Transactional Relationships Between Maternal Requesting and Responsivity, Adolescent Compliance, and Task Demands in Fragile X Syndrome

Abstract

Background:

Males with fragile X syndrome (FXS) may have difficulty with social communication and language in addition to behavioral concerns such as noncompliance to maternal requests. Mothers vary in how they pose requests to their children, and contextual demands may also be a factor in compliance. This study examined the relationships between maternal requesting behaviors, maternal responsivity, child compliance, and task demands in adolescent males with FXS.

Methods:

35 mother-son dyads completed three interactive tasks during in-home visits (puzzle, iPad games, snack). The adolescents also completed assessments of language and autism symptoms. The three interactive tasks were video-recorded, and behavior-by-behavior coding and transcription was completed.

Results:

Comparing between tasks, mothers requested a behavior most often during the snack task and used a higher proportion of indirect requests during the iPad task. Adolescents were largely compliant across tasks, with average compliance equal to 70%. Adolescent compliance was predicted by maternal request frequency such that mothers who used more requests had adolescent sons who were less compliant. Maternal responsivity was higher for mothers of adolescents with FXS-only compared to those with FXS and high autism symptoms.

Conclusions and Implications:

Task demands may have influenced maternal requesting, which in turn may have impacted adolescent compliance. Compliance with requests is important for both social and educational aspects of life, and further study of requesting and compliance in FXS is needed to identify effective teaching and behavioral intervention methods.

1. Introduction

Throughout development, children are generally expected to comply with parental requests to the extent that when children are noncompliant, they may be labeled as difficult or defiant, rather than seen as developing autonomy and self-identity. Certainly, a degree of compliance with parental requests is how children learn sociocultural norms and how to be successful members of society (Bornstein, 2009). Within Sameroff’s transactional framework for child development (see further Sameroff, 2009), parents, children, and context interact to shape one another, and as such parental behavior and parenting may influence child behavior and development, and vice versa. The frequency of parental requests and the language parents use when requesting may shape the likelihood of child compliance. Additionally, task demands may influence parental requesting strategies, child compliance, or general behavioral expectations and inclinations. This becomes particularly relevant when considering adolescent compliance, as adolescence is a time of identity-creation and autonomy within a set of parental expectations. This study focused on the transactional relationships between maternal responsivity and requesting, adolescent compliance, and task demands in mother-son dyads with fragile X syndrome (FXS). Specifically, we examined (1) whether task demands influenced maternal request frequency and directness, maternal responsivity, and adolescent compliance, and (2) which, if any, adolescent characteristics (i.e., language, autism symptomology) and maternal characteristics (i.e., directness, responsivity) predicted adolescent compliance. We also probed whether maternal requesting behavior, maternal responsivity, and adolescent compliance differed between boys with FXS-only (FXS-O) and those with FXS and co-occurring autism (FXS+ASD).

1.1. Fragile X Syndrome

Fragile X syndrome (FXS) is a heritable, monogenic disorder that often presents with deficits in language and social communication, cognition, daily living skills, and other adaptive behaviors, as well as elevated incidence of challenging behaviors (Finestack et al., 2009; Hardiman & McGill, 2018; Lewis et al., 2006; McDuffie et al., 2010; Warren et al., 2017). Because the gene responsible for FXS is located on the X chromosome, males and females are differentially affected, with boys typically having more severe deficits and behavioral challenges than girls, albeit with considerable heterogeneity. Due to the sex differences in FXS symptomology, this study focused on males only. As many as 50–75% of boys with FXS also meet diagnostic criteria for autism spectrum disorder (ASD; Abbeduto et al., 2019; Klusek et al., 2014b). Those who meet ASD criteria often demonstrate lower intellectual function and more severe deficits in language (Hinton et al., 2013; McDuffie et al., 2012) as well as characteristic deficits in social communication (Klusek et al., 2014a; Lorang & Sterling, 2017; Losh et al., 2012; Martin et al., 2017). Co-morbid ASD is also associated with elevated problem behaviors and noncompliance (Abbeduto et al., 2014; Hall et al., 2016; Hustyi et al., 2014; Newman et al., 2015; Smith et al., 2012). However, maternal responsivity may moderate the effects of autism symptomology on language, social communication, and problem behaviors in children with FXS (Brady et al., 2014; Warren et al., 2017; Warren et al., 2010).

