Introduction

The recent societal increase in concerns about fathers’ participation in childcare has prompted substantial research focus on paternal parenting behavior, which has led to greater understanding of the relationships between infants and their fathers1,2. Paternal participation in parenting can reduce maternal stress and have positive effects on the physical (including neuronal) development of infants3,4. A lack of paternal parenting assistance is associated with a higher risk of maternal postpartum depression and parenting anxiety5, and of problem behaviors in children6. This association between paternal parenting and both maternal mental health and infant development highlights the importance of identifying ways to encourage fathers to participate in childcare.

Several papers have discussed the factors associated with the absence of paternal parenting behavior. The risk of fathers not participating in child-rearing is increased if the family live with grandparents7,8, if mothers do not work or work part-time7,9,10, if fathers work long hours10, and younger paternal age10,11. Psychological health problems also affect paternal parenting behavior. Depression in fathers is associated with less positive parenting (e.g., lower frequency of reading, singing songs, and telling stories12;), and more negative parenting13. Attention-deficit/hyperactivity disorder symptoms in fathers are associated with negative parenting, child-rearing disagreements14, lax or permissive parenting15,16, and less involvement with their children17. However, comparatively few studies have investigated the parenting behaviors of autistic fathers18.

According to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, autism spectrum disorder (ASD) is characterized by persistent deficits in social communication and interaction, as well as a restricted, repetitive patterns of behaviors, interests, or activities19. Dissanayake et al.20 found the association between parents with ASD and more difficulties in modelling and teaching their child behaviors, understanding the needs of their child, being spontaneous in parenting situations, and coping with sensory stimuli around their child. According to cross-sectional study, there was the association between mothers with ASD and a lax parenting style16. In interview to mothers with ASD, they may experience stress owing to both higher sensory stimulation during lactation and higher auditory stimulation when children cry or play21. In interview or internet survey, mother with ASD answered that they may have difficulties when required to multitask or ask an expert for child-rearing support, feel isolated from others, and experience high anxiety18,22. Compared with research on mothers with ASD, parenting behavior in fathers with ASD is not elucidated. Although the study investigated that autistic fathers associated with lower parenting efficacy23, there were still unknown. Because there is a higher incidence of ASD among men and that men with ASD show more restricted, repetitive, and stereotyped behaviors than women15,24, it is important that the parenting styles or behaviors of father with ASD are investigate the population-based study.

However, in previous study, there were studies about parents with ASD using interview to mothers and small size questionnaire survey, there are no big population-based studies on the parenting behavior of parents with autism traits. It has been suggested that research on childcare and ASD should focus not only on parents with diagnosed autism but also on those with milder ASD-like cognitive and personality traits but no autism diagnosis23,25. Thus, we investigated the association between autism traits and parenting behaviors in fathers using Japanese general populations.

Results

In this study, we measured autistic characteristics in fathers regardless of the presence of an ASD diagnosis. A total of 6.7% of fathers reported autism traits as screened by the AQ-J-10 (Table 1, supplementary Fig. 1). There was significant difference between existence of father’s autism traits and depression, education, number of children and house income, however, there were no difference between existence of father’s autism traits and marriage status, mother’s working and living with grandparents (Table2). Caregivers/mothers reported that 25.9% of fathers with autism traits did not prepare meals for the child or assist the child in eating; 22.3% did not change the child’s diapers or clean up the waste; 12.0% did not bathe with the child; 18.1% did not help the child to change clothes; and 4.7% did not play with the child (Table 3).

Table 1 Father’s AQ-10 score.
Full size table
Table 2 Characteristics of participants.
Full size table
Table 3 Fathers’ participation in parenting behaviors.
Full size table

To investigate which parenting behaviors were performed by fathers with autism traits compared with fathers without autism traits, logistic regression analysis was performed. We found that fathers with autism traits had a significant tendency to not prepare meals for the child, not assist the child in eating, not help the child to change clothes; and not play with the child (crude OR = 1.10 [95% CI: 1.01, 1.21]; crude OR = 1.18; 95% CI [1.06, 1.30]; crude OR = 1.24; 95% CI [1.03, 1.49]) (Table 3). Adjusted model, which included all the confounders, showed an association between autism traits in fathers and not preparing meals for the child or helping the child to eat and not helping the child to change clothes (adjusted OR = 1.11; 95% CI [1.01, 1.23]; adjusted OR = 1.17; 95% CI [1.04, 1.31]). However, we found no significant association between the other paternal parenting behaviors and autism traits (not changing the child’s diapers and cleaning up the waste; not bathing with the child; and not playing with the child) (Table 4).

