized by low persistence (see Figs. 1 and 2). When the children with ASD were low in persistence, children’s ratings of the positive quality of the sibling relationship and parents’ ratings of the satisfaction of the typically developing children evidenced relatively steep slope lines, such that sibling relationship quality was perceived to be more positive and satisfaction was judged to be higher when the persistence of the sibling was higher. Examining the regression lines for children with ASD who were high in persistence, children’s ratings of positive relationship quality and parents’ ratings of sibling satisfaction stayed relatively stable as the persistence level of the siblings increased.
https://static-content.springer.com/image/art%3A10.1007%2Fs10803-008-0560-z/MediaObjects/10803_2008_560_Fig1_HTML.gif
Fig. 1

Interaction between parents’ ratings of the persistence of the children with ASD and the persistence of their siblings in the hierarchical regression model predicting typical siblings’ reports of positive sibling relationship quality

https://static-content.springer.com/image/art%3A10.1007%2Fs10803-008-0560-z/MediaObjects/10803_2008_560_Fig2_HTML.gif
Fig. 2

Interaction between parents’ ratings of the persistence of the children with ASD and the persistence of their siblings in the hierarchical regression model predicting parents’ reports of the typical siblings’ satisfaction with the sibling relationship

Descriptive Analyses for Sibling Satisfaction with Differential Parenting

Children’s and parents’ reports of sibling satisfaction with differential parenting were positively correlated, r = .30, p < .05. Nonetheless, when mean differences were examined, parents and their children had significantly different views, t(46) = 2.86, p < .006. Parents were more likely than their children to report that the children were unhappy about their differential parenting (mean ratings: child = 7.28, parent = 6.21). On the first item assessing differential parenting, parents were over four times more likely than their children to report that the children were unhappy about how the parents distributed their attention (35% of parents; 8% of children; data are the percent of respondents who circled either “unhappy” or “very unhappy” in response to this item). On the second item, parents were over twice as likely to report that the children were unhappy with the general level of differential treatment (37% of parents, 15% of children).

Sibling Temperaments and Satisfaction with Differential Parenting

Parents reported greater sibling satisfaction with the level of differential parenting when the typically developing siblings were rated as being higher in persistence (see top of Table 3). Only one significant relationship was detected between the temperaments of the children with ASD and ratings of sibling satisfaction with differential parenting: when the children with ASD had temperaments characterized by higher persistence, their siblings indicated greater satisfaction with differential parenting, r = .36, p < .05.
Table 3

Associations between sibling satisfaction with differential parenting and child temperament and the quality of the sibling relationship

Child temperament and sibling relationship quality

Sibling satisfaction with differential parenting

Child report

Parent report

Temperament of typically developing siblings

    Activity

−.24

−.12

    Persistence

.28

.55****

    Emotionality

−.28

−.28

Child reports of sibling relationship quality

    Satisfaction

.47****

.14

    Positive quality

.54****

.15

    Negative quality

−.48****

−.34**

Parent reports of sibling relationship quality

    Satisfaction

.38***

.46****

    Positive quality

.21

.18

    Negative quality

−.30*

−.33*

p < .05; ** p < .01; *** p < .005; **** p < .001;  p < .06

Two hierarchical multiple regressions were executed with children’s and parents’ ratings of sibling satisfaction with differential parenting as outcome variables. The predictor variables and order of entry were identical to that for regressions predicting sibling relationship quality. Results are shown at the bottom of Table 2. For children’s and parents’ ratings of sibling satisfaction with differential parenting, the persistence of the siblings contributed significant variance over and above that contributed by the persistence of the children with autism. The interaction terms were not significant in either model. An additional set of regression analyses, parallel to those above, were run with emotional intensity as the temperament variable. Neither model was significant.

Satisfaction with Differential Parenting and Sibling Relationship Quality

To address the final research question, parents’ and siblings’ ratings of sibling satisfaction with differential parenting were correlated with the parents’ and siblings’ ratings of the quality of the sibling relationship (see Table 3). Children who expressed more satisfaction with differential parenting rated their sibling relationships as being of higher quality across all relationship dimensions. Parents’ ratings of sibling satisfaction with differential parenting were correlated only with the children’s ratings of negative relationship quality. When parents and siblings reported that siblings were more satisfied with differential parenting, both parents and the siblings, themselves, rated the siblings as being more satisfied overall with the sibling relationship and characterized sibling relationship quality as less negative.