1.2. Maternal Responsivity and Requesting

Within FXS research, maternal responsivity has been measured during mother-child interactions as the degree to which mothers use comments, requests for verbal compliance (e.g., questions that introduce or maintain topics, etc.), and recodes (i.e., verbal interpretations of the child’s communication act) (Brady et al., 2014; Warren et al., 2010). Behaviors that mark responsive parenting styles include contingent responses to child bids for social interaction and following the child’s lead. Children with FXS whose mothers are more responsive (i.e., have more positive affect and are warmer during mother-child interactions) tend to have better linguistic and behavioral outcomes than children whose mothers are less responsive or warm (Brady et al., 2014; Warren & Brady, 2007; Warren et al., 2017; Warren et al., 2010).

While maternal responsivity encompasses positive parenting behaviors, maternal requests for behavioral compliance are not as clearly positive or negative. Requests for behavioral compliance can be teaching tools, or they may be overly directive and insensitive to the child’s abilities and independence. The specific linguistic structure of the request for behavioral compliance may impact the likelihood of compliance. Request directness is the degree to which the syntactic form and pragmatic function of the request align. Direct requests are ones where the form is an imperative statement, and the function is clearly, and unambiguously, a request, such as Pass me the butter (Clark, 1979). In contrast, indirect requests are ones in which the form is not an imperative (e.g., a declarative or interrogative) but the intended function is still a request. In indirect requests, the function may be ambiguous or unclear to the listener (Clark, 1979; Searle, 1969), as in Do you mind passing me the butter? or My roll needs butter + reach gesture. Because indirect requests are ambiguous, they may be more difficult to interpret and comply with than their direct counterparts (Clark, 1979; Ledbetter & Dent, 1988; Pinker et al., 2008; Searle, 1969).

It was once believed that practitioners working with children with language delays, such as those with neurodevelopmental disorders, should not use indirect requests, as they may be overly confusing (Blue, 1981). Studies addressing this concern have found mixed results (note: there have been no studies on this topic in FXS, so we reference other neurodevelopmental disorders in which this has been examined). Despite known difficulties with non-literal or figurative language (Kalandadze et al., 2018), children with ASD are able to use contextual cues during a contrived act-out task to arrive at the correct interpretation of indirect requests (Kissine et al., 2015). In contrast, Bryce and Jahromi (2013) reported that, compared to language-matched typically developing children, children with ASD were significantly less likely to comply following an indirect parent request during a clean-up activity. Differences in context or task may be responsible for the contrasting findings. Nonetheless, compliance with a request may be compromised if the listener does not understand the request, which could be a product of language deficits (Ledbetter & Dent, 1988) or of the inherent social communication deficits in ASD (Bryce & Jahromi, 2013).

1.3. Effect of Task Demands

Predetermined interactive contexts provide structure to parent-child interactions, and depending on the perception of the task structure, maternal requesting behaviors and child compliance may vary. Mothers may adapt their behavior based on the task at hand with each requiring different amounts of direction and guidance. For instance, mothers of children with Down syndrome (DS) adjust their requesting behaviors based on social context and the child’s need for supports in a given task (Landry et al., 1994). When the mother’s request follows the child’s lead, compliance is higher, suggesting that children with DS are more likely to comply for parents who are more flexible.

Children too may be sensitive to task demands. Young typically developing children do well with routinized requests and with action requests (Ervin-Tripp & Gordon, 1986), and task familiarity and action norms may impact their request comprehension and compliance. Similarly, many individuals with ASD prefer routine and may have difficulty with transitions or changes in routine (American Psychiatric Association, 2013). As such, for individuals with symptoms of ASD, context and routine familiarity may help clarify task expectations, guide routinized action sequences, provide structure to an activity, and ultimately increase likelihood of complying with parental requests. When provided with a structured act-out task in a lab setting, children with ASD were compliant with indirect requests (Kissine et al., 2015), whereas when they were doing a clean-up task with a parent, they were not (Bryce & Jahromi, 2013). This suggests that children with ASD are sensitive to task and context expectations, to a certain degree.