Table 4 Association Between Paternal Autism Traits and Non-Participation in Parenting Behaviors.
Full size table

To determine whether particular types of autism traits were associated with specific parenting behaviors, we performed logistic regression analyses of the associations between each AQ-J-10 subscale score and each paternal parenting behavior. We found some associations between subscale scores and parenting behaviors (Table 5). Fathers with poorer communication skills had a tendency not to perform most parenting behaviors. Interestingly, fathers with deficits in social skills and deficits in attention-switching had a tendency to change diapers in a score-dependent manner (for social skills: adjusted OR = 0.90, 95% CI [0.85, 0.97] for a score of 1; 0.85 [0.79, 0.92] for a score of 2; and 0.77 [0.69, 0.87] for a score of 3; for attention-switching: adjusted OR = 0.93, 95% CI [0.87, 0.99] for a score of 1).

Table 5 Association Between Autism Trait Subscale Scores and Non-Participation in Parenting Behaviors.
Full size table

Discussion

To our knowledge, this is the first study to examine the associations between autism traits and paternal parenting behaviors. We found that although fathers with autism traits had significantly different parenting tendencies to fathers without autism traits. Father’s autism traits (i.e., those who scored ≥ 7 on the AQ-J-10) were significantly associated with less likely to assist the child in putting on or taking off clothes compared with fathers without autism traits, and to prepare meals for their 2-year-old children, help the child to eat. However, the analysis of AQ-J-10 subscale scores showed that fathers who reported difficulties with attention-switching and social skills associated with changing diapers.

One possible reason why the association between autism traits and not to help the child change clothes may relate to the fact that children develop a sense of self at approximately 2 years. Children at this age are less likely to listen to parental instructions26. This may lead to more hostile parent–child relationships, resulting in behaviors such as the child resisting when changing clothes, not sitting still, running around naked, or being particular about preferred clothing (reflecting the development of the ego). Because individuals with autism traits may have difficulty understanding the emotions and intentions of others, solving problems flexibly, and educating their children about appropriate behavior20,27, it may be difficult for fathers with autism traits to understand why their child is agitated when changing clothes, to respond to their child’s unpredictable movements, or to persuade the child to change their clothes. Additionally, effective child-rearing requires parents to be sensitive to children’s needs according to their behaviors in particular situations, and to respond to accordingly28. Fathers with autism traits may find it difficult to quickly recognize the needs of their children (e.g., when clothes are dirty from eating or sweating) and to react spontaneously; thus, they may be less able to encourage and help their child to change clothes by themselves. Therefore, it may be necessary for mothers or other caregivers to inform fathers with autism when they need to change a child’s clothes. One possible reason why the association between autism traits and not to prepare meals for their 2-year-old children and help the child to eat is that children gradually develop the ability to eat on their own; fathers with autism traits may be less likely to help meals for their child because they may find it difficult to judge the appropriate level of help the child needs. Most parents of 2-years-olds direct them to focus on eating rather than playing at mealtimes and encourage food intake if it is insufficient29. However, there was a report about the association between parents with autism traits and difficulty to attend to and instruct children20, and it may therefore be less able to help their children at mealtimes.

We found no evidence for an association between paternal autism traits and the extent to which fathers played with their children, the most interactive of the evaluated parenting behaviors. Although these results were unexpected, it is favorable for parents and children. This is because play is important in promoting children's development and emotional stability, and fathers' play with their children also helps reduce the mother's burden of housework and childcare. In addition, they may contribute to children's development in a different way to mother’s play30,31. However, it is possible that fathers with autism traits simply engage in their own preferred activities in the same space as their child. Thus, further study to understanding father’s parenting behavior with autism traits should be needed.