Discussion

One child temperament dimension, persistence, was found to be an important predictor of the quality of relationships between children with ASD and their typically developing siblings. Negative relationship quality, as rated by both the siblings and their parents, was lower when children had temperaments characterized by high persistence. In addition, parents perceived that their typically developing children were more satisfied with their relationship with the children with ASD when the siblings were higher in persistence. This finding was not reflected in the ratings provided by the siblings, themselves. Interactions were found between the persistence ratings of the two siblings for child ratings of positive relationship quality and parent ratings of child satisfaction with the sibling relationship. Both of these measures were at their lowest when both children were low in persistence. The deleterious impact of each child’s non-persistent temperament was exacerbated by the non-persistent temperament of the other sibling, seemingly placing the siblings at heightened risk for decreased positive relationship quality, as perceived by the siblings, and satisfaction, as perceived by parents. Both of the sibling measures were at moderate levels when both siblings were high in persistence.

Evidence for buffering effects was found. Temperament buffering occurs when one sibling has a difficult temperament and the other does not, and the more positive temperament of one child buffers the impact of the more temperamentally difficult child, minimizing negative effects on the sibling relationship. Positive relationship quality, as perceived by the siblings, was quite high for sibling pairs with dissimilar levels of persistence, regardless of whether it was the child with autism or his/her sibling who was rated as high in persistence. For parents’ ratings of children’s satisfaction with the sibling relationship, high persistence in the sibling was a more effective buffer then was high persistence in the child with ASD.

The concentration of findings for child persistence, with no similar associations for the temperamental variables of negative emotionality or activity, was unexpected. The current results suggest that children’s relationships with their siblings with ASD are enhanced when at least one child in the sibling pair continues to work on tasks even when they are difficult or frustrating, returns to tasks after being interrupted, and becomes so absorbed in tasks that they ignore outside distractions. These are the types of behaviors that define high persistence. Children with ASD who are high in persistence may spend more time actively involved in tasks and may be less likely to engage in behaviors that annoy their siblings. Siblings who are high in persistence may be more able to cope with distracting verbalizations, disturbing behavior, and other intrusions from their siblings with ASD because they ignore these potentially irritating events when possible and return to their preferred tasks more easily after being interrupted.

In addition, typically developing siblings who are highly persistent may be better able to negotiate spending mutually pleasing time with their siblings with ASD. Guralnick and Groom (1990) suggested that temperamentally persistent children have more effective relationships with peers than less persistent children because they are more likely to engage in repeated efforts to begin or reinstate social interactions. Children with autism often take longer to orient to social stimuli (Dawson et al. 1998) and can be nonresponsive to the social bids of other children (Jackson et al. 2003). Persistent children may be more tenacious in attempting to engage their siblings with ASD in social interactions, continuing their social bids until their siblings respond. Similarly, children with ASD who are high in persistence may be better able to initiate interactions and respond to the social bids of their brothers and sisters.

In addition to focusing on the direct effects of child temperaments on the sibling relationship, this study also examined the effects of child temperaments on siblings’ satisfaction with differential parenting, and the associations between siblings’ satisfaction with differential parenting and qualitative aspects of the sibling relationship. Lower persistence in siblings was associated with greater sibling dissatisfaction with differential parenting, as viewed by both parents and siblings. The persistence of the siblings predicted sibling satisfaction with differential parenting over and above the effects of the persistence levels of the children with ASD. It is plausible that when siblings spend time enthralled in activities and in learning new skills, they are less concerned with the amount of attention that parents give them as compared to their siblings with ASD. These siblings may be quite content to be left alone by parents to pursue their own activities. Conversely, siblings who tend not to lose themselves in tasks and activities may have a higher demand for parent attention and interaction, causing the children to be less satisfied with the attention that their parents give to the children with ASD. When children with ASD engage in constructive tasks for long periods of time, their parents may be more able to divide their attention between the children, resulting in greater sibling satisfaction with differential parenting.

