Profiles of the Parenting Experience in Families of Autistic Children

Abstract

Parents of autistic children experience more parenting stress and are at increased risk for poor mental and physical health compared to parents of neurotypical children; however, not all parents are distressed. The present study used a person-centered analytic approach to identify profiles of the parenting experience in a sample of 183 mothers and fathers of an autistic child (5–12 years old) and to examine associations between profile membership and child outcomes. Results indicated three profiles for mothers: Adaptive (41.1%; high authoritative parenting, lowest stress, and highest competence), Average (42.1%; sample average of all parenting indicators), and Distressed (16.8%; high stress, low competence, maladaptive parenting strategies). Fathers were classified into four profiles: Adaptive (33.3%), Average (37.7%), Distressed - Permissive (15.3%; high stress, low competence, permissive parenting strategies), and Distressed – Authoritarian (13.6%; some stress, lowest competence, authoritarian parenting strategies). The profiles differed on child internalizing and externalizing symptoms and autism symptom severity. Comparative analysis also revealed that children did better when at least one parent was included in the Adaptive group. Implications of these findings are discussed and include fostering empowering messages to parents as well as providing useful new insight in the context of family-focused interventions.

Autism is a neurodevelopmental condition that is present early in life, characterized by impairments in social-communication and repetitive behaviors and/or restricted interests (American Psychology Association, 2013). Over the past twenty years, a substantial body of research has examined the subjective experiences of parents of autistic children (Corcoran, Berry, & Hill, 2015; Karst & Van Hecke, 2012; Keenan et al., 2010). This research has largely taken a variable-centered approach, describing the mean score on measures of parenting experiences such as parenting stress, competence, and styles. This analytic (and theoretical) approach is useful for describing the parenting experiences of the ‘average’ parent of an autistic child. The limitation with this approach, however, is that a focus on the sample average obscures meaningful variability among the experiences and outcomes of parents of autistic children. The goal of the current study was to use a person-centered analytic approach, as opposed to variable-centered approach, to identify different profiles of parenting experiences in a sample of 183 mothers and fathers of an autistic child aged 5–12 years and to examine associations between profile membership and child internalizing, externalizing, and autism symptoms.

Parenting in the context of autism can be stressful. High caregiving demands (Kuhlthau et al., 2014; Sawyer et al., 2010), challenges with child behaviors (Davis & Carter, 2008), inflexible family routines (Boyd, McCarty, & Sethi, 2014; Marquenie et al., 2011), reduced career opportunities and more career interruptions (Gray, 2002; Redd, Fewster, & Gurayah, 2019; Stoner & Stoner, 2016), limited and less satisfying leisure time (Walton, 2019), social isolation (Dunn et al., 2001; Kinnear et al., 2016), increased financial burden (Horlin et al., 2014; Parish et al., 2012), and stigma associated with caring for an autistic individual (Papadopoulos et al., 2019) have all been implicated as sources of stress for parents of autistic children. The average parent of an autistic child has been found to be highly stressed compared to both parents of neurotypical children as well as parents of children with other types of disabilities (Dabrowska & Pisula, 2010; DesChamps et al., 2020; Estes et al., 2013; Hayes & Watson, 2013). However, not all parents of autistic children report high stress, and there may be other important resilience factors that reduce negative effects of high stress.

One factor that may contribute to resilience in the face of stressful parenting environments is parent’s sense of competence. Perceived parenting competence, or an individual’s behavioral, affective, and cognitive belief in their effectiveness as parents (Coleman & Karraker, 1998), is related to outcomes for both parents and children (see Kurzrok, McBride, & Grossman for an in-depth review of this literature). Research on sense of parenting competence, sometimes referred to as parental self-efficacy, in parents of autistic children generally mirrors findings from the neurotypical literature (e.g., Hastings & Brown, 2002; Kuhn & Carter, 2006; Lindsey & Barry, 2018). For instance, one study of families of children with a range of neurodevelopmental disabilities found a negative correlation between parents’ sense of competence and stress (Hassall et al., 2005), but little research has examined variability in perceived sense of parenting competence and how that sense of parenting competence is associated with other indicators of the parenting experience in the context of autism.

Parenting practices and attitudes, or parenting styles, are also an important aspect of the parenting experience (Baumrind, 2009). Parent-centered attitudes and parenting practices that emphasize parent authority and child obedience (authoritarian style) or parenting styles that lack boundaries (permissive style) have been consistently associated with poorer child outcomes in both neurotypical and autistic children (e.g., Greenlee et al., 2021; Johnston, Park, & Miller, 2018). In addition, parenting stress has been positively associated with greater use of both authoritarian and permissive parenting in neurotypical children and autistic children (e.g., Hutchison et al., 2016). A recent meta-analysis of observed parenting behaviors found parents of autistic children to be more controlling and exhibit more negative parenting behaviors compared to parents of neurotypical children (Ku, Stinson, & MacDonald, 2019). Additionally, different parenting styles were found by child in a study of mothers of children with disabilities. mothers of a school-aged child (9–11 years old) were found, on average, to use less authoritative and more permissive parenting styles compared to mothers of young children (3–5 years) and mothers of neurotypical children. Mothers of children with disabilities who used the highest levels of controlling parenting behaviors also reported the highest levels of parenting stress (Phillips, Conners, & Cutner-Smith, 2017).