In this study, we asked participants with FXS and their mothers to complete three different activities together which provided them with a variety of tasks during which they were expected to interact. These tasks were chosen because we wanted activities that were naturalistic and varied. We selected tasks that we thought would provide multiple opportunities for interaction between members of the dyad—past observations indicated that without some scaffolding we frequently saw very low rates of interaction. The iPad task (an interactive game with the mother) was expected to be a familiar interface for many adolescents that would be rewarding and engaging. The puzzle task (Castle Logix ^™) was relatively structured (i.e., there were solutions to the puzzles provided) and unlikely to be very familiar. The snack task was also fairly structured, as we provided a list of ingredients and a numbered list of steps to complete to make the snack. Additionally, the snack task was expected to be inherently rewarding and engaging due to the nature of the snack itself —chocolate pudding, gummy worms, and Oreos.

1.4. Current Study

We investigated how maternal behaviors and task demands related to child compliance because compliance is instrumental in teaching and learning. Parents teach their children about the world and about sociocultural norms through interactions which require, or at least expect, compliance (Bornstein, 2009). In addition, noncompliance may be associated with problem behaviors and poorer school functioning (Menezes & Mazurek, 2021). It has been well-established that males with FXS demonstrate elevated levels of problem behavior and noncompliance (Abbeduto et al., 2014; Hall et al., 2016; Hustyi et al., 2014; Newman et al., 2015; Smith et al., 2012). The effects of child problem behavior permeate into maternal mental health and physical well-being, as well as family dynamics and parenting stress (Baker et al., 2012; Fielding-Gebhardt et al., 2020; Mailick et al., 2017), and noncompliance can negatively affect the child’s educational success (Austin & Agar, 2005; Menezes & Mazurek, 2021). As such, interventions and parent support aimed at alleviating noncompliance and other challenging behaviors is an important area of focus (O’Nions et al., 2018)—one which has had little attention in the FXS literature. By studying the variables surrounding compliance, we may identify positive targets for future interventions that aim to increase compliance and thereby reduce noncompliance.

In examining compliance in adolescent males with FXS, this study considered three variables affecting a dyadic interaction—the mother, the child, and the task demands. Previous investigations in FXS that have employed the transactional model of development (Sameroff, 2009) have demonstrated the interconnectedness between mother and child. However, the integration of task demands into mother-child interactions has not been adequately considered in FXS. Our research aims were to determine (1) whether task type influenced maternal request frequency and directness, maternal responsivity, and/or child compliance, (2) examine differences in maternal behaviors and child compliance in adolescents with FXS-O and those with FXS+ASD, and (3) identify child and maternal characteristics associated with child compliance.

2. Method

2.1. Participants

Thirty-five mother-son dyads provided data for these analyses. Each dyad consisted of an adolescent or preadolescent son (12 to 18 years old) diagnosed with full mutation FXS and their biological mother. On average, the sons were 16.13 years old, and the mothers were 46.68 years old. Of the 35 mothers, 33 were FMR1 premutation carriers, one carried both the full mutation and premutation alleles (mosaicism), and one carried the full mutation.

Participants were enrolled in an ongoing longitudinal study (PIs: Warren & Brady, see publications). They were recruited through advertisements at national conventions, a national research registry, and Fragile X family networks within local communities. The sample included families from across the United States but represents a sample of convenience due to the rareness of the disorder. As such, racial, ethnic, and economic diversity were limited, see Table 1. Prior to participating in research activities, all mothers provided informed consent for their child and themselves to participate in the study, in accordance with the University of Kansas - Lawrence Institutional Review Board, and all procedures were performed in accordance with the ethical standards of the University of Kansas - Lawrence Institutional Review Board.

Table 1:

Sample Demographics

Characteristic	Level	Count	Percentage
Child Ethnicity	Non-Hispanic Hispanic	32 3	91.43 8.57
Child Race	White Black Asian Native American Native Hawaiian or Other Pacific Islander	33 4 0 0 1	94.29 11.43 0.00 0.00 0.00
Mother Ethnicity	Non-Hispanic Hispanic	32 3	91.43 8.57
Mother Race	White Black Asian Native American Native Hawaiian or Other Pacific Islander	33 2 0 0 1	94.29 5.71 0.00 0.00 2.86
Mother Work	No Yes	11 24	31.43 68.57
Mother Education	High School Diploma or Equivalent Some Post High School Associate’s or Technical Degree Bachelor’s Degree Some Post Bachelor’s Master’s Degree 2 Master’s or Doctoral or Professional Degree	3 8 0 14 4 4 2	8.57 22.86 0.00 40.00 11.43 11.43 5.71
Mother Marital Status	Engaged Separated Divorced Married Single, never married	1 0 8 24 2	2.86 0.00 22.86 68.57 5.71
Household Income	> $100,000 $80,000 - $100,000 $50,000 - $79,999 $30,000 - $49,999 $15,000 - $29,999 < $15,000	23 2 6 0 2 2	65.71 5.71 17.14 0.00 5.71 5.71