As predicted, when siblings were dissatisfied with differential parenting, the quality of the sibling relationship suffered. Associations were the strongest between child reports of satisfaction with differential parenting and child reports of sibling relationship quality. However, findings in support of this hypothesis were not restricted to within-rater associations or to child-report measures. These findings support the findings of other studies (Kowal et al. 2002; McHale et al. 2000) that the subjective satisfaction of siblings with differential parenting is important for the sibling relationship. This study extends those findings to siblings of children with ASD.

One of the strengths of the current study is that it incorporated siblings’ and parents’ perspectives of the sibling relationship. Olson (1977) referred to different views of a relationship as representing insider and outsider perspectives. Throughout the study, differences emerged between the perspectives of parents and siblings. Furman (1993) pointed out that there is not just one accurate perspective of the sibling relationship, arguing for the importance of studying multiple perspectives. Parent and sibling perspectives would not be expected to be identical, and both perspectives have validity. It is interesting to note that the typically developing siblings who participated in the study were happier with their parents’ distribution of attention than their parents believed them to be. Nonetheless, both siblings’ and parents’ views of sibling satisfaction with differential parenting were predictive of aspects of the quality of sibling relationships.

Study findings involving the temperaments of children with ASD were less evident than those involving the temperaments of their siblings. It is plausible that individual characteristics other than temperament, such as behavior problems, may have been more predictive for children with ASD. In addition, more is still to be learned about the meaning of different temperament dimensions for children with ASD. Temperament instruments were developed and validated on typically developing children; the meaning of persistence and other temperament dimensions for children with ASD may differ from the normative sample. Although several researchers (i.e. Bailey et al. 2000; Hepburn and Stone 2006) have found that children with ASD had less persistent temperaments than comparison children, behaviors diagnostically associated with ASD would suggest high levels of persistence, albeit concentrated on a narrow range of objects or activities. This creates an interesting paradox alluded to by Hepburn and Stone (2006). The child with ASD who spends long periods of time in stereotyped perseverative play, lining up toys or twirling preferred objects, is demonstrating persistence, but not in a normative fashion. Temperament assessments focus on persistence as demonstrated across a variety of activities and toys, including a desire to persevere in learning difficult new skills. Thus, although children with ASD may demonstrate extremely high levels of persistence in certain splinter activities, many of these children might be expected to score low on persistence as measured by traditional temperament assessments.

This study had several important limitations. The sample was not randomly selected and the broad distribution of research packets precluded an analysis of the response rate. The sample size was relatively small and was skewed toward middle class, better educated, Euro-American families. Thus, generalization of study findings is limited to those families demographically similar to the study sample. Since all but two of the children had diagnoses of autism, there was no opportunity to examine sibling relationships separately for children with other ASD diagnoses. Although most of the sample was recruited from university centers focusing on ASD, with strong diagnostic protocols, there was no independent confirmation of the diagnostic status of participating children. Similarly, all of the siblings participating in the study were described by their parents as being typically developing, with no diagnosed disabilities, but it is possible, given evidence for the heritability of a broad ASD phenotype (i.e. Bailey et al. 1995), that some of these children possessed ASD-like characteristics in a milder form. A relatively simple, summative measure of satisfaction with differential treatment was used. Participants were not asked about the actual level of differential parenting, nor were observations made of parenting. Although hypotheses were derived from the premise that temperament and differential parenting influence the sibling relationship, contemporaneous data preclude causal attributions.

Although this study has limitations, findings reinforce the roles of temperament and sibling satisfaction with differential parenting as contributing to the quality of relationships between children with ASD and their siblings. They also underscore the importance of considering the characteristics of both children in the sibling relationship, rather than attributing variations in relationship quality solely to the child with a disability. Travis and Sigman (1998) suggested that siblings may be especially important for children with autism because they provide opportunities to socially interact with other children in maximally supportive conditions. Understanding the processes that promote positive, supportive relationships between children with ASD and their siblings can provide meaningful information about how to further strengthen these relationships.

Copyright information

© Springer Science+Business Media, LLC 2008