Understanding the ‘average’ parent experience is critical for effective intervention development; however, a focus on sample means has created an image of the ‘stressed’ parent with low subjective parental competence and maladaptive parenting styles. Yet, there are likely more diverse parenting experiences present, with research supporting a multidimensional view of the parenting experience (Zheng et al., 2017). It may also be that different parenting experiences do not always go hand-in-hand. For instance, some parents may not report high levels of stress while also feeling a lack of competence in their parenting. Moreover, some parents could feel highly competent but use maladaptive parenting styles. It is also worth considering that mothers and fathers may have different subjective parenting experiences. Most of our current understanding of parenting in the context of autism relies on maternal reports but contemporary views of parenting highlight the importance of fathers as important contributors to the social and emotional development of their children (Keown, Franke, & Kaur, 2018). The few studies that have included fathers suggests both similarities and differences in their parenting experiences compared to mothers (Hartley et al., 2011; Hickey et al., 2020; Hickey, Nix, & Hartley, 2019). For example, mothers of autistic children commonly report higher stress compared to fathers (Foody et al., 2015; Grebe et al., 2022) and some evidence suggests mothers endorse more authoritative parenting styles compared to fathers who are more likely to endorse authoritarian parenting (Greenlee et al., 2021).

Finally, little is known about how different parenting experiences may map onto child functioning in the context of autism. Parents plays a central role in child outcomes across theories of child development (e.g., Bornstein, 2015; Darling & Steinberg, 1993) with parenting hypothesized to be a key mechanism through which parents exert influence in both positive and negative ways (e.g., Greenlee et al., 2021). While there is a preponderance of evidence pointing to ways in which having an autistic child impacts parents (e.g., mental health) and the parenting experience (e.g., stress, self-efficacy, parenting styles), the link between parent factors and outcomes in autistic children remains understudied. In fact, it may be that parents play an especially important role in the lives of autistic children. In addition to acting as a source of nurturance, physical and emotional care and support, education, and socialization (Sanders & Turner, 2018), parents of autistic children are also advocates and social mediators (Schopler & Mesibov, 2013), and they play a primary role in decision making around child intervention, particularly during early and middle childhood (Wilson et al., 2021). Thus, understanding the parenting experience has implications for a host of child outcomes in the context of autism.

The current study addresses critical gaps in the literature by using a person-centered analytic approach to understand the parenting experience of 183 mothers and fathers of an autistic child aged 5–12 years. Our first aim was to use latent profile analysis to identify different parenting profiles in mothers and fathers drawing on indicators of parenting stress, parenting competence, and parenting style. Applied to parenting an autistic child, this approach provides a more nuanced understanding of the parenting experience and has implications for identifying which parents are in most need of support as well as what areas of the parenting experiences are ripe for intervention. Our second aim was to examine associations between the identified parenting profiles and the autistic child’s level of autism symptoms and behavioral and emotional challenges. In line with previous profile-based research in autism samples (e.g., Hickey, Nix, & Hartley, 2019), we hypothesized that profiles characterized by high levels of parenting stress, low levels of competence, and/or maladaptive parenting styles would be associated with more severe internalizing, externalizing, and child autism symptoms. The goal of this study is not to identify parents or parenting as a cause autistic childrens’ challenges, both theoretically and methodologically given the cross-sectional nature of the study. We aim to highlight the important ways that parents and parenting make a difference in children’s lives to better understand how we can help families and autistic children be well.

Method

Participants

Data were drawn from the baseline visit of the Families in Autism project (e.g., Hickey et al., 2020) project, a longitudinal study designed to characterize families of autistic children. Eligible families included 188 cohabitating, heterosexual couples in a relationship for at least three years with a child between 5 and 12 years of age with a documented diagnosis of autism spectrum condition from a medical or education professional. Documentation had to include a score above the Autism Spectrum cutoff (69% reported scores above the Autism cutoff, with the remaining participants meeting criteria for an Autism Spectrum diagnosis) on the Autism Diagnostic Observation Schedule (ADOS-2; Lord et al., 2012). In cases where multiple children had been diagnosed with autism, the oldest child with autism was selected as the target child. Couples did not have to be married to participate nor was it a requirement to be the biological parent of the autistic child. Due to missing data, one family was excluded from mother profiles (final N = 185) and five families were excluded from father profiles (final N = 183). The sociodemographic characteristics reflect those of the area from which the sample was drawn. According to the United States Census (US Census Bureau, 2020), 79% of the area identifies as Caucasian (84% of the current sample), with a median income of $75,000 (34.6 % of the sample had an income less than $70,000; 38.2% of the sample had an income over $100,00 compared to 27% according to US Census), and is fairly well educated (31% have a college degree compared to 37% of mothers and 33.2% of fathers in this sample). A full description of parent and child sociodemographic information can be found in Table 1. There was no community involvement in the reported study.

Table 1.