2.2. Procedure

Dyads were visited in their homes between 2016 and 2019 by two researchers. The lead examiner attended every visit, and the secondary examiners were graduate students and post-baccalaureate assistants who alternated visits. During the home visit, which lasted around 4 hours, the secondary examiner would conduct a series of cognitive and linguistic assessments with the adolescent, and the lead examiner would complete several interviews and assessments with the mother. After the interviews and assessments were finished, the mother and adolescent would come together to participate in three interactive tasks. Each task lasted 10 minutes. Dyads were instructed to work on the task together for 10 minutes and told that the study team was interested in their communication with one another, based on a standard script read aloud to dyads before each task. For the first task, the dyad completed Castle Logix^™ puzzles together. The dyad was provided with the puzzle pieces and a booklet of examples they could attempt along with solutions to those examples. For the second task, the dyad played two-player games on an iPad that was provided by the study team which had pre-selected two-person games installed and no internet access or data plan. Finally, the third task was to make a snack together. The dyad was supplied with the materials for dirt cups—chocolate pudding, sandwich cookies, and gummy worms—along with a recipe card with step-by-step instructions that were written in simple text alongside pictures. The tasks were completed in the same order for each dyad to ensure that fatigue effects would be constant across dyads. In general, tasks were completed back-to-back without extended breaks between tasks. On occasion, there were short breaks if the child needed time/space to regulate their behaviors before returning to the tasks.

Each mother-son interaction was videotaped and audio-recorded. Following the visit, each interaction was transcribed by utterances and coded behavior-by-behavior for maternal communication and child compliance (see Table 2).

Table 2:

Coding Definitions and Examples

Maternal Communication	Definition	Examples
Request for Behavioral Compliance	A directive to which the child can comply behaviorally. Can be (1) an imperative statement, (2) an indirect request with clear intent/expectation, (3) a gesture, usually paired with an expectant pause, (4) hand-over-hand guidance/manipulation, and (5) imperative statement directing the child to read.	• I need your help. *indirect • You should sit down. *indirect • Give me the milk. *direct • Can you throw this away? *indirect
Request for Verbal Comply	A question or statement intended to elicit a verbal response from the child. The mother must provide the child with at least 2 seconds to respond to the request, otherwise default to comment.	• Say thank you. • Tell me when. • Where do you want to go for lunch? • What games are there?
Recode	A correction or expansion of a child’s unclear or incomplete communication act.	• Yes, the yellow ball. • That’s your airplane.
Comment	Any utterance by the mother that does not fit into one of the other categories described. Includes verbatim imitation, abandoned utterances, and statements to draw the child’s attention.	• Thank you for helping. • I know you like this flavor of pudding. • Tomorrow we can go to the store.
Praise	A statement that makes a positive appraisal of the child; a child-directed compliment.	• Good job. • You’re so sweet.
Child Compliance	Definition	Examples
Compliant	The child responds in a socially appropriate manner to the mother’s request for behavioral comply and an attempt is made to complete the request (i.e., does not need to successfully complete the request, but makes an attempt).	• Child comes towards the mother (but does not sit down fully) when asked to sit down. • Child throws aways the trash when asked to do so.
Noncompliant	The child ignores the request or otherwise does not comply or declines the mother’s request.	• Child leaves the room when asked to sit down. • Child yells “no” when asked to throw away the trash.
No Opportunity to Comply	The mother completes the request herself, uses hand-over-hand prompts/assistance, or does not give the child enough time to complete the request before her next request for behavioral compliance.	• Mother throws away the trash herself. • Mother takes child’s hand to manipulate the iPad.

2.3. Child Measures

2.3.1. Language.

The Peabody Picture Vocabulary Test (PPVT-4; Dunn & Dunn, 2007) was used to measure adolescent receptive vocabulary and the Expressive Vocabulary Test (EVT-2; Williams, 2007) was used to measure adolescent expressive vocabulary. In the PPVT-4 participants are asked to identify a picture from among an array of four pictures according to the examiner’s prompt (e.g., point to puppy). Similarly, the EVT-2 asks participants to label a target word or provide a synonym based on the examiner’s prompt and paired image. The PPVT-4 and the EVT-2 both yield raw scores which can then be converted into standard scores and age equivalency scores. Together, these two assessments are widely used to determine general language ability in children and adolescents with FXS and other neurodevelopmental disorders.

