The Centers for Disease Control and Prevention (CDC) estimate the prevalence of autism at one in 36 children in the United States (Maenner, 2023). In addition to variation in the degree and presentation of core autism features, including difficulties with social communication and the presence of restricted interests and/or repetitive behaviors, autistic children are at increased risk for experiencing high rates of co-occurring externalizing behavior problems, including aggression and non-compliance (Brookman-Frazee et al., 2018). Externalizing behavior problems often confer elevated levels of stress within the family system (Rodriguez et al., 2019). Indeed, within this population, empirical work has found bidirectional associations between behavior problems and corresponding parenting stress levels, suggesting a transactional relationship wherein behavior problems and parenting stress exacerbate each other over time, which poses cascading risks to child and family functioning (Lin et al., 2021; Rodriguez et al., 2019). Despite these well-established findings, no study, to our knowledge, has explored how these associations may differ for Latino families of autistic children, who comprise the fastest growing ethnic minority subpopulation of autistic children in the United States, but remain empirically underexamined (Nevison & Zahorodny, 2019). Compared to non-Latino counterparts, Latino families face greater diagnostic disparities and barriers to service acquisition, and in turn, may differ from non-Latino peers in their reports of parenting stress and behavior problems (Rivera-Figueroa et al., 2022). Moreover, few studies have leveraged daily diary methodology to examine parenting stress and child behavior problem associations at a within-person level, thus limiting the field’s understanding of such dynamics.

Parenting Stress and Child Behavior Problems among Families of Autistic Children

Externalizing behavior problems, such as aggression and non-compliance, are more common among autistic children compared to neurotypical counterparts (Quetsch et al., 2023). Co-occurring psychiatric disorders (e.g., attention-deficit/hyperactivity disorder, oppositional defiant disorder, and conduct disorder) are also highly prevalent for autistic children; one recent study found that nearly all school-age autistic children in the sample met diagnostic criteria for at least one co-occurring psychiatric disorder, with an average of more than two co-occurring diagnoses per child (Brookman-Frazee et al., 2018). The presence of these co-occurring mental health needs further elevates the risk and severity of behavioral problems, with the potential to impede long-term child functioning and parental well-being (Fitzgerald et al., 2023). Indeed, a growing body of literature has highlighted child externalizing behavior problems as a primary predictor of parenting stress, over and above other parental stressors (e.g., differences in social interaction/communication and adaptive behaviors; Olson et al., 2022), making the relationship between child externalizing behavior and parenting stress a critical area of focus.

Parents of autistic children tend to report higher levels of parenting stress compared to parents of allistic peers (Barroso et al., 2018), which has adverse implications for parents’ mental, physical, and relational health, and may ultimately compromise parenting behaviors (Hickey et al., 2020; Shawler & Sullivan, 2017; Stephenson et al., 2022). Carefully controlled longitudinal designs provide compelling evidence of transactional associations between child behavior problems and parenting stress (e.g., Lin et al., 2021; Rodriguez et al., 2019; Zaidman-Zait et al., 2014), with evidence for both child- and parent-driven processes across a range of child ages and developmental stages. Relatively few studies have specifically focused on families of preschool-aged autistic children, despite the developmental significance of this period for early intervention. Moreover, these transactional associations have rarely been examined at a daily level. The use of daily report (e.g., daily diary) methodology permits examination of within-person variability (e.g., how a change in functioning from a person’s own average impacts immediate outcomes) and offers distinct advantages, including more accurate assessments of momentary and variable experiences, minimized recall bias, and greater ecological validity compared to laboratory-based methods, affording a finer-grained understanding of these temporal processes (Bamberger, 2016; Horstmann, 2021).

The Experience of Latino Families of Autistic Children

Very few studies of parenting stress and child behavior problems have specifically focused on the experience of Latino families of autistic children. As the number of Latino autistic children in the United States continues to rise (Gallin et al., 2024; Maenner, 2023), health inequities among Latino families of autistic children present an increasing concern. These inequities manifest in various ways, including disparities in healthcare service use (Walensky et al., 2021), greater unmet service needs (e.g., lack of access to a personal physician; Smith et al., 2020), and challenges in obtaining referrals for diagnostic and intervention services (Luelmo et al., 2022). Contributing factors include insufficient autism-related information for Latino families (Zuckerman et al., 2017), language barriers for Spanish-speaking households (Smith et al., 2020), and the lower quality of care often reported by Latino families compared to non-Latino counterparts (Shaw et al., 2022). As a result, Latino children are typically diagnosed with autism an average of 2.5 years later than their non-Latino peers, and often present with higher autism symptom levels and/or co-occurring intellectual disability by the time of diagnosis (Christensen et al., 2018; Gallin et al., 2024; Hickey et al., 2021). These heightened features are thought to stem from delayed diagnosis, which impedes timely delivery of early intervention and may affect long-term outcomes (Zwaigenbaum et al., 2015). Given the growing evidence of health disparities and associated stressors faced by Latino families of autistic children, understanding daily family dynamics specifically among Latino families is crucial.

