For fathers raising children with autism, do coping strategies mediate

Research in Developmental Disabilities 36 (2015) 620–629

Contents lists available at ScienceDirect

Research in Developmental Disabilities

For fathers raising children with autism, do coping strategies mediate or moderate the relationship between parenting stress and quality of life? Latefa A. Dardas a,*, Muayyad M. Ahmad b,1 a b

School of Nursing, Duke University, Durham, NC 27708, United States Faculty of Nursing, The University of Jordan, Amman 11942, Jordan

A R T I C L E I N F O

A B S T R A C T

Article history: Received 26 July 2014 Received in revised form 28 October 2014 Accepted 29 October 2014 Available online

In response to the dramatic change in the perception of fatherhood and the significant expansion in fatherhood research, this study came to fill the gap in literature and examine the possible mediation and moderation effects of coping in the relationship between fathers’ of children with autism parenting stress and quality of life (QoL). Mediation and moderation effects were examined using multiple programs and software which included hierarchical regression, structural equation modeling and special Macros added to the analysis programs to confirm the findings. None of the investigated coping strategies could mediate or moderate the stress–QoL relationship among the 101 participating fathers. This study provides interesting information on how the stress-coping-QoL relationship among fathers of children with autism can be affected by the nature of their stress provoking situation, their individual characteristics, the environment and its demands and resources, and the way fathers perceive and apply their coping responses. ß 2014 Elsevier Ltd. All rights reserved.

Keywords: Mediation Moderation Coping Parenting stress Quality of life Fathers Autism

1. Introduction The birth of a new child is considered a pleasant time for most fathers. Unfortunately, nearly one out of 68 fathers receives upsetting news that his child will demonstrate delays across all areas of development due to a pervasive chronic disability. Autism is currently considered the number one occurring developmental disability that can impose significant impacts on fathers’ ordinary roles and render them despaired, powerless, and at risk for many psychosocial problems (Lyons, Leon, & Roecker Phelps, 2010). The literature addressing fathers of children with autism suggested that once the child is introduced into the family, the whole fatherhood experience changes (Donaldson, Elder, Self, & Christie, 2011). Fathers’ chronic caregiving responsibilities may threaten their emotional, physical, and social well-being. For instance, fathers of children with autism usually experience higher levels of psychological distress and poorer quality of life (QoL) than fathers of typical children do (Ericzon, Frazee, & Stahmer, 2005; Mungo, Ruta, Arrigo, & Mazzona, 2007). Further, fathers may feel emotionally charged because of their inability to protect their children from pain or harm (Pelchat, Levert, & Bourgeois-Gue´rin, 2009). Dealing with health services may also be physically strenuous for most fathers. Moreover, some studies report that many

* Corresponding author. E-mail addresses: [email protected], [email protected] (L.A. Dardas), [email protected], [email protected] (M.M. Ahmad). 1 Tel.: (919) 684-9198. http://dx.doi.org/10.1016/j.ridd.2014.10.047 0891-4222/ß 2014 Elsevier Ltd. All rights reserved.

L.A. Dardas, M.M. Ahmad / Research in Developmental Disabilities 36 (2015) 620–629