Family, parent, and child socio-demographic characteristics

Demographic Items
Family
Annual Household Income, n (%)
$19,999 or less	2(1.1)
$20,001 – 39,999	13(7.4)
$40,000 – 59,999	27(14.9)
$60,000 – 79,999	33(18.2)
$80,000 – 99,999	34(18.8)
$100,001 and greater	72(39.8)
Missing (n=8, 4.2%)
Additional child with SHCN (yes), n (%)	67(35.6)
Length of relationship, M years (SD, range)	14.55(5.59, 1–31)
Child
Age, M years (SD, range)	7.92 (2.25, 5–12)
Sex (male), n (%)	160 (85.6)
ID status (yes), n (%)	65(34.8)
Parent	Mother	Father
Age, M years (SD, range)	38.73(5.61, 24–54)	40.76(6.19, 22–60)
Relationship to child, n (%)
Biological parent	167(88.4)	167(88.4)
Step-parent	5(2.6)	10(5.3)
Adoptive parent	5(2.6)	5(2.6)
Previously married (yes), n (%)	25(13.2)	25(13.2)
Race/Ethnicity, n (%)
White, non-Hispanic	158(84)	154(82)
Other	30(16)	34(18)
Education, n (%)
Less than high school	0(0.0)	4(2.1)
Some high school	2(1.1)	4(2.1)
High school degree or GED	9(4.8)	22(11.8)
Associates or technical school	19(10.3)	18(9.6)
Some college	30(16.3)	29(15.5)
College degree	68(37.0)	62(33.2)
Some graduate school	10(5.4)	11(5.9)
Master’s degree	32(17.4)	29(15.5)
Doctoral or Law degree	6(3.3)	5(2.7)

Note: SHCN = special healthcare needs; M = mean; SD = standard deviation

Procedure

The larger longitudinal study was approved by the Institutional Review Board at a university in the Midwestern U.S. Participants were recruited from local schools, ASD clinics, community centers, and ASD research registries. All participants provided informed consent. Families completed a 2-hour in-home or laboratory visit that included questionnaires about the parenting experience, the child’s emotional and behavioral functioning, and the child’s autism symptoms. Parents each received $50 for this portion of the study.

Measures

Parenting Styles

Each parent completed the 62-item Parenting Styles and Dimensions Questionnaire (PSDQ; Robinson et al., 2001). Three subscales were created reflecting the three original dimensions of Baumrind’s parenting styles: authoritative, authoritarian, and permissive parenting. Parents rated how often they engaged in parenting behaviors on a 5-point scale (1 ‘Never’ to 5 ‘Always’). A mean score for each style was computed for each parent, with good subscale reliability in the current study (Cronbach’s α = .73–.90 in mother report; Cronbach’s α =.74−.92 in father report).

Perceived Parenting Competence

A subscale of the Family Experiences Questionnaire (FEQ; Frank, Jacobson, & Aver, 1988) was used to assess the extent to which parents felt they were doing a good job as a parent. Mothers and fathers independently completed the Perceived Competence subscale of the FEQ, which consists of 15-items (e.g., “I have the knowledge I need to be a good parent”, “I know that I am doing a good job as a parent”) rated on a scale from 1 (Strongly Agree) to 4 (Strongly disagree). The subscale exhibited good internal reliability for both mothers (α = .88) and fathers (α = .89).

Parenting Stress

Mothers and fathers independently completed the Burden Interview (Zarit et al., 1980), a 29-item measure of personal distress associated with parenting. Items were rated on a four-point scale from 0 (Not at all) to 3 (Extremely). The current study used the total score as an indicator of global parenting stress (α = .87 in mothers; α= .86 in fathers), with higher scores suggesting greater stress. Previous studies have used this measure in parents of autistic children (e.g., Hartley et al., 2016; Rodriguez et al., 2019).

Child Emotional & Behavioral Symptoms

The Child Behavior Checklist (CBCL; Achenbach & Rescorla, 2001) is a widely used measure of child emotional and behavioral symptoms. The CBCL includes 140 questions rated on a three-point scale from 0 (not true) to 2 (very true or often true) and raw scores of the Externalizing (α = .92 and .90 in mothers and fathers, respectively) and Internalizing (α = .84 in mothers; α= .82 in fathers) subscales were used in the current analysis.

Child Autism Symptoms

Parents independently completed the Social Responsiveness Scale-2 (SRS-2; Constantino & Gruber, 2005). The SRS-2 is a 65-item measure assessing a child’s level of social-communication impairment and severity of restricted and repetitive behaviors. Items are rated on a scale of 1 (Not True) to 4 (Almost Always True) regarding the child’s behavior in the past six months. Items were summed to create two subscale scores: Social Communication Index (SCI; a measure of the severity of social-communication impairment; Cronbach’s α mother SCI = .93, father SCI = .92) and RRBs (a measure of the severity of restricted interests and repetitive behaviors associated with ASD; Cronbach’s α mother RRB = .83, father RRB = .80). Higher scores reflect more impairment.