Latino families of autistic children are highly diverse, with considerable within-group variation influenced by multiple intersecting sociodemographic and cultural factors (Rivera-Figueroa et al., 2022). In research on Latino populations, language use in the home is a common proxy for acculturation—the process of adjusting cultural beliefs, practices, values, and behaviors from a person’s culture of origin to those of a different culture (Scholaske et al., 2021; Schumann et al., 2020). In the current sample, Latinos were primarily English-speaking, with the vast majority reporting English as the primary language spoken at home and as their child’s first language. English use at home, in particular, generally correlates with higher levels of acculturation and is often linked to being U.S.-born, having longer residence in the United States, younger age at immigration, and U.S.-based education (Scholaske et al., 2021). Theorists suggest that that English-speaking Latino parents of autistic children may experience greater stress relative to Spanish-speaking counterparts due to increased access to their child’s health information and a deeper understanding of the implications of their child’s diagnosis (Mercado et al., 2021). Interviews with English-speaking and bilingual Latino caregivers of autistic children have also illuminated additional challenges faced by this Latino subpopulation, such as dealing with stigma, isolation from family and community, and the lack of culturally appropriate information for Latino families (Blanche et al., 2015). These factors may be associated with elevated frequency and severity of co-occurring child behavior problems (e.g., as a function of delayed autism diagnosis and related intervention; Zwaigenbaum et al., 2015), and elevated levels of parenting stress (Iadarola et al., 2019). Given the additional challenges faced by Latino families of autistic children—particularly English-speaking Latinos—the effect of Latino ethnicity on the transactional relationship between child behavior problems and parenting stress warrants research attention. We posit that these relationships may be more pronounced for Latino families due to the added burdens that confer risk to both caregiver mental health and child behavioral outcomes. Such insight may help identify population-specific intervention strategies aimed at optimizing parent and child functioning and attenuating the impact of health inequities on minoritized families within the autism community.

Daily Diary Approaches to Examining Within-Person Family Processes

Existing longitudinal research on the relationship between child behavior problems and parenting stress has primarily examined lagged effects across months or years. While these approaches help in identifying broader temporal trends, there is an increasing appreciation for daily diary and other intensive longitudinal methods for advancing the study of transactional family dynamics (Bamberger, 2016). By using a within-person, daily-level design, questions can be answered about how a change in functioning from a person’s own average (e.g., being more stressed than usual) impacts immediate outcomes (e.g., next-day reports of child behavior). This allows for the estimation of rapid changes within family processes and offers a more robust basis for causal inferences than between-person analyses, which answer questions about cross-sectional or group effects (e.g., how a person’s functioning, relative to the group average, predicts next-day outcomes) and often do not adequately reflect the dynamic nature of developmental processes (Hamaker et al., 2015; Berry & Willoughby, 2017). Although the bulk of existing developmental literature has drawn conclusions from between-person data, methodologists contend that within-person level data are preferable when attempting to draw causal inferences and avoid biased model effects (Hamaker et al., 2015). Utilizing a within-person daily diary approach, one preliminary exploration involving a subgroup (n = 26 participants) from the current sample found associations between shorter child sleep, more child behavior problems, and greater parenting stress at the between-person level but found that such relations differed at the within-person level (Iwamoto et al., 2023). Using a larger, multi-ethnic sample, the present study uses the same methodology to examine between- and within-person transactional associations between daily parenting stress and reported child behavior problem severity, and expands on previous preliminary findings by exploring the moderating role of Latino ethnicity on such relations.