621

fathers feel strain in building up their social networks either because of their demanding roles as care providers or because of the lack of public understanding of their unique experiences (Lewis, Skirton, & Jones, 2010). One of the most commonly searched psychological outcomes for parenting a child with autism is parenting stress. Studies have long revealed significant high levels of stress among fathers of children with autism (Dardas & Ahmad, 2014a,b; Ericzon, Frazee, & Stahmer, 2005; Seltzer, Abbeduto, Krauss, Greenberg, & Swe, 2004). Fathers report high stress usually because of their children daily management, the need to search for and deal with available services and educational programs, and the potential financial strains (Parish, Seltzer, Greenberg, & Floyd, 2004). Furthermore, fathers’ lack of knowledge about their children’s disability (Kucuker, 2006), and the lack of active planning for fathers’ engagement in all areas of service delivery (Premberg, Hellstrom, & Berg, 2008) can lead to high patenting stress. The prolonged nature of parenting stress highlights the significant short and long term risks to fathers’ mental and physical health. The entire QoL of fathers of children with autism is believed to be threatened with the presence of a child with autism (Dardas & Ahmad, 2014c,d). Poor QoL reported by those fathers reflects a multidimensional outcome that incorporates their physical health, psychological state, social relationships, and relationships with salient features of the environment. Overall, fathers raising children with autism face several challenges that make their coping and adjustment to the child’s disability an on-going complex task. Based on the prior findings, it becomes clear that fathers of children with autism are fair much better if they can have means to cope with their inevitable stressors. In fact, relevant studies reported that the ways fathers of children with autism cope with stress is more influential to their health and QoL than the stressors themselves (Dabrowska & Pisula, 2010; Hastings et al., 2005). Several links between the types of coping strategies used by fathers of children with autism and their levels of stress and QoL have been revealed (Dabrowska & Pisula, 2010; Gray, 2006; Hastings et al., 2005). For instance, the use of less emotion-focused coping and more problem-focused coping was associated with better QoL (Lyons et al., 2010). More specifically, King et al. (2006) found that adopting positive meanings to the lived experiences is a popular coping strategy among fathers experiencing stress. Other adaptive strategies included participating in religious and individual activities and emphasizing strong moral and religious standards, planful problem solving, and self-control (Sivberg, 2002). On the other hand, the reliance on passive and avoidant coping strategies was found ineffective with chronic stressors such as raising a child with autism (Hastings et al., 2005; King et al., 2006). The application of the stress-coping theories on parenting children with pervasive disabilities has yielded lots of interesting yet mixed results. In addition, there is a lag in coping research on fathers raising children with autism when compared to mothers, making it difficult to draw firm conclusions about their coping outcomes (Vacca, 2006). In their extensive review, Hastings et al. (2005) stated that it was difficult to locate published papers describing coping strategies adopted by fathers of children with autism. Further, Premberg et al. (2008) have stated that although available studies could clarify some issues for fathers, a small number of fathers were usually included and their voices were mixed with those of mothers. Therefore, it is of critical importance to understand the underlying structure of coping used specifically by fathers of children with autism. 1.1. Why this study? There are compelling practical and theoretical reasons for conducting this study. Research generally suggests that the way fathers cope with life stressors can significantly affect their children’s, partners’, families’, and ultimately their own health and well-being (Gray, 2006; King et al., 2006; Sivberg, 2002). Fathers, although labeled by some as the ‘peripheral parent’, can set the tone for the whole family’s reaction to the child with a disability (Donaldson et al., 2011). When fathers’ coping strategies fall short, it becomes difficult for them to support other family members. To date, many gaps are left unresolved in the arena of coping behaviors adopted by fathers of children with chronic disabilities in general and autism in particular. It is not well understood how fathers’ empirically derived coping responses are linked to their psychological functioning and QoL. Available literature includes an investigation for the concepts of stress, coping and QoL through bivariate relationships. However, relationships among variables are often considered more complex than simple bivariate relationships with the presence of potential effects of other variables that may work as mediators or moderators in the existing relationships (Baron & Kenny, 1986; Bennett, 2000). A mediator is a third variable that links a cause and an effect. A moderator is a third variable that modifies a causal effect (Wu & Zumbo, 2008). In other words, mediation explains the process of ‘‘why’’ and ‘‘how’’ a cause-and-effect happens, whereas moderation postulates ‘‘when’’ or ‘‘for whom’’ an independent variable most strongly (or weakly) causes a dependent variable (Frazier, Tix, & Baron, 2004). Investigating the role of such variables in psychology not only verifies researchers’ substantive theories around a phenomenon but also answers practical questions about whether an intervention or treatment program has the expected effect (Wu & Zumbo, 2008). Nevertheless, experts in this field claim that studying mediation and moderation causal nature effect is often overlooked or simply misunderstood leading to misapplication and misinterpretation in much of applied research (Frazier et al., 2004; Rose, Holmbeck, Coakley, & Franks, 2004). In this study, we aimed at investigating eight coping strategies as possible mediators or moderators in the relationship between fathers’ of children with autism parenting stress and QoL. 1.2. Theoretical framework and hypothetical model This study was guided by the Lazarus and Folkman’s Transactional Model of Stress and Coping (1984). The model addresses the concepts of stress and coping. Stress is perceived as ‘‘a relationship between the person and the environment

622


that is appraised by the person as taxing or exceeding his or her resources and endangering his or her well-being’’ (Lazarus & Folkman, 1984, p. 21). Coping is defined as the cognitive and behavioral efforts that are constantly changing to master, reduce, or tolerate a specific stressor appraised as exceeding one’s available resources and abilities. Thus, stressful experiences are perceived in this model as dynamic and reciprocal person–environment transactions. A striking point in the model is that coping strategies are the intermediate process between stressors and health outcome. Lazarus and Folkman (1984) suggested that the process of coping can change the relationship between life stressors and health. Generally, ineffective coping strategies can generate negative emotions, harmful long-term effects, and calls for additional and/or different coping efforts. On the other hand, effective coping leads to positive emotions and beneficial long-term effects. We assumed that the style of coping fathers of children with autism use to address their stressors can significantly affect their parenting experience outcomes (QoL) by working as mediators or moderators in an existing stress–QoL relationship.