2.3.2. Autism Symptoms.

Two observational assessments were used to evaluate autism symptoms and determine autism status. The Childhood Autism Rating Scale – 2nd Edition (CARS; Schopler et al., 2010) was completed by the two examiners who attended the home visit immediately upon completion of the visit. THE THIRD AUTHOR, a speech-language pathologist with a doctoral degree and over 20 years’ experience in the field, completed all CARS assessments along with graduate students who were studying language development and speech language pathology with extensive training on identification of autism. The CARS is an observational rating scale that provides a general impression of autism symptomology based on 15 items scored along a 7-point scale. The items capture various behaviors consistent with autism, including verbal and nonverbal language skills, adaptation to change, emotional response, unusual body use, unusual sensory interest, and self-injurious behaviors. It is also designed for use with individuals with impaired cognitive function. When completing this assessment, the examiners can use multiple sources of information including direct observation, parent report, and other assessments.

The Autism Diagnostic Observation Schedule – 2nd Edition (ADOS-2; Lord et al., 2012) is a play-based assessment during which the examiner creates opportunities for social interaction with the participant. The ADOS-2 includes many activities designed to create a social world between the examiner and participant including free play, puzzle completion, story-telling, bubbles, snack, and pretend birthday party or act-out stories. In this study, participants received Modules 1 – 4, based on their expressive language abilities. Per the ADOS-2 administration guidelines, only behaviors that occurred during the 45–60-minute assessment were considered when scoring the assessment results. All ADOS-2 administrations were performed by research staff who were trained by the test makers and research-reliable. Of the 35 participants, 9 received Module 1, 11 received Module 2, 8 received Module 3, and 5 received Module 4. Two had not completed an ADOS-2 at the time of analysis.

Autism status was determined based on ADOS-2 classification and CARS total score. Participants who had CARS scores ≥ 28.5 and whose ADOS-2 scores indicated autism spectrum or autism were in the FXS+ASD group (n=14), the rest of the participants were in the FXS-O group (n=21). This included participants whose CARS scores were < 28.5 but whose ADOS-2 scores indicated autism spectrum or autism due to discrepancies in classifications between assessments.

2.2.4. Nonverbal Cognition.

The Leiter International Performance Scale – Revised (Leiter-R; Roid et al., 2013) was used to assess nonverbal cognition. This assessment is administered in-person and while administering the assessment, the examiner cannot use verbal communication. Participants completed the Brief IQ test comprising four subtests: figure ground, form completion, repeated patterns, and sequential ordering. The Brief IQ test yields an IQ score, an age equivalency, and a growth scale value which are derived from the raw score.

2.4. Mother-Son Interaction Coding and Transcription

The 10-minute mother-son interactions were transcribed by utterances and coded for behavior using the Observer XT software, version 14 (Noldus Technology, 2016). Maternal communication was coded turn-by-turn. Each turn was assigned a behavior according to the function of the transcribed utterance(s), see Table 2. Maternal turn boundaries were marked by one of the following: the child taking a communication turn, the mother changing the function of her utterance(s), or a three or more second break between maternal utterances. Each maternal turn was a single function/behavior, and if that function/behavior changed (i.e., from comment to praise), then a new turn began. For example, if the mother said “I like pudding. Can you get that for me?” two turns were recorded because of the change in function (comment + request for behavioral comply). Transcripts were used to assign directness to each maternal request for behavioral compliance. In addition, each maternal turn was assigned a function/behavior, see Table 2, from which the following maternal constructs were derived.

2.4.1. Maternal Responsivity.

Previous investigations of maternal responsivity in FXS have defined maternal responsivity as a combination of maternal comments, recodes (verbal interpretations of the child’s communication act), and requests for verbal compliance (see Warren et al., 2010, Brady et al., 2014). See Table 2 for specific definitions. As such, maternal responsivity was the sum of turns that were scored as comments, recodes, or requests for verbal compliance which came from the behavioral coding. A total count was used for this construct so that there was no constraint on the maximum value.