Data Analysis Plan

The primary analysis for the current study was latent profile analysis (LPA) conducted in Mplus (Version 8.3; Muthén & Muthén, 1998–2017). LPA is a person-centered analytic technique in which patterns of observed scores are used to identify meaningful unobserved subgroups in a sample. Person-centered analysis tests the for the existence of sample subgroups in which individuals within a subgroup are more alike than individuals across groups (Jung & Wickrama, 2008). In other words, the “average” experience is not assumed to most accurately reflect all members of a given sample. Rather, each identified subgroup has higher or lower levels of certain characteristics that that are both quantitatively and qualitatively different from other subgroups in the sample. For this study, we used a three-step approach to LPA to classify mothers and fathers into conceptually and empirically distinct profiles based on their reports on three indicators of the parenting experience: parenting style, parenting competence, and parenting stress. Because of documented differences in their parenting experiences (Hartley et al., 2011; Hickey et al., 2020; Hickey et al., 2019), stress levels (Foody et al., 2015; Grebe et al., 2021), and parenting styles (Greenlee et al., 2021), mothers and fathers were run in separate LPA analyses.

The first step in the LPA was to determine the number of differentiated profiles in the sample by fitting a series of models beginning with the specification of a one class model. The number of classes were then incrementally increased until adding classes no longer improved the model. Model selection criteria were based on guidelines proposed by Ram and Grim (2009). We compared all models using Information Criteria based on a number of fit statistics, including Sample-Size Adjusted Bayesian Information Criteria (ABIC) and the Akaike Information Criteria (AIC). Lower values indicate a better fit. Second, we examined Entropy values to assess classification accuracy. Entropy is a measure of data-model fit used in person-centered analyses to empirically gauge the separation between membership profiles or classes. Values closer to 1 indicate more accurate results (i.e., more separation between profiles), with scores above 0.70 indicating greater accuracy in class membership prediction. Next, we used the Bootstrap Likelihood Ratio Test (BLRT) to evaluate whether a more complex model fit the data better than a more parsimonious model of k-1 classes. In other words, the BLRT provides a p-value to determine whether the addition of another class provides a statistically significant improvement in model fit. Finally, we looked at the sample size of the smallest class to determine if that class represented a meaningful proportion of the overall sample. We considered classes with fewer than 9 participants (or less than 5% of the sample) potentially spurious unless grounded in theory or previous research (Baur & Curran, 2004).

After identifying the optimal number of profiles for mothers and fathers, the Bolck-Croon-Hagenaars (BCH) method was used to determine individual class membership probabilities and evaluate the effects of four child-related covariates: externalizing symptoms, internalizing symptoms, severity of autism-related social communication challenges, and restricted and repetitive behaviors (Asparouhov & Muthén, 2014; Bolck, Croon, & Hagenaars, 2004). The advantage of the BCH method is that it accounts for individual-level uncertainty in profile classification while modeling the effects of covariates with profiles (Ferguson, Moore, & Hull, 2020).

Results

Means, standard deviations, and ranges for all study variables can be found in Table 2 and correlations amongst study variables are presented in Table 3. Correlations among the five observed measures of the parenting experience were strong but not overlapping, suggesting that measures provided unique information.

Table 2.

Descriptive statistics for study variables used in the profile analysis

	Mean (SD)		Min-Max		Skew		Kurtosis
	Mother	Father	Mother	Father	Mother	Father	Mother	Father
Authoritative Parenting	3.96 (.48)	3.77 (.56)	2.19–4.85	2.04–4.85	−.644	−.486	.782	.019
Authoritarian Parenting	1.84 (.38)	2.03 (.48)	1.10–2.90	1.10–3.80	.283	.697	−.554	.358
Permissive Parenting	2.10 (.46)	2.04 (.43)	1.07–3.53	1.07–3.40	.463	.410	.272	.163
Parenting Competence	1.93 (.48)	1.95 (.49)	.47 −2.87	.53–3.00	−.382	−.301	.091	−.082
Parenting Stress	19.18 (9.87)	17.85 (8.71)	3.00–51.00	0–49.00	.478	.508	−.478	.054

Note. For all measures, a higher score indicates “more” of that construct. For instance, a higher score on authoritarian parenting indicates higher levels of authoritarian parenting; a higher parenting stress score indicates a parent endorses more parenting stress.

Table 3.

Correlations among parenting variables and child outcomes.

		1	2	3	4	5	6	7	8	9
1.	Authoritative Parenting		−.405**	−.389**	.509**	−.313**	−.230**	−199**	−.217**	−.211**
2.	Authoritarian Parenting	−.128		.344**	−.400**	.387**	.270**	.116	.132	.166*
3.	Permissive Parenting	−.408**	.165*		−.387**	.379**	.269**	.205**	.190**	.210**
4.	Parenting Competence	.415**	−.322**	−471**		−.635**	−.284**	−.324**	−.241**	−.216**
5.	Parenting Stress	−.395**	.367**	.432**	−.668**		.523**	.427**	471**	.452**
6.	Child Ext symptoms	−.154*	.276**	.289**	−.290**	.476**		.621**	474**	.469**
7.	Child Int symptoms	−.060	.248**	.154*	−.242**	.403**	.620**		.536**	.444**
8.	Social Comm.	−.192**	.168*	.157*	−.234**	.454**	.548**	.549**		.795**
9.	Restricted & Rep Bs	−.100	.140	.159*	−.222**	.456**	.583**	.523**	.839**

Note. Mother correlations are below the diagonal and father correlations are above the diagonal; Ext = externalizing;

Int = internalizing; Social Comm = social communication skills; Rep Bs = repetitive behaviors;

^*p < .05;

^**p < .01;

for all measures, a higher score indicates “more” of that construct. For instance, a higher score on child internalizing symptoms indicates the endorsement of more internalizing symptoms in the child; a higher score on social communication indicates and restricted/repetitive behaviors indicates more challenges with these autism symptoms.