In this study, we also analyzed the associations between paternal parenting behaviors and scores on the four subscales of the AQ-J-10, which measure difficulty in social skills, communication, imagination, and attention-switching. Father’s autism traits in social skills associated with changing diapers. This indicates that these fathers, who may prefer to be alone or have difficulty understanding others, may find the mechanical behavior of diaper-changing easier to perform because it requires less interaction with children. Men with ASD tend to repeat the same behaviors24,32 and to do so accurately33,34; both characteristics are likely helpful when changing diapers. In addition, attention-switching deficits also associated with changing diapers. Fathers with attention-switching difficulty may have a strong interest in a particular task or subject and feel uncomfortable if they are not immersed in it. Although it is difficult to fully understand the mechanisms underlying parenting behaviors, it may be that fathers with autism traits recognized these characteristics in themselves and so chose to change diapers because this task does not require immediate attention and takes little time to perform. It is also possible that others understood the father’s characteristics and suggested that they take on that particular task. Whatever the reason, it is notable that we were able to identify an aspect of childcare that makes good use of the strengths of fathers who have difficulty with social skills and attention-switching.

In contrast, fathers with greater difficulty in communication skills significantly associated with not to perform most parenting behaviors. Generally, people with autism traits may experience problems with interpersonal communication, either because they phrase things in a way that others find rude or because they do not like to interact with others. Although infancy (birth to 12 months) requires mostly basic childcare (e.g., feeding and bathing), toddlerhood (12 to 35 months) requires more social interaction30. Therefore, fathers with communication difficulties may find it hard to perform parenting behaviors both during and beyond toddlerhood. It is important to respect the childcare that autistic fathers are able to do, while supporting them in childcare that they are significantly less able to do compared to non-autistic fathers.

There were three main study limitations. First, we used a simple evaluation of the frequency with which each paternal parenting behavior was performed (categorized as yes or no); therefore, the ways in which the parenting behaviors were actually performed were not assessed. To elucidate this, future studies should examine the specific ways and situations in which parenting behaviors are performed. Second, because the AQ-J-10 questionnaire was self-administered, autism traits may have been overestimated or underestimated. Third, it is unknown whether fathers with autism traits perform multiple types of parenting behavior simultaneously. Mothers with autism traits can experience difficulty multitasking while performing parenting behaviors22, so additional studies on this issue are warranted. Despite these limitations, this study is the first to examine parenting behavior in fathers with autism traits. Notably, we were also able to identify specific characteristics that are advantageous to parenting behavior, which may extend the parenting tasks in which fathers with ASD can excel. More research is needed to identify the specific child-rearing situations of fathers with ASD using research designs that can overcome the above-mentioned limitations.

Conclusions

Fathers with autism traits were significantly less likely than those without autism traits to help their children in changing clothes and at mealtimes. In contrast, fathers who had difficulties with social skills and attention-switching were more likely to change diapers. There is a need to promote the aspects of child-rearing that capitalize on the strengths of fathers with autism, while compensating for the areas in which they perform less well.

Methods

Study design and participants

The Japan Environment and Children’s Study (JECS) is a nationwide birth cohort study funded by the Japanese Ministry of Environment. The purpose of the JECS is to identify specific environmental factors that affect children’s health and development. Kawamoto et al.35 have described the details and protocol of the JECS. In brief, the JECS recruited 97,413 mothers who lived in the study areas (15 regional centers across Japan) between January 2011 and March 2014.

In this study, we analyzed the dataset jecs-ta-20190930. This dataset was released in October 2019 and contains information from 104,062 records, including the data collected from 2011 to 2014. This dataset does not contain any information that could identify participants.

Participants

Of the 104,062 records in the JECS dataset, we excluded 55,599 records that had missing data because of non-participation of the father, the father had participated more than twice, or the presence of multiple births or stillbirths. We excluded a further 6904 records because the children were not living with their parents at age 2 years or because of missing data on family structure. We excluded 1695 records with missing data for more than one question on the Autism Spectrum Quotient (AQ10). Therefore, the sample comprised 39,864 records containing the data for single-birth children who lived with both parents, and whose parents had answered all the AQ-J-10 questions (Fig. 1). In analysis, we performed complete-case analysis.

Figure 1
figure 1

Flowchart of Participant Selection Process.

Full size image

Measurement of autism traits

The AQ was originally developed by Baron-Cohen and his colleagues36 to screen for ASDs in adults. The AQ10 Japanese version (AQ-J-10) uses 10 questions from the original AQ to screen for high-functioning pervasive developmental disorder in Japanese adults36. We measured the autism traits of fathers using the AQ-J-1037. AQ-J-10 was answered by fathers.