The Current Study

Longitudinal research on child behavior problems and parenting stress among Latino families of young autistic children is scarce, and very few studies have examined such constructs at a daily within-person level. Hence, the current study sought to attain a finer-grained understanding of the associations between child behavior problems and parenting stress using a within-person, daily-diary design among Latino and non-Latino parents of young autistic children. This study had two aims: (1) to bidirectionally examine the relationship between child behavior problems and parenting stress at a daily level among families of autistic children ages 3–5 years; and (2) to examine whether Latino ethnicity moderates the relationship between child behavior problems and parenting stress. In investigating these relations, we hypothesized parallel processes for between- and within-person effects. Specifically, we expected that child behavior problems would positively predict next-day parenting stress. We similarly anticipated that greater parenting stress would predict increased next-day child behavior problems. Lastly, we hypothesized that Latino ethnicity would moderate these associations at the between-person level, such that predictive effects would be stronger for families identifying as Latino (versus non-Latino).

Method

Participants

Participants (N = 70) were recruited from a larger randomized controlled trial (RCT) examining the efficacy of Mindfulness-Based Stress Reduction (MBSR) compared to Psychoeducational Support (PE) in reducing parenting stress in families of preschool-aged autistic children (Fenning et al., 2023; Neece et al., 2024). These families were primarily recruited from September 2020 to March 2021 through the Inland Regional Center, a government agency that contracts services for individuals with developmental disabilities. Additional recruitment was done through participating universities, community groups, local agencies that provide services for autistic children, local preschools, and community events for families of children with disabilities.

The RCT included 117 parent-child dyads. Inclusionary criteria for the RCT were: (a) child community diagnosis of autism verified by study-administered assessments; (b) child age 3 to 5 years at enrollment; and (c) parent ability to complete study procedures in English. Exclusionary criteria included: (a) positive screen for active parental psychosis, substance abuse, or suicidality according to the associated modules of the Structured Clinical Interview for DSM Disorders, Research Version Non-Patient Edition (First, et al., 2002); (b) parent participation in an auxiliary mental health treatment or support group, given potential confounds with the interventions in the larger clinical trial; and (c) child motor impairment that would prevent participation in laboratory assessments (e.g., difficulty sitting independently).

Of the 117 parent-child dyads who participated in the RCT, 70 were eligible and enrolled in the current daily diary sub-study, which was completed during the baseline timepoint of the larger study, prior to any intervention delivery. Inclusionary criteria for the sub-study were: (a) primary caregiver eligibility and enrollment in the broader study, (b) primary caregiver willingness to participate in the daily diary sub-study, (c) primary caregiver currently in a long-term relationship (i.e., dating or married for at least 6 months), and (d) primary caregiver currently living with their partner or spouse. The sub-study only included primary caregivers that were in a long-term relationship and living with their partner or spouse, as the daily diary study was designed to examine additional processes related to romantic partnerships. Table 1 presents current daily diary sub-study demographics.

Procedure

Procedures for the current study were approved by the Institutional Review Board at participating universities. Interested parents completed a phone screening to determine initial eligibility related to demographics and parent mental health. During screenings, families identified one caregiver as the “primary caregiver,” who was then the primary participant in the study. Families who met preliminary eligibility were scheduled for a baseline pre-randomization laboratory assessment where study staff obtained informed consent, gathered demographic and service utilization information, and conducted structured clinical interviews with parents and a battery of standardized psychological assessments to evaluate the child’s autism symptoms (Autism Diagnostic Observation Schedule, 2nd Edition; Lord et al., 2012), intellectual functioning (Stanford-Binet Intelligence Scales, 5th Edition Abbreviated Battery IQ; Roid, 2003), receptive vocabulary (Peabody Picture Vocabulary Test, 4th Edition; Dunn & Dunn, 2007), and adaptive behavior skills (Vineland-3 Comprehensive Interview Form; Sparrow et al., 2016). Parents also completed a packet of questionnaires, including measures of child behavior problems, child developmental functioning, parent and child mental health, and family functioning.

At the end of the baseline assessment, parents were screened to participate in the elective daily diary sub-study. Parents enrolled in the daily diary sub-study were asked to answer questions daily for 14 typical days (i.e., not during vacations or holidays). Parents were instructed to complete the surveys around the same time each day and respond to questions in terms of the previous 24 h. Participants were compensated for their participation contingent on the number of daily diary entries completed, with a bonus amount offered to incentivize consecutive diary entries. Participants were compensated $40 for completing at least 10 consecutive days of diary entries, with an added $10 bonus if they completed all 14 days. If participants completed fewer than 10 consecutive days of diary entries, they were allowed to complete additional days to reach a minimum of 10 consecutive days.