Latent Profile Analysis

The fit indices for all models are summarized in Table 4. For mothers, fit indices indicated similarities in a 3-, 4-, and 5-class solution, with the lowest AIC and ABIC in the 5-class solution; however, the addition of a fifth class did not significantly improve model fit above the 4-class solution (BLRT p = .26). Thus, the 5-class solution was dropped from consideration. Entropy values were above 0.7 for both the 3- and 4-class models, indicating good accuracy in membership classification. Additionally, the addition of a fourth class significantly improved model fit above the 3-class solution (BLRT p = .024). An examination of the sample sizes in each profile of the 4-class solution, however, revealed that the smallest profile included less than 5% of the overall sample, suggesting potentially spurious findings. After visually comparing the 3- and 4-class profiles, we determined that the additional fourth class did not provide meaningful information about this small group of parents (n = 7). Thus, the 3-class solution was chosen for mothers.

Table 4.

Estimated Models with 1–5 Classes and Model Fit Indices

Mother Profiles
Class	AIC	BIC	ABIC	Entropy	BLRT p=value	% in smallest class
1	2640.023	2672.226	2640.553	-----	------	------
2	2461.408	2512.934	2462.257	0.779	0.001	40%
3	2427.688	2498.535	2428.855	0.742	0.001	16.7%
4	2413.563	2503.733	2415.048	0.801	0.024	4%
5	2409.901	2519.393	2411.705	0.792	0.26	2%
Father Profiles
Class	AIC	BIC	ABIC	Entropy	BLRT p=value	% in smallest class
1	2604.541	2636.636	2604.964	-----	------	------
2	2426.467	2477.819	2427.144	0.744	0.001	41%
3	2398.412	2469.02	2399.342	0.728	0.001	21%
4	2386.007	2475.872	2387.191	0.746	0.012	14%
5	2381.882	2491.004	2382.320	0.783	0.252	5%

Note. BLTR=Bootstrapped Likelihood Ration Test. BLTR is a log-likelihood difference test. A value of p<.05 indicates that a k-class model provides better fit than a k-1 class model. Bolded values indicate the optimal model.

The fit indices for the father LPA indicated a 5-class solution to fit best according to both the AIC and ABIC. Similar to the mother LPA models, all Entropy values indicated good classification accuracy (i.e., values above 0.7), with the highest value in the 5-class solution. However, the BLRT was not significant for the 5-class solution (p = .252) and the smallest class contained only 5% of the sample (n =9) suggesting that the additional fifth class did not add statistically or conceptually to the 4-class model. For these reasons, the 4-class solution was chosen for fathers.

Mother Profiles

The latent profiles of the 3-class model for mothers are presented in Figure 1 as profiles 1–3 (solid lines). Standardized scores on the five observed measures for each profile can be found in Table 5. Profile 1 (41.1%) consisted of an Adaptive Parenting Experience; mothers displayed high levels of authoritative parenting, high parenting competence, and low levels of parenting stress. Profile 2 (42.1%) included mothers who reported an Average Parenting Experience; they exhibited (sample) mean levels of self-reported competence and stress and used a mixture of all three parenting styles. Finally, profile 3, representing 16.8% of the sample, consisted of mothers who reported high levels of stress, low competence, and used more authoritarian and permissive parenting strategies. This profile was considered the Distressed Parenting Experience.

Figure 1.

Visual representation of both mothers (lines labeled 1–3) and fathers (dashed or dotted lines labeled 4–6) parenting experience profiles from the latent profile analysis. Profile 1: Mother Adaptive. Profile 2: Mother Average. Profile 3: Mother Distressed. Profile 4: Father Adaptive. Profile 5: Father Average. Profile 6: Father Permissive-Distressed. Profile 7: Father Authoritarian-Distressed.

Table 5.

Means of Parent Profile Variables

Mother Profiles
	Average n = 78 (42.1%)		Adaptive n = 76 (41.1%)		Distressed n = 31 (16.8%)
Parenting Variable	Mean	SE	Mean	SE	Mean	SE
Authoritative Parenting	−0.115	0.119	0.508	0.149	−0.953	0.501
Authoritarian Parenting	0.244	0.235	−0.486	0.119	0.585	0.247
Permissive Parenting	0.357	0.185	−0.657	0.14	0.726	0.245
Parenting Competence	−0.187	0.228	0.75	0.143	−1.366	0.279
Parenting Stress	0.224	0.273	−0.811	0.121	1.424	0.245
Father Profiles
	Average n = 69 (37.7%)		Adaptive n = 61 (33.3%)		Distressed – Permissive n = 28 (15.3%)		Distressed – Authoritarian n = 25 (13.6%)
Parenting Variable	Mean	SE	Mean	SE	Mean	SE	Mean	SE
Authoritative Parenting	−0.183	0.153	0.729	0.122	0.139	0.181	−1.246	1.042
Authoritarian Parenting	0.238	0.156	−0.646	0.208	−0.423	0.702	1.378	0.205
Permissive Parenting	0.194	0.167	−0.709	0.112	0.284	0.618	0.826	0.162
Parenting Competence	−0.034	0.114	0.901	0.112	−0.73	0.597	−1.198	0.335
Parenting Stress	−0.123	0.289	−0.831	0.197	1.187	0.184	1.00	0.23

Note. Parenting variables are standardized so that all variables have a mean of 0 and a standard deviation of 1.