2.4.2. Maternal Requests for Behavioral Compliance.

Rather than counting the total number of requests for behavioral compliance and including a total count in the analyses, we calculated the ratio of requests for behavioral compliance to all communication turns (see Table 2) because we wanted to get a sense of the frequency with which mothers used requests for behavioral compliance relative to the frequency with which they used all other communication behaviors. This approach controls for differences between mothers in total number of communication acts. For example, mothers who used nine requests for behavioral compliance out of a total 45 communication turns would have a request frequency score of 0.20 or 20%, indicating that relative to their total number of communication turns (i.e., comments + requests for behavioral compliance + recodes + requests for verbal compliance), they used requests for behavioral compliance 20% of the time. Mothers with higher request frequency scores were those who used a higher proportion of requests for behavioral compliance relative to their total number of communication behaviors.

2.4.3. Maternal Request Directness.

For each turn, including the transcribed maternal utterances, maternal request directness referred to whether a request for behavioral compliance was direct or indirect. A direct request was syntactically an imperative statement that had a single, readily interpretable meaning and as such was not ambiguous, e.g., “Give me the spoon.” An indirect request was one in which the syntactic form and pragmatic function of the request did not align, such that the intended request was syntactically phrased as a question or a suggestive statement rather than an imperative statement, see Table 2 for examples. With the same rationale as above, maternal request directness was quantified as the ratio between direct and indirect requests, with .20 or 20% indicating that mothers used direct requests 20% of the time.

2.4.4. Adolescent Compliance.

For each maternal request for behavioral compliance, the adolescent could respond in a compliant or noncompliant manner, or they could have no opportunity to respond, see Table 2. The response was compliant when the adolescent completed or attempted to complete the mother’s request, either in part or in full. A noncompliant response occurred when the adolescent did not attempt to complete the mother’s request, ignored the mother’s request, or refused the mother’s request. Finally, the adolescent had no opportunity to respond when the mother completed the action herself, physically assisted the adolescent in completing the action (i.e., hand-over-hand manipulation), or when she did not provide the adolescent with sufficient time (at least three seconds) to complete the action before issuing another request. If there were multiple requests for behavioral compliance in the same maternal communication turn (e.g., Put away the milk. Go put it over there. Can you put the milk away?), the adolescent’s compliance with the final request was considered. Because we were interested in how adolescents responded when provided an opportunity to do so, we did not include “no opportunity” in predictor analyses. However, we report between-group differences and associations for “no opportunity” requests.

2.4.5. Reliability.

Transcription and coding were done using Noldus Observer XT software (Noldus Technology, 2016). Reliability procedures for maternal communication and adolescent compliance differed from reliability procedures for request directness. Maternal communication (including requests for behavioral compliance) and adolescent compliance were coded during Noldus observational coding and were scored for reliability using a consensus scoring model. A primary coder transcribed and coded maternal communication acts, a secondary coder then reviewed the transcript and codes and noted any disagreements. Then, the two coders met to resolve disagreements through consensus. There were seven separate coders who rotated through as primary and secondary coders. Three of the coders were full-time post-baccalaureate research assistants, three were doctoral graduate students studying language development and developmental disabilities, and the final coder was the lead examiner who was a full-time Research Associate with a Ph.D. in speech-language pathology. Before coding independently, all coders were trained one-on-one by senior research staff to reach >85% reliability.

Request directness was coded after all Noldus observational coding was completed, as this was a variable unique to this analysis. All transcripts were independently reviewed and scored for request directness, with one-third of files scored by two independent coders. Percent agreement for request directness was high at 94% agreement, and the intraclass correlation coefficient (ICC) was also strong, with ICC=0.89.

2.5. Analyses

We used a combination of SAS^™ software, Version 9.4 of the SAS System for Windows (SAS Institute Inc., 2013), and SPSS^™ software, Version 27.0 (IMB Corp., 2020). Specifically, descriptives and tests for between-task differences were conducted using SPSS, and regressions, correlations, and figures were generated using SAS software. Differences between FXS-O and FXS+ASD were examined using analyses of variance (ANOVAs), while between-task, within-person differences were examined using repeated measures ANOVAs with task as the within-subjects factor. A priori post-hoc tests were used to determine the direction and significance of the differences in each variable between tasks. To determine predictors of adolescent compliance, a linear regression was conducted using PROC REG of the SAS software (SAS Institute Inc., 2013). Predictors were entered into the model using a stepwise method and were retained if they had p-values ≤ 0.15.