Minor adaptations to the procedures were made for participants (n = 26) who joined the study after the onset of COVID-19. Due to social distancing requirements, for this cohort of individuals, baseline pre-randomization assessments were completed via phone and Zoom, and materials were delivered to the family’s home. Standardized assessments of the child’s autism symptoms (Autism Diagnostic Interview-Revised; Rutter et al., 2003) and adaptive behavior (Vineland-3) were completed via parental interview on Zoom. Standardized direct assessments of the child (SB-5 ABIQ and PPVT-4) were postponed until post-intervention (September 2021), when we received institutional permission to resume in-person activities. This subgroup of participants did not demographically differ from participants who joined the study prior to the onset of COVID-19, nor did they differ on key study variables (i.e., parenting stress, behavior problems), per ANOVA and independent samples t tests.

Measures

Demographic Information

All primary caregivers completed a demographic interview with study personnel to obtain information about parent and child characteristics including child and parent ages, ethnicities, child diagnoses, family income, education level, language background, and services received.

Parenting Stress

Parents used a 7-point scale (1 “not stressful” to 7 “extremely stressful”) in response to the item “Overall how stressful were your parenting experiences with your child?”. Higher scores were indicative of more parenting stress. In previous research involving parents of autistic children (Hartley et al., 2018), this item has been significantly correlated with other global measures of parenting stress, including scores on the Burden Interview (Zarit et al., 1980), r = 0.66, p < 0.01. The current daily diary study sample provides additional validity evidence for the single-item stress measure through significant positive correlations with three established and widely used measures of parenting stress: the Parenting Stress Index Fourth Edition-Short Form Parental Distress score (Abidin et al., 2006), r = 0.61, p = 0.002, the Parenting Daily Hassles Total Intensity score (Crnic & Greenberg, 1990), r = 0.28, p = 0.020, and the Family Impact Questionnaire Negative Impact score (Donenberg & Baker, 1993), r = 0.24, p = 0.045.

Behavior Problems

The Scales of Independent Behavior (SIB-R; Bruininks et al., 1996) was used to examine the occurrence and severity of eight different types of problem behaviors: hurtful to self, unusual or repetitive habits, hurtful to others, socially offensive behavior, destructive to property, withdrawal or inattentive behavior, disruptive behavior, and uncooperative behavior. These items have good reliability and validity and have strong correlations with externally validated measures of the same constructs (Hartley et al., 2018). The current analyses utilized the severity subscale, which demonstrated adequate reliability in the present sample (α = 0.74). Higher scores were indicative of greater child behavior problems. In the current daily diary study sample, severity scores were positively correlated with Child Behavior Checklist (CBCL; Achenbach, 2000) Total Problems t-scores (r = 0.52, p = 0.011).

Data Analytic Plan

Prior to running the main analyses, we computed descriptive statistics for each study variable (Table 1). We also ran independent samples t tests to identify Latino/non-Latino group differences in each variable (Table 1). We then computed bivariate correlations between key study variables and potential demographic covariates (Table 2). Demographic variables were included as covariates in the models if they were significantly associated with initial (i.e., day 1) levels of both parenting stress and child behavior problem severity.

Parents in the current study completed an average of 13.03 (SD = 2.62) days of the total 14 days of the daily diary study. There were no demographic differences between participants who completed all 14 days of the daily diary study (n = 59) and those who did not (n = 11; ps > 0.05). To assess whether missed items on parenting stress and child behavior problem measures were related to demographic features, we computed a composite variable that summed each participant’s total number of missing parenting stress items across all daily diary study days and a subsequent composite variable for participants’ summed number of missing behavior problem severity items. We then ran bivariate correlations between these composite variables, demographic variables, and primary study variables. Bivariate correlations indicated that missingness was not significantly related to any demographic characteristics—child age, gender, IQ, adaptive behavior, intellectual disability status, family income, ethnicity—nor any of the primary outcomes (ps > 0.05). Our main analyses relating parenting stress and child behavior problem severity were estimated using full maximum likelihood estimation (Singer & Willett, 2003), an approach that fits the model using the incomplete data in a way that does not bias estimates and produces better estimates than other missing data estimation approaches (e.g., multiple imputation, listwise deletion; Little & Rubin, 2002). In longitudinal multilevel analysis, missing data are not treated as missing; instead, estimates are based on all available data (e.g., one data point to the next). Although there are approaches in development to handle missing data within nested data structures, there is not yet consensus as to an ideal solution (see Hox et al., 2015, for a review).