Father Profiles

The latent profiles of the 4-class model for fathers can be found in Figure 1 as profiles 4–7 (dashed and dotted lines), with standardized scores for each profile in Table 5. Similar to the mother profiles, we found evidence for an Adaptive Parenting Experience (33.3%; profile 4 in Figure 1) and an Average Parenting Experience profile (37.7%; profile 5 in Figure 1). Two additional profiles were identified for fathers, both characterized by high levels of stress but were differentiated by their dominant parenting style. In addition to high levels of stress, fathers in the Distressed – Permissive Parenting Experience profile (15.3%; profile 6 in Figure 1) also reported low self-reported parenting competence and permissive parenting as a primary parenting style. In contrast, fathers in the Distressed – Authoritarian Parenting Experience profile (13.6%; profile 7 in Figure 1) endorsed the highest levels of authoritarian parenting compared to all other parent profiles.

Association between Parenting Profiles and Child Mental Health & Autism Symptoms

The LPA was followed by an investigation of differences in child mental health and autism symptoms based on parenting profiles using the BCH method. Ratings of child symptoms from the respective parent (i.e., mother ratings of child symptoms and mother profiles) were used in this analysis. This method was used because children can behave differently with different parents and in part, this may be because parents are using different parenting styles and/or have different levels of stress, etc.¹ No differences were found between any profiles based on child age, child sex, intellectual disability status, parent age, or parent education. The standardized mean and standard error for each child outcome by parenting profile are presented in Table 6. The overall omnibus test revealed statistically significant differences among parenting profiles in both mothers and fathers for all four of the child outcomes measured. Follow-up pairwise contrasts showed that the three mother profiles differed significantly across child outcomes. Specifically, children of mothers in the Adaptive group were rated as having the fewest externalizing and internalizing symptoms and lower severity of social communication challenges and RRBs compared to all other groups. Children of mothers in the Distressed group exhibited the highest externalizing and internalizing symptoms and highest severity of social communication challenges compared to all other group. The Distressed and the Average profiles were not statistically different in ratings of child restricted and repetitive behaviors (p = .06).

Table 6.

Mean comparisons between parental profiles on child outcomes

	Mothers				Fathers
Child Outcome	Adaptive a	Distressed b	Average c	Child Outcome	Adaptive d	Average e	Permissive f	Authoritarian g
Externalizing Sx χ2(2) = 29.967***	−.471b, c (.126)	.799a, c (.215)	.146a,b (.112)	Externalizing Sx χ2(3) = 36.120***	−.464f, g (.145)	−.113f, g (.129)	.670d, e (.220)	.669d, e (.179)
Internalizing Sx χ2(2)= 17.677***	−.318b (.139)	.704a, c (.204)	.034b (.112)	Internalizing Sx χ2(3) = 32.981***	−.486e, f, g (.162)	−.014d, f, g* (.130)	.761d,e (.169)	.375d, e* (.169)
Social Communi. χ2(2)= 27.774***	−.390b, c (.146)	.643a, c (.135)	.119a, b (.115)	Social Communi. χ2(3) = 34.477***	−.472e*, f, g (.166)	−.073d*, f, g (.120)	.770d, e (.215)	.483d, e (.130)
Restrict & Rep Bs χ2(2)= 21.124***	−.389b, c (.137)	.571a, c* (.166)	.147a, b* (.123)	Restrict & Rep Bs χ2(3) = 34.563***	−.387f, g (.163)	-.155f, g (.131)	.816d, e (.195)	.450d, e (.133)

Note. Estimated means from BCH method with auxiliary variable are presented; Standard errors are in parentheses. Means that differ significantly (p<.05) between parental profiles are identified by different superscripts;

^*p<.001; Sx = symptoms; Communi. = Communication; Rep Bs = Repetitive Behaviors. No differences were found between any mother or father profiles based on child age, sex, intellectual disability status, parent age, or parent education.

Pairwise contrasts also showed a similar pattern of differences in the father profiles. Children of fathers in the Adaptive and the Average profiles were rated as having fewer externalizing and internalizing symptoms and lower severity of social communication challenges and RRBs compared to both Distressed profiles; however, the only statistically significant difference between the Adaptive and the Average father profiles was seen in child internalizing symptoms (see Table 5; differences in social communication challenges showed a similar, nonsignificant pattern, p = .07). No differences were found when comparing the Distressed – Permissive and the Distressed – Authoritarian father profiles.