Adolescent receptive language and expressive language were highly significantly correlated (r=0.92, p<0.001), and as such, only receptive language was included in predictor models due to collinearity concerns and limitations in sample size. Additionally, receptive language was most closely related to the research question because it gives an indication of the participant’s ability to understand verbal commands. Similarly, nonverbal cognition scores were strongly associated with receptive and expressive language (all rs ≥0.70, ps <.001), and so nonverbal cognition was not included as an additional predictor due to limitations in power.

3. Results

3.1. Sample Description & Characteristics

The participants for this set of analyses included 35 adolescent males with FXS and their biological mothers. As expected, there was a considerable level of heterogeneity in language and nonverbal cognition among the adolescents, see Table 3. Receptive and expressive vocabulary were impaired across the group. On average, the adolescents had receptive vocabulary scores equivalent to 5.23 years of age (n=34), and for expressive vocabulary scores, the age equivalency was slightly higher at 5.72 years of age (n=28). Autism symptom severity, measured by the CARS was variable as well, with 60% of the sample having scores lower than the mild-to-moderate symptoms of autism cutoff (28.5). When considering CARS score along with ADOS-2 score, 14 of the 35 adolescents were considered to have FXS+ASD. Compared to adolescents with FXS-O, those with co-morbid ASD had significantly lower PPVT standard and raw scores, EVT raw scores, and Leiter-R growth scores. Mothers of adolescents with FXS+ASD were slightly less responsive (p=0.56), but both groups of mothers provided similar frequency of requests for behavioral compliance and adolescents were similarly compliant across groups. Finally, mothers of adolescents with FXS-O provided more opportunities to respond than mothers of children with FXS+ASD (p=.01).

Table 3:

Sample Characteristics

N			Mean	SD	Range	P-value
Adolescent Characteristics
Age (yrs)	FXS-Only	21	16.2	1.3	12.4 – 18.0	0.88
	FXS+ASD	14	16.1	0.8	14.3 – 17.5
CARS Score	FXS-Only	21	22	2.5	18 – 27.5	<.001
	FXS+ASD	14	34.3	4.3	28.5 – 40.5
ADOS-2 Comparison Score	FXS-Only	19	5.9	2.3	3 – 10	0.31
	FXS+ASD	14	6.7	1.8	3 – 10
PPVT Raw Score	FXS-Only	20	95.7	36	24 – 147	0.01
	FXS+ASD	14	61	36.7	28 – 130
PPVT Standard Score	FXS-Only	20	44	17.1	20 – 71	0.02
	FXS+ASD	14	29.6	17.1	20 – 61
PPVT Age Equivalency (yrs)	FXS-Only	20	6	2.1	2.25 – 9.5	0.01
	FXS+ASD	14	4.1	2	2.42 – 8.1
EVT Raw Score	FXS-Only	20	76.1	22.9	31 – 144	0.03
	FXS+ASD	8	50.5	35.7	6 – 106
EVT Standard Score	FXS-Only	20	51.7	17.5	23 – 78	0.10
	FXS+ASD	8	38.1	21.4	20 – 74
EVT Age Equivalency (yrs)	FXS-Only	20	6.2	1.9	3.1 – 9.9	0.06
	FXS+ASD	8	4.5	2.5	1.9 – 8.9
Leiter-R Brief IQ Score	FXS-Only	20	39	4.5	36 – 52	0.10
	FXS+ASD	12	37	0.9	36 – 38
Leiter-R Growth Score	FXS-Only	20	457	11.9	430 – 474	0.01
	FXS+ASD	12	433	18.3	404 – 466
Total # Compliant Responses	FXS-Only	21	72.9	28.3	35 – 121	0.81
	FXS+ASD	14	70.5	25.8	35 – 107
Maternal Behaviors
Total # Requests for Behavior	FXS-Only	21	27	17.2	2 – 60	0.05
	FXS+ASD	14	39.6	19.1	11 – 79
Prop. Direct Requests	FXS-Only	21	0.76	0.15	0.46 – 1.0	0.94
	FXS+ASD	14	0.76	0.15	0.44 – 0.96
Total # Responsivity Behaviors	FXS-Only	21	317.6	57.5	236 – 425	0.01
	FXS+ASD	14	247.4	95.4	102 – 403
Total # No Opportunity	FXS-Only	21	6.7	7.1	0 – 26	0.01
	FXS+ASD	14	20.2	21	1 – 75

^*Note: Autism status was defined based on CARS total score and ADOS-2 classification, see Method section. P-value based on ANOVAs comparing FXS-Only and FXS+ASD