2. Methods 2.1. Participants Participants consisted of 101 fathers aged 25–69 years, with a mean of 39.3 years (SD = 6.9). Almost all were married (98%, n = 99), and the majority had a secondary school or higher. Family income levels for the total sample ranged from 150 to 3800 JD (1 JD = 1.4 US Dollar). Around 84% of the fathers were employed and represented a variety of occupations (e.g., laborers, professionals, and government services). The children with autism were 85 boys and 16 girls aged 2–13 years (M = 5.9; SD = 3.2). The mean number of children fathers had, beside the child with autism, was 3.6 (SD = 1.6). 2.2. Measures A demographic questionnaire designed for the present study was used along with three other measures; the Parent Stress Index-Short Form (PSI-SF), The Revised Ways of Coping Checklist (WCC-R), and the World Health Organization Quality of Life Assessment-BREF (WHOQOL-BREF). 2.2.1. The Parenting Stress Index-Short Form (Abidin, 1995) ‘‘Parenting stress is a complex construct that involves behavioral, cognitive, and affective components and is a combination of child and parent characteristics, as well as family situational components as they relate to the person‘s appraisal of his or her role as a parent’’ (Whiteside-Mansell et al., 2007, p. 27). Based on this notion, we selected the PSI-SF scale for this study to assess fathers’ perceptions of parenting stress. The PSI-SF is a widely used measure with parents of children up to 13 years of age and is considered a robust diagnostic measure for parenting stress. The PSI-SF contains questions derived from the 102-item original PSI full-length that could be administered within a limited amount of time (Abidin, 1995). Respondents were instructed to select an answer questions on a 5-point Likert-type scale ranging from ‘‘strongly agree’’ to ‘‘strongly disagree’’. We used a validated version for the PSI-SF with parents of children with autism (Dardas & Ahmad, 2014e). Cronbach’s alpha for the total stress score revealed reliable results (.92). According to Abidin (1995), parents who obtain a total stress score above a raw score of 90 (at or above the 90th percentile) are experiencing clinically significant levels of stress. In this study, the mean score for the total stress score was 94.35 (SD = 19.28). 2.2.2. The Ways of Coping Checklist-Revised (WCC-R) (Folkman & Lazarus, 1988) The WCC-R is an empirically derived inventory composed of 66-item coping strategies that people use to deal with the internal and/or external demands of specific stressful encounters (Folkman & Lazarus, 1988). The measure was adapted to provide a clear description for coping through deriving eight classifications, including: positive reappraisal, confrontive coping, planful problem-solving, seeking social support, distancing, self-controlling, escape-avoidance, and accepting responsibility. Those categories provide a clear description for the thoughts and actions an individual uses to cope with a specific stressful encounter. This study had one of its strengths in utilizing the eight coping subscales rather than the general two-classification model (problem versus emotion-focused coping) that was used in previous relevant research. Indeed, Folkman and Moskowitz (2004) have argued that relying on the two classifications for the synthesis of findings across studies may run the risk of distorting important differences within categories. Fathers were required to respond to a specific stressor (raising a child with autism) and subsequently respond to each statement in the scale by expressing the extent to which they used the strategy in coping with that stressful situation on a 4-point scale from 0 (not used) to 3 (used a great deal). Scores are additively derived from individual items and divided by a total score to provide relative scores for a total of eight scales. The WCC-R is considered one of the most widely used comprehensive and psychometrically robust measures of coping (Lundqvist & Ahlstro¨m, 2006). Table 1 describes fathers’ coping responses. The internal consistency ranges from 0.68 to 0.79 for the eight subscales (Folkman & Lazarus, 1988). In this sample, Cronbach’s alphas for the eight subscales ranged from moderate to high.