To address the first aim regarding the relationship between parenting stress and child behavior, we ran bivariate correlational analyses to confirm significant relationships between measures as hypothesized. Individual fluctuation in parenting stress and child behavioral variables across the 14 data points was then estimated with multilevel models using full maximum likelihood estimation in SPSS to account for missingness and the within-person nested structure of daily diary data. Unconditional means models were tested to examine average daily levels of parenting stress and average daily severity of behavior problems reported by parents. We then conducted a series of time-lagged MLM models, using lagged day variables, with parenting stress predicting next-day severity of child behavior problems, and with severity of child behavior problems predicting next-day level of parenting stress. In each model, we controlled for the autoregressive effect of the previous-day within-person measurement of the outcome variable of interest. To capture changes in these relationships both within- and between-individuals, in each model the predictor of interest was partitioned into its between- and within-person components.

To address the second aim regarding the potential moderating role of ethnicity in the relationship between parenting stress and child behavior problems, we ran these models again, adding ethnicity as well as the interaction of ethnicity with the predictor variable of interest within each model. Simple slopes and post-hoc subgroup regression analyses were conducted to interpret significant moderation. In all models, Level 1 predictors were within-person centered, and Level 2 continuous variables were grand-mean centered.

Results

Parents reported a weighted average daily parenting stress score of 1.81 (SD = 1.16), with non-Latino parents rating their daily parenting stress (M = 1.99, SD = 1.29) at a significantly higher level on average than Latino parents (M = 1.59, SD = 0.92), t(910) = 5.56, p < 0.001, d = 0.37. Across days, parents reported a weighted average child behavior severity score of 7.82 (SD = 5.20), but there was no significant difference in the severity of behaviors as reported by Latino (M = 7.75, SD = 4.58) as compared to non-Latino parents (M = 7.87, SD = 5.65), t(910) = 0.34, p > 0.05. Demographic and clinical characteristics of Latino and non-Latino participants are shown in Table 1. Of note, Latino and non-Latino participants differed in terms of their child’s adaptive functioning, such that Latino parents reported lower Vineland Adaptive Behavior Scales Third Edition (Vineland-3; Sparrow et al., 2016), Adaptive Behavior Composite (ABC) Scores than did non-Latino parents, with all participants scoring in the moderately low to low range (i.e., below a score of 85), indicating significant deficits in adaptive functioning.

Bivariate correlations between major study variables and demographic characteristics are shown in Table 2. Only ethnicity (Latino or non-Latino) was significantly related to the initial status of both dependent variables, such that it was negatively related to initial parenting stress and positively related to initial child behavior problem severity, ps < 0.05. Of note, autism symptom level was positively associated with initial child behavior problem severity but not with initial parenting stress. Similarly, parent education was positively related to initial levels of parenting stress but not with initial child behavior problem severity. Family income, child intellectual disability status, and child adaptive functioning were not significantly associated with either outcome variable.

An unconditional means model indicated that, on average, parents reported daily parenting stress scores of 1.81 (SE = 0.14) and daily behavior problem severity sum scores of 7.74 (SE = 0.61). Intraclass correlation coefficients for the unconditional model indicated that 45.8% of the variance in parenting stress and 55.0% of the variance in behavior problem severity wass estimated to be between participants.

Aim 1: Effect of Child Behavior Problem Severity on Parenting Stress

Table 3 presents the time-lagged models in which child behavior severity was used to predict next-day parenting stress, controlling for the autoregressive effect of previous-day within-person centered parenting stress (Aim 1). On average, parents’ within-person centered level of parenting stress was only marginally predictive of their next day parenting stress, p = 0.081, suggesting notable day-to-day variation.

Between-person effects (across the group)

Controlling for within-parent centered ratings of previous-day parenting stress, parents who reported greater than average child behavior problems also reported significantly greater levels of next-day parenting stress, p < 0.001.

Within-person effects

Parent-reported greater than usual child behavior problem severity was not significantly associated with fluctuations in reported next-day parenting stress, p = 0.230.

Aim 1: Effect of Parenting Stress on Child Behavior Problem Severity

Next, we ran a time-lagged model in which parenting stress was used to predict next-day level of child behavior severity, controlling for the autoregressive effect of previous-day within-person centered child behavior severity (Table 3). We found stability effects for child behavior severity wherein child behavior severity significantly predicted next-day behavior severity, p < 0.001.

Between-person effects (across the group)

Controlling for within-parent centered ratings of child behavior severity, parents who reported experiencing greater parenting stress on average also reported experiencing significantly greater next-day child behavior severity, p < 0.001.

Within-person effects

Unexpectedly, we found the opposite effect when parents reported experiencing greater than usual parenting stress. Specifically, higher than usual ratings of parenting stress predicted a significant next-day decrease in reported child behavior problem severity, p = 0.012.