Exploratory Within-family Comparison of Adaptive Profiles

Given the similarities in the mother and father Adaptive profiles, we wanted to explore whether differences (or similarities) in the parenting experience within families had implications for child functioning. In other words, we wanted to know if there was a difference in child functioning in families with both parents classified as Adaptive, with only one parent in the Adaptive profile, or if neither parent fell into that category. This post-hoc, exploratory analysis was conducted using a multivariate analysis of variance (MANOVA) to test for mean differences in the four child outcome measures based on a within-family similarity code. The data were dummy coded to represent within-family similarities in profile membership [1 = ‘both adaptive’ (20.1% of dyads), 2 = ‘mother adaptive, father other’ (21.2% of dyads), 3 = ‘father adaptive, mother other’ (13.6% of dyads), 4 = ‘neither parent adaptive’ (45.1% of dyads)]. A Bonferroni correction was used to account for multiple post-hoc comparisons and to minimize Type-I error.

Results of the MANOVA (Figure 2) suggest a significant difference between groups (Wilks’Λ = .807, F (12, 175) = 3.264, p < .001, Cohen’s f = .24). Multiple follow-up comparisons using a Bonferroni correction revealed that children with neither parent in the Adaptive profile had higher scores on all four outcomes (externalizing symptoms, internalizing symptoms, level of social-communication impairment, and severity of restricted and repetitive behaviors) compared to children with both parents in the Adaptive profile (all p’s < .01). In addition, the ‘neither’ category reported higher externalizing (M_difference = .613, p = .006) and internalizing symptoms (M_difference = .535, p = .03) compared to the ‘mom only’ category. Children in families with both parents in the Adaptive profile had lower externalizing symptoms compared to the ‘dad only’ group (M_difference = −.656, p = .04) but not the ‘mom only’ group (M_difference = −.341, p = .70). No other individual comparisons were statistically significant.

Figure 2.

Mean differences in child outcomes based on whether both parents were most likely to fall into the adaptive profile (Both), the mother was categorized as adaptive, but the father was not (Mom), the father was categorized as adaptive resilient, but the mother was not (Dad), or neither parent in the same family part of the adaptive parenting profile. Sx = symptoms; Bx = behaviors. Scores were significantly higher in the ‘neither’ category (* p < .05) across all child outcomes compared to all other groups. There was not a statistically significant differences in child outcomes between both, mom, or dad categories.

Discussion

Our aim was to provide a multidimensional description of the varied parenting experience in the context of autism by using a person-centered analytic approach to identify profiles of the parenting experience of mothers and fathers of an autistic child. Results indicated three profiles in mothers and four profiles in fathers with varying levels of parenting stress, self-reported parenting competence, and parenting styles. A prominent subgroup of both mothers (41%) and fathers (33%) reported lower than sample average parenting stress, high levels of parenting competence, and used primarily authoritative parenting techniques (i.e., the Adaptive Parenting profiles). In addition, another subgroup was identified (42% of mothers and 38% of fathers) who reported using around mean levels of all three parenting styles, average levels of stress, and average levels of parenting competence. Finally, a subgroup of parents was also labeled as “distressed” (17% of mothers) in that they reported experiencing high levels of stress, low levels of parenting competence, and primarily using authoritarian (Distressed – Authoritarian Parenting; 14% fathers) or permissive parenting (Distressed – Permissive Parenting; 15% fathers) techniques. Following the identification of classes, subgroups were then compared across a range of child outcome measures, indicating differences in child autism symptoms, behavioral, and emotional challenges, particularly between the Distressed profiles and the Adaptive profiles. Overall, these correlational findings add to the growing body of literature highlighting the important role parents have in the lives of autistic children.

We found much overlap amongst mother and father profiles. The majority of parents were classified as Adaptive or Average (83.3% of mothers and 71% of fathers) and these profiles were characterized by similar variable patterns across, suggesting that what constitutes a positive, or even average parenting experience as measured via parenting style, stress, and competence, is similar in mothers and fathers. Differences in the parenting experience were most readily found in the Distressed profiles. Previous research has generally found mothers to report more psychological distress than fathers (e.g., Jones et al., 2013), although not universally (e.g., Ozturk, Riccadonna, & Venuti, 2014); however, in the current sample, almost twice as many fathers (n = 53) were characterized as distressed compared to mothers (n =31). This may suggest that focusing on sample means has been “hiding” significant variability in parenting distress for fathers in particular. Furthermore, the best fitting model for fathers included two Distressed profiles. Fathers in both groups reported high levels of stress and low competence but differing parenting strategies, although there was not a statically significant difference between the two Distressed father profiles on any child variables. Thus, children with Distressed fathers tended to demonstrate more behavior challenges and autism characteristics regardless of which Distressed profile their father fit into (i.e., Distressed – Permissive or Distressed – Authoritarian).

Findings from this study suggest that the parenting experience in the context of autism is not monolithically characterized by stress. Similar to other profile-based research on parenting in autism (e.g., Schwartzman, Hardan, & Gengoux, 2021), we identified a subgroup of parents (33% of fathers and 41% of mothers) who reported low levels of stress and who endorsed high levels of authoritative parenting (i.e., indicating that they are “responsive to child’s feelings or needs”, “encourages the child to talk about their troubles”, “explains the consequences of the child’s behavior”, etc.) and high subjective feelings of parent competency. While this study represents a snapshot of the parenting experience and should be replicated, results add to a growing body of work highlighting the positive contribution and important role parents play in the lives of autistic children.