623

Table 1 Description for the eight coping subscales. Coping subscales (number of items)

Mean

Percentage

Standard deviation

Range

Positive reappraisal (7) Confrontive coping (6) Planful problem solving (6) Seeking social support (6) Distancing (6) Self-control (7) Escape avoidance (8) Accepting responsibility (4)

20.13 12.96 15.95 16.34 13.85 17.56 17.74 9.04

72 54 66 68 58 63 55 57

3.29 3.07 3.48 3.88 3.17 3.40 4.76 2.57

12–27 7–21 6–24 8–24 8–23 9–25 10–29 4–16

2.2.3. The World Health Organization Quality of Life Assessment-BREF Self-administered Instrument (WHO, 1996) The WHOQOL-BREF assessment instrument was developed by the WHOQOL Group with fifteen international field centers in an attempt to develop a QoL assessment that would be applicable cross-culturally (WHO, 1996). The instrument has been developed to provide a short form QoL assessment that looks at domain level profiles, using data from the WHOQOL-100. The WHOQOL-BREF contains a total of 26 questions inquiring about respondents’ overall perception of QoL and health beside their physical, psychological, social, and environmental health. The measure scores are scaled in a positive direction, with higher scores denoting higher levels of QoL. The score 60 out of 100 is considered as the midpoint where QoL is judged on the WHOQOL-BREF measure to be neither good nor poor (Skevington, Lotfy, & O’Connell, 2004). The mean score for the total QoL score in this study was 74.17. A validated Arabic version of the WHOQOL-BREF by Dardas and Ahmad (2014f) was used in this study. Cronbach’s alpha for the total QoL score was .91. 2.3. Analyses Prior to conducting any type of analyses, we performed basic data screening activities to ensure the accuracy and legibility of data entry and assess the normality of the continuous variables. Missing data were replaced in accordance with their manuals’ scoring procedure that allows for calculating missing data. That is, the mean of the items from the subscale from which the item was missing was calculated and then rounded to the nearest whole number, and assigned to the missing item. However, there were only seven cases with few items unanswered. The data were also inspected for outliers, defined as values greater than 3.5 standard deviation units from the sample mean for a given variable. Univariate outlier analysis using box plots indicated the absence of univariate outliers for all the continuous variables. Descriptive and inferential analyses were applied to the cleaned data set to describe the sample and to address the research hypotheses using the Statistical Package for Social Sciences (IBM, SPSS, 21). We conducted moderation and mediation tests using multiple programs and software which included SEM, SPSS, AMOS, SOBEL test, and PROCESS Macros for SPSS. In order to confirm if a variable is making a moderation effect in the relationship between an independent variable and a dependent variable, the nature of this relationship must change once the moderator variable changes. The inclusion of the interaction in the model should lead to a significant explanation in the outcome variance better than without. The following two steps were followed to examine the moderation effect: in step one of the regression model, we entered the predictor and the moderator. Both effects as well as the explained variance (R2) of the model should be significant. In step two of the regression model, we entered the interaction effect and checked if it was significant as well as the change in R2 of the model. Moderation is considered to occur if previous results were significant. In other words, moderation occurs when the direction, strength, or both of the relation between an independent and a dependent variable are affected by a third variable, which is termed a moderator (Baron & Kenny, 1986; Wu & Zumbo, 2008). We evaluated the accuracy of the mediation effect using several types of tests. We conducted the traditional hypothesized method using SPSS regression analysis (Baron & Kenny, 1986) and then, we did the analysis with bootstrap resampling. In addition, SPSS macros for computing mediated effects with Sobel test was performed (Preacher & Hayes, 2004). Structural equation modeling using AMOS was also used. Furthermore, the Sobel test was used to test the significance of a mediation effect. The Sobel test provides means to determine whether the reduction in the effect of the independent variable, after including the mediator, is a significant reduction and therefore whether the mediation effect is statistically significant (Sobel, 1982). It is important to report here that the literature on the process of testing moderation effect still ambiguous. Some researchers hypothesized that a regression model is designed to predict the outcome variable from both the predictor variable and the moderator variable in first step and then adding the interaction variable in the second step. The effects of the variables as well as the model in general (R2) should be significant (Aiken & West, 1991; ORSP, 2013). However, Bennett (2000) stated that in the first step of the regression, the independent variables and the moderator are entered into the model as predictors. Bennett claimed that the independent variables are not necessary to be significant predictors of the outcome variable in order to test for an interaction effect. In the next step, if the interaction outcome explains a significant amount of variance in the dependent variable, then the moderator effect exists (Bennett, 2000).