Aim 2: Effect of Parent Ethnicity on the Effect of Child Behavior Problem Severity on Parenting Stress

We then ran the same models, adding in the effects of ethnicity (Latino/Non-Latino) and the interaction of ethnicity with the relevant predictor variable (Table 3). In the first model, we ran the time-lagged models in which ethnicity, child behavior severity and the interaction of these were used to predict parenting stress, controlling for the autoregressive effect of previous-day parenting stress (Table 3). As in the previous iterations, on average, parents’ within-person centered level of parenting stress was only marginally predictive of their next day parenting stress, p = 0.083; in this model, ethnicity itself did not have a significant main effect on parenting stress, p = 0.379. The between- and within-person analyses yielded similar results to the previous models.

Between-person interaction effects (across the group)

The interaction of ethnicity and child behavior severity had a significant effect on next-day parenting stress at the between-person level. Simple slopes analyses found significant positive associations between child behavior problem severity and next-day parenting stress for both groups, ßLatino = 0.17, ßNon-Latino = 0.19, ps < 0.05. The slope for Latino families was significantly less steep compared to the slope for non-Latino families, t = −2.48, df = 60, p = 0.016.

Within-person interaction effects

When parents reported greater severity of their child’s behavior than usual, Latino and non-Latino parents reported similar fluctuations in their next-day parenting stress, p = 0.505.

Aim 2: Effect of Parent Ethnicity on the Effect of Parenting Stress on Child Behavior Problem Severity

In the next model, parenting stress, ethnicity, and the interaction of ethnicity and parenting stress were used to predict next-day level of child behavior problem severity, controlling for the autoregressive effect of previous-day within-person centered child behavior problem severity. Again, we found stability effects for child behavior problems wherein reported child behavior severity significantly predicted reported next-day behavior severity, p < 0.001. Ethnicity was marginally predictive of child behavior problems, such that parents identifying as Latino reported somewhat greater severity of child behavior problems on average across days, p = 0.079. The between- and within-person analyses yielded similar results to the previous models (see Table 3).

Between-person interaction effects (across the group)

The interaction of ethnicity and parenting stress had a marginal effect on next-day child behavior severity such that parenting stress and next-day child behavior problem severity were positively associated, and this link was slightly weaker among Latino (versus non-Latino) families, p = 0.081. As ethnicity was not a statistically significant moderator in this model, the main effect of parenting stress stands; parenting stress positively predicted next-day child behavior problem severity, p < 0.001.

Within-person interaction effects

When parents reported greater parenting stress than usual, Latino and non-Latino parents reported similar fluctuations in their child’s behavior severity, p = 0.985.

Discussion

This study bidirectionally examined day-to-day temporal relations between child behavior problems and parenting stress in ethnically diverse families of autistic children. Child behavior problems and parenting stress were positively and bidirectionally associated at the between-person level. The bidirectional relationship between child behavior problems and parenting stress was moderated by Latino ethnicity, such that, the positive association between reported child behavior problems and next-day parenting stress was slightly weaker for Latino than for non-Latino families. The reverse effect (i.e., parenting stress positively predicting next-day child behavior problems) was also less pronounced among Latino families, although not significantly so. At the within-person level, parenting stress was associated with fluctuations in next-day child behavior severity, but in the opposite direction than expected. Specifically, higher than usual ratings of parenting stress, relative to an individual’s own average parenting stress, predicted a significant decrease in reported next-day child behavior problem severity. Reported severity of child behavior problems was not significantly associated with fluctuations in next-day parenting stress at the within-person level.

Child Behavior Problems and Parenting Stress: Between-Person Effects and Latino/non-Latino Differences

At the between-person level, where predictor levels reflect an individual’s daily functioning relative to the group average across all time points, evidence of a daily positive bidirectional relationship between child behavior problems and parenting stress was found. In particular, parents of autistic children who reported a greater severity of behavior problems, compared to the group average across all time points, also tended to have higher next-day levels of parenting stress. Conversely, parents with greater levels of parenting stress, compared to the group average across all time points, also tended to report greater next-day severity of child behavior problems, substantiating bidirectional trends found in previous research on families of autistic youth (e.g., Iwamoto et al., 2023; Lin et al., 2021; Rodriguez et al., 2019; Zaidman-Zait et al., 2014).