This study contributes to a more nuanced picture of the parenting experience in the context of autism. Results indicate that parenting an autistic child includes a diversity of experiences for both mothers and fathers, suggesting that not all parents need intervention services. A strength of identifying subgroups of parents is that we can use this information to maximize resources and target parents who most need services. If similar profiles or subgroups of the parenting experience were found in a larger, more diverse sample, the question then becomes how to more accurately identify parents in the Distressed subgroups and devote resources to those particular families. For instance, preliminary evidence of a brief parenting program implemented in a primary care setting suggests short-term (6-month) improvement in parenting styles, parenting confidence, and a significant reduction in parenting stress (Tellegen & Sanders, 2014). It may be that parents who would benefit most from these types of programs are those who would be classified in the Distressed profiles; thus, screening for a combination of high parenting stress, low parenting competence, and ineffective parenting strategies (i.e., permissive and authoritarian) during routine child check-ups may offer a promising way for clinicians to provide more targeted support services.

Our findings have implications not just for parents but for autistic children. Children of parents identified as Distressed had the highest autism symptom severity scores and the most parent-reported behavior and emotional challenges. It could be that these parents are “distressed” because they are parenting children with more severe autism presentations (e.g., behavior problems, inattention, more severe restricted and repetitive behaviors, etc.) and the cross-sectional nature of the current study precludes any suggestion of a casual relationship. There may also be additional influences (e.g., contextual variables such as socioeconomic conditions, educational challenges, social support) not measured in the current study that contribute to both parent ‘dis’stress as well as child behavioral profiles. We are also not suggesting that parents or parenting causes autistic children’s challenges; however, we know that parents and parenting make a difference in children’s lives and when parents are not doing well, ultimately it is more difficult for children to do well, including and maybe particularly for autistic children. For example, evidence suggests that parenting stress can reduce the effectiveness of early intervention for autism (Osborne et al., 2008). Previous research in children with neurodevelopmental disabilities, including autistic children, has reported a beneficial effect of positive parenting, primarily characterized by sensitive, responsive, and directive (i.e., authoritative) parenting strategies, on child outcomes (Dyches et al., 2012). Regardless of the “causal chain” in the association between the parenting experience and child characteristics, identifying those parents in need of extra support may confer a downstream benefit for children as well. We can use our understanding of the reciprocal nature of parent-child relationships to our advantage – if we can help one (i.e., parents), theoretically we can help the other (i.e., children), which may, in turn, help further improve the parenting experience.

Within families, our results indicate that (1) it is not uncommon for mothers and fathers to have different parenting experiences and (2) child behavioral symptoms are correlated with having at least one parent doing well. In other words, children with two parents in the Adaptive group and children with one parent in the Adaptive group did not differ statistically in their autism symptom severity, behavioral, or emotional challenges. There was also no significant difference between which parent (mother or father) was in the Adaptive group except in the case of child externalizing symptoms. This can be an empowering message for parents, particularly in families in which one parent may have better strategies for stress-management, parenting skills, or confidence. Similarly, this finding may be of particular importance for single parents, and these patterns should be investigated in single-parent families. Results also provide support for the growing body of literature on fathers of autistic children and suggest that additional research examining within family experiences in the context of autism is warranted.

This study provides meaningful information about the parenting experience in the context of autism, but it is important to acknowledge study limitations. One drawback to person-centered analysis is that results may be specific to the analytic sample, therefore limiting generalizability to other groups. While the current sample represented the sociodemographics of the area from which it was taken, issues of generalizability to the broader population (i.e., racial and socioeconomic diversity) of families with autistic children remain. The profiles of the parenting experience identified in the current study should be replicated in a larger, more diverse sample. A second limitation of person-centered approaches is the subjectivity associated with choosing the final number of profiles. While we followed best analytic practices by using multiple fit indices, considering small sample cut-offs as indicators of error in the model, and visually inspecting all profile solutions, there is still uncertainty about the best solution. It is possible that additional classes (e.g., a fourth class for mothers) could be found in replication studies, again suggesting the need to confirm these findings in larger, more diverse samples. In addition, associations presented here are based on cross-sectional evidence, precluding any claims of directionality or causality. It is likely that the associations between the parenting experience and child variables are bidirectional and future longitudinal research should address this issue. Future research should also consider a multi-reporter and/or a multi-method approach to address issues such as rater bias (e.g., mothers reporting on their own and their child’s symptoms), self-report bias (e.g., using alternatives to questionnaires when measuring parenting styles). This study contributes to the literature by examining three aspects of the parenting experience; however, parenting stress, competence, and styles are not the only indicators of the parenting experience and future research should consider additional variables that may impact parenting both within the parent (e.g., mental health) and within the environment (e.g., relationship satisfaction). It is also worth noting that this study did not include a control group of parents without an autistic child or parents of children with other neurodevelopmental disabilities. Nevertheless, the current study provides evidence for a diverse set of parenting experiences for both mothers and fathers in the context of autism. Understanding the characteristics of different parenting experiences as well as how those experiences relate to child characteristics provides insight for intervention development as well as for whom intervention may be helpful.