Participants rated each of the 10 questions on a 4-point Likert scale with the following options: definitely agree, slightly agree, slightly disagree, and definitely disagree. Agree responses (definitely or slightly) were given a score of 1; disagree responses (definitely or slightly) were given a score of 0. For reversed items, the scores were reversed: slightly disagree and definitely disagree were scored as 1; definitely agree and slightly agree were scored as 0. The cutoff value for the AQ-J-10 is 6/7 and the diagnostic accuracy is 88%36. The AQ-J-10 comprises four subscales measuring social skills, communication, imagination, and attention-switching. The social skills items are “I prefer to do things with others rather than on my own” (reverse content), “I find it easy to work out what someone is thinking or feeling just by looking at their face” (reverse content), and “I find it difficult to work out people’s intentions.” The communication items are “Other people frequently tell me that what I’ve said is impolite, even though I think it is polite” and “I am often the last to understand the point of a joke.” The imagination items are “When I’m reading a story, I find it difficult to work out the character’s intentions,” “I would rather go to the theater than to a museum” (reverse content), “I like to collect information about categories of things (e.g., types of cars, birds, trains, plants)” and “I find it difficult to imagine what it would be like to be someone else.” The attention-switching items are “I tend to have very strong interests, which I get upset about if I can’t pursue”37,38.

Measurement of paternal parenting behavior

Paternal parenting behavior was evaluated when children were 2 years old using a questionnaire completed by mothers or caregivers, who were asked to rate the frequency with which fathers performed five different childcare behaviors: “Father’s frequency of preparing meals for the child and helping her or him eat”; “Father’s frequency of changing the child’s diapers and cleaning up the waste”; “Father’s frequency of taking a bath with the child”; “Father’s frequency of helping the child change clothes”; and “Father’s frequency of playing with the child.” Responses of always or sometimes were given a dummy variable value of 0 and counted as a yes, and responses of seldom or not at all were given a dummy variable value of 1 and counted as a no.

Covariates

Based on previous studies about related factors about the father’s parenting behavior, we selected the covariates. Furthermore, it should be noted that autism and psychosis tend to co-occur39, that autism has been suggested to have genetic factors40, and that children may have developmental issues. Therefore we selected the following items as covariates: for fathers, psychological distress2,12 (Kessler Screening Scale for Psychological Distress [K6] score of < 13 or ≥ 13)41, daily working hours 10,42 (≤ 9 h or > 9 h), education42 (university or not), age10,11 (< 16 years or ≥ 16 years); for mothers, employment status7,10 (employed/not employed), psychological distress43,44 (K6 score of < 13 or ≥ 13)38; for children, birth order45 (1st born or ≥ 2nd born), sex9,10 (male/female), cognitive developmental delay46 (yes/no); and for socioeconomic factors, marital status47 (married or not married), household income41 (< 2 million yen or ≥ 2 million yen; 1 million yen = approximately 6500 US dollars), and living with grandparents7,8 (no or yes). There was no multicollinearity between these factors (VIF range is 1.000–1.032).

Statistical analysis

Participants status was summarized by descriptive statistics. To investigate the difference between fathers with autism traits and without autism traits, we performed Chi-square test. To calculate crude ORs, we performed a univariate logistic regression analysis using paternal parenting behaviors as objective variables and fathers’ autism traits as the explanatory variable. To calculate adjusted ORs, we then performed a multivariate logistic regression analysis using the following model: Adjusted for fathers’ psychological distress, length of fathers’ workday, fathers’ education, and fathers’ age, mothers’ working status, living with grandparents, marital status, and household income, mothers’ psychological distress, firstborn, sex of child, and presence of cognitive developmental delay. To analyze the autism traits, we summed the scores on each AQ-J-10 subscale. We defined a score of 0 as the reference category for each subscale, generated the dummy variables, and calculated the ORs of each score. IBM SPSS Statistics 27 (IBM Corp., Armonk, NY, USA) was used for the analysis, with the significance level set at p < 0.05.

Ethical declarations

The study was conducted according to the principles of the Declaration of Helsinki. The JECS protocol was reviewed and approved by the Ministry of the Environment’s Institutional Review Board for Epidemiological Studies and the ethics committees of all participating institutions. Written informed consent was obtained from all study participants.