These bidirectional effects were moderated by Latino ethnicity; however, not in the way we expected. Interestingly, the association between reported child behavior problem severity and next-day parenting stress, while positive for both groups, was weaker for Latino families of autistic children. The association between parenting stress and reported next-day child behavior problem severity was also weaker among Latino (versus non-Latino) families of autistic children, although not significantly so.

While Latino ethnicity was initially hypothesized to strengthen the positive relationship between parenting stress and child behavior problems due to a number of health inequities faced by Latino families of autistic children, including the 2.5-year diagnostic disparity between Latinos and non-Latino peers, these sociodemographic factors may also help to explain the weakened links among Latino families. Health inequities faced by Latino families of autistic children, including disparities in healthcare service use (Walensky et al., 2021), unmet service needs (Smith et al., 2020), and diagnostic disparities (Gallin et al., 2024; Hickey et al., 2021; Luelmo et al., 2022), alongside cultural stigma and feelings of isolation (Blanche et al., 2015) may introduce varied sources of parenting stress, which may be further compounded for English-speaking Latinos like those in the current study (Mercado et al., 2021). As such, child behavior problems may be one of many sources of stress facing Latino parents of autistic children, and thus may emerge as a less robust direct predictor of daily parenting stress levels.

In addition, protective Latino cultural factors like familism—characterized by strong family loyalty, reciprocity and solidarity (Campos et al., 2016)—may foster resilience, diminishing the impact of daily child behavior problem severity on parenting stress among Latino families of autistic children. Indeed, recent studies have highlighted familism’s protective role among Latino caregivers of autistic children, suggesting that family cohesion may buffer negative impacts of diagnoses and attenuate adverse perceptions of child behaviors (e.g., Magaña & Smith, 2006, as cited in Lopez & Magaña, 2020). Hence, it is plausible that cultural factors may influence the association between child behavior problems and parenting stress in this population, perhaps by attenuating negative perceptions of child behaviors (e.g., Blacher et al., 2013). Examination of familism and other resilience factors as moderators of the link between child behavior problems and parenting stress is a critical area for future research.

Within-Person Effects of Daily Parenting Stress on Child Behavior Problems

At the within-person level, where predictor levels reflect an individual’s daily functioning relative to their own average functioning across all time points, parenting stress was unidirectionally, negatively related to next-day child behavior problems, replicating initial findings from Iwamoto et al.’s (2023) study examining child behavior, sleep duration, and parenting stress relations among a subgroup from the current sample. That is, surprisingly, on days when parents reported greater parenting stress than usual (i.e., relative to their own average parenting stress), next day child behavior problem severity was reported to be lower. Though seemingly paradoxical, we speculate that a number of factors may underlie the surprisingly inverse associations between parenting stress and child behavior problems. On the one hand, these findings may suggest a potential adaptive or compensatory mechanism at play within the family unit. For instance, parents of autistic children may be responding to their parenting stress and child’s behavior problems with adaptive coping skills, conferring more effective parenting strategies to manage their child’s behavior (Menezes et al., 2021) and resulting in fewer behavior problems the following day. This could indicate a form of resilience where parenting stress motivates more effective coping or intervention efforts. Indeed, recent research evidence from our larger sample suggests that parental adaptive coping, especially active planning, may buffer the association between child externalizing problems and parenting behavior (Alostaz et al., 2021). Alternatively, the inverse associations found in this study may reflect use of negative parenting behaviors in the wake of child behavior problems. Recent evidence from the our larger sample suggests spillover of parental distress to negative parenting behavior for families of young autistic preschoolers (Baker et al., 2023). Although negative parenting may appear to lead to short-term child compliance and thus parent stress reduction, it may actually portend escalating cycles of parenting stress and child behavior problems over time (Bader & Barry, 2014; Hirschler-Guttenberg et al., 2015).

Of note, in contrast to the between-person findings, there was no effect of child behavior problems on next-day parenting stress at the within-person level. That is, while greater reported child behavior problems, relative to the group average across all time points, was predictive of next-day parenting stress, greater-than-usual reported child behavior problems, relative to one’s own average, was not. This may reflect the complex, multifaceted nature of parenting stress itself within each family. Among parents of young autistic children, parenting stress is likely multiply determined by a number of stressors, all of which may contribute to further within-person variability (Ilias et al., 2018; Stanojević et al. 2017). Hence, at a within-individual level, child behavior problems may not consistently relate to daily parenting stress levels. Future studies should include measures of other daily stressors to elucidate the mechanisms underlying within-person changes in parenting stress among families of autistic children.