624


2.4. Study protocol The population (fathers) for this study was drawn from public and private Jordanian centers that provide specialized diagnostic, remedial, and teaching services in the field of disabilities (including autism). The Jordanian Ministry of Social Development and the Higher Council for the Affairs of Persons with Disabilities provided the researchers with the names of the licensed centers. Prior to collecting the data, the study received ethics approval from the University and Faculty of Nursing and the participating centers. The first author personally contacted the head of each center in order to gain approval for obtaining participants. The study was described to center officials and they were asked for their approval for fathers’ participation. The head of each center was then asked to invite eligible fathers to meet the first author in their children’s centers. In addition, we sent an invitation letter to the fathers including all the information regarding the researchers and the study objectives, inclusion criteria, and rights of the participant. The letter also confirmed that participation was voluntary and that the participant could withdraw from the study at any point. Since the PSI-SF used in this study is standardized and valid for use with parents of children up to 13 years of age, we invited only fathers of children aged up to 13 years to participate in this study. Fathers also had to have just one child diagnosed with autism to ensure that fathers would focus on the same child when they completed the questionnaire. Further, fathers must be living in the same household with the child with autism. Fathers who agreed to participate met the researcher (first author) at their children’s centers where the researcher confirmed the study’s protocol and inclusion criteria. Fathers were informed that the researcher will have an access to their children’s records. They were also assured that this access was only for the purpose of confirming their children’s clinical diagnosis. The researcher was then responsible for distributing and collecting the study questionnaires, and making sure they were voluntarily given to eligible fathers. The first author, rather than any other data collector, chose to be responsible for gathering the data in order to assure that the study purpose was clarified accurately to the fathers, and increase the number of the participants by allowing them a convenient access to the researcher to answer any question they may raise regarding the study. Indeed, it was evident in this study that fathers were willing to participate when they were personally invited and interviewed by the researcher. Data collection was conducted on Saturdays to avoid difficulty in obtaining responses from fathers due to job obligations. Some of the eligible centers agreed to participate but did not permit any direct contact with the fathers. For those, the research materials were sent home by the classroom teacher via the child. Special closed packages were prepared including the questionnaire set, an empty envelop, and a letter including the following clarification points: (A) the inclusion criteria for the study (fathers were asked not to complete the questionnaires if they did not meet any of the criteria), (B) completing all the questionnaires and returning it back in a sealed envelope to the center, and (C) avoiding sharing or discussing responses with wives or any other person to assure the reliability and validity of the responses. Returning the completed questionnaire was considered to be the respondent’s consent to participate in the study. The self-reported questionnaire utilized in this study had its pros over several other data collection methods. For instance, the questionnaire was properly quick to administer and took into consideration the fathers’ time constraints. Second, the use of such method permitted fathers who could not attend the meetings with the researcher to examine and complete the questionnaire in the comfort and familiarity of their own home. Moreover, the questionnaires ensured the confidentiality to all the participants. Data collecting procedures also ensured that no personally identifiable information would be associated with the study questionnaires. Only coded data were entered and analyzed. 3. Results 3.1. Mediation models We conducted the mediation tests using multiple analysis techniques and statistical software to validate the results. We ran multiple regression analyses to examine the eight coping strategies as mediators between stress and QoL using SPSS (IBM, 21; 2012) (Table 2). We conducted mediation analyses using the bootstrapping method with 5000 bootstrap resamples and bias-corrected confidence estimates (MacKinnon, Lockwood, & Williams, 2004; Preacher & Hayes, 2004). We used structural equation modeling (SEM) with Analysis of Moment Structure (AMOS, version 21.0) because it allows testing all components of the mediation model simultaneously (SEM) (Arbuckle, 2012). In addition, we used multiple mediation analyses with special software for testing the significance of the indirect effects and contrasting the strengths of indirect effects through eight mediators (Preacher & Hayes, 2008). For this analysis, the SPSS macro program developed and updated by Preacher and Hayes (2008) was used. In the multiple regression analysis, we assessed all the coping strategies separately for their mediation effect. First, each mediator was regressed on the independent variable, and each dependent variable was regressed on the suggested mediator. As shown in Table 2, in each model, one of these equations was not significant. Furthermore, when controlling for mediators effect, the relationship between the independent variable and the dependent variable remained significant. Thus, results of the mediation analyses in the eight models confirmed that none of the coping strategies had a full direct mediating effect. Model fit through SEM was examined using the chi-square statistic, the Comparative Fit Index (CFI), the Tucker–Lewis Index (TLI), and the root mean square error of approximation (RMSEA). The mediation path model showed a good fit with the data: x2 (df = 164) = 583, p < .001, CFI = .91, TLI = .90, RMSEA = .06 (CI = .05, .07). Overall model fit ranged from moderate to


625

Table 2 Testing coping strategies in 8 mediation models between stress and QoL with bootstrapping. Mediators

B

t

p value

95% CI

Positive reappraisal A–B B–C A–C0

.02 1.16 .34

.92 3.18 5.49

.360 .002