The nuances in the within-person findings highlight the importance of partitioning between- and within-person effects when examining dynamic family processes. In particular, the within-person findings afford a finer-grained understanding of daily within-person variability in parenting stress and child behavior problem severity and capture how dynamic behavioral and stress processes unfold at a more micro, intrapersonal level. The within-person findings also signal potential intervention targets. For instance, interventions that promote stress-reduction (e.g., Mindfulness-Based Stress Reduction; Kabat-Zinn, 2003) may benefit from understanding how child behavior problems and parenting stress vary in the moment within families of autistic children, which may help interventionists to tailor and individualize group discussions. Meanwhile, the between-person findings highlight broader trends across the sample as a whole, replicating previous research trends suggesting, on average, a positive and mutually escalating relationship between parenting stress and reported child behavior problem severity, which may offer insight into how these processes accumulate over time. Stress-reduction interventions may benefit from added psychoeducation on the long- and short-term consequences of transactional dynamics in families of autistic children. The mechanisms underlying the differing within- and between-person findings warrant further investigation.

Strengths, Limitations, and Future Directions

The present study had several strengths. First, the daily-diary design provided a finer-grained understanding of the transactional relationship between child behavior problems and parenting stress at a day-to-day level. Furthermore, this study used advanced between- and within-person methodological approaches to examine these trends in parent and child functioning, which have only been used in a few prior studies on autistic populations (e.g., Hartley et al., 2018; Iwamoto et al., 2023), affording a more comprehensive understanding of transactional dynamics in families of autistic children. Additionally, this study drew from a diverse, multi-ethnic, multicultural sample, with a particular focus on the experience of Latino parents—a historically marginalized group that is often overlooked in existing literature on autism. In focusing specifically on Latino parents of autistic children, relative to non-Latino counterparts, we are better able to identify relational processes that are group-specific in order to inform culturally congruent intervention targets.

Although promising, the results of the current study should be considered in light of the following limitations. First, parenting stress was assessed using a single-item measure. This item has been used successfully in prior research on families of autistic individuals (Hartley et al., 2018) and was significantly correlated with multiple established measures of parenting stress within the current study. Nonetheless, the single item does not fully capture the complexity of parenting stress. This item was selected due to its use and validity in prior research and the dearth of other short-form measures of parenting stress that are both valid and feasible for a daily diary study. Our choice of measure, while not perfect, reflects the best available option at the time and underscores the need for the development of short-form measures of parenting stress for rapid, repeated assessments. We highly encourage future psychometric research efforts to distill more comprehensive measures of parenting stress into shorter, validated forms suitable for daily use. Moreover, all measures of child behavior and parenting stress in the current study were self-reported by a single informant and are therefore susceptible to response biases. Although parent perceptions are important (e.g., Haney et al., 2018), future research would benefit from multi-informant measurement of child behavior (e.g., including behavioral observations from teachers or clinicians) and use of objective indicators of parenting stress (e.g., physiological biomarkers like cortisol). Additionally, the participant inclusion criteria dependent on parents’ relationship status and specific child age range (i.e., 3–5 years) may limit generalization to the larger population. We also acknowledge potential limitations associated with the time-scale of the daily diary study. While the current study captured over 1800 observations of parenting stress and behavior problem severity across a 14-day period, it is possible that we would see additional lagged associations unfold if observations were recorded over a longer time period (e.g. a month or greater), which may be especially useful in examining within-person effects. Finally, while we speculate that parenting behaviors and coping strategies may account for some of the observed effects of parenting stress on child behavior problems and vice versa, we did not measure such constructs in the current daily-diary study. Future studies should include measures of more proximal predictors of child behavior problems (e.g., parenting strategies) to empirically test these theorized mediational models.

Conclusions

Using daily diary methodology, this study found further evidence of a positive bidirectional relationship between child behavior problems and parenting stress at the between-person level. This relationship was moderated by ethnicity, such that the positive relationship between child behavior problems and parenting stress was weaker for Latino (versus non-Latino) families, potentially indicating partial buffering within Latino families. Meanwhile, at the within-person level, greater parenting stress than usual, compared to an individual’s own average, was associated with lower next-day child behavior problem severity. While the within-person findings may reveal a more immediate transactional dynamic, possibly indicating short-term effects of parenting behaviors or coping mechanisms following child behavior problems within a given family, the between-person findings portend how these processes may accumulate over time. Findings underscore the value of using daily diary methodology to decompose between- and within-person effects, revealing potential intervention targets to improve parent and child functioning and address health inequities among minoritized families of autistic children.