3.2. Interactive Effects of Task, Maternal Requesting, and Adolescent Compliance

To determine whether maternal requesting behaviors (maternal responsivity, request frequency, and request directness) and adolescent compliance differed by task, we ran a series of repeated measure ANOVAs with task as the within-subjects factor. A priori post-hoc tests between tasks indicated the direction and significance of the differences in each variable between tasks. On average, maternal responsivity was highest during the puzzle task, see Figure 1 and Table 4, and there was a modest main effect of task on maternal responsivity (F_2,102=3.14, p=.05). The between-task post-hoc testing indicated that responsivity was significantly higher during the puzzle than snack (p=.04), but there were no differences between iPad and snack or iPad and puzzle, see Table 4.

Figure 1:

Maternal Responsivity, Request Frequency, Request Directness, and Child Compliance by Context and Autism Group

Table 4:

Contextual Differences in Maternal Responsivity and Requesting and Child Compliance

	N	Mean	SD	Range	F	P	Effect Size (Η2)
Responsivity					3.14	.05	.06
1. iPad	35	94.63	32.10	22–165	1v2	.24
2. Puzzle	35	106.29	30.62	46–163	1v3	.68
3. Snack	35	88.60	27.08	34–149	2v3	.04
Request Frequency					3.00	.05	.06
1. iPad	35	.24	.13	.02-.59	1v2	.98
2. Puzzle	35	.25	.08	.10-.42	1v3	.08
3. Snack	35	.30	.10	.11-.53	2v3	.11
Request Directness					7.12	.001	.13
1. iPad	35	.76	.15	.44–1	1v2	<.001
2. Puzzle	35	.62	.17	.26-.91	1v3	.04
3. Snack	35	.67	.14	.32-.91	2v3	.41
Compliance					1.32	.27	.03
1. iPad	35	.70	.26	.13–1	1v2	.66
2. Puzzle	35	.74	.18	.33–1	1v3	.24
3. Snack	35	.78	.15	.33–1	2v3	.73

There was a main effect of task on maternal request frequency (F_2,102=3.00, p=.05). Mothers had relatively more requests for behavioral compliance (i.e., gave more requests for behavioral compliance relative to other communication bids) during the snack task than the iPad or puzzle tasks, see Figure 2 and Table 4, however there were no statistical differences between tasks. On average, the requests were more often direct than not (>60% direct for each task). Mothers’ requests were most direct during the iPad activity, with significant differences emerging in directness between iPad and puzzle and between iPad and snack, see Figure 3 and Table 4. There was not a significant difference between snack and puzzle for directness. Adolescent compliance was relatively high across tasks, with all three tasks having at least 70% compliance, see Figure 4 and Table 4. Adolescent compliance did not vary by task (F_2,99=1.33, p=.27).

Fig 2:

Maternal Responsivity, Request Frequency, Request Directness, and Child Compliance by Context and Autism Group.

Fig 3:

Maternal Responsivity, Request Frequency, Request Directness, and Child Compliance by Context and Autism Group.

Fig 4:

Maternal Responsivity, Request Frequency, Request Directness, and Child Compliance by Context and Autism Group.

Upon visual inspection of the distribution of maternal requesting behaviors and adolescent compliance by task and autism status (Figs. 1–4), some patterns emerged. Specifically, it appeared that mothers with lower responsivity and mothers who requested behavioral compliance more had sons who were more likely to have autism. Indeed, correlations revealed that mothers who were less responsive had sons with significantly higher CARS scores (r=-0.50, p=.002), mothers who used a higher proportion of requests for behavioral compliance had sons with higher CARS scores (r=0.59, p=.0002), and mothers who provided more “no opportunity” requests had sons with higher CARS scores (r=0.34, p=.04), see Table 5. Importantly, autism status, as reported in the figures, was determined through both the CARS and ADOS-2 scores, so correlations between CARS scores and other variables do not completely align with autism status.

Table 5:

Correlations Between Maternal Responsivity and Requests, and Child Compliance, CARS Score, and Language Raw Scores

Correlations	Compliance	CARS score	PPVT	EVT
Responsivity	0.31	−0.50	0.44	0.33
p-value	0.08	.002	0.02	0.09
Request Frequency	−0.61	0.59	−0.71	−0.69
p-value	0.0001	0.0002	0.000	0.000
Request Directness	−0.15	−0.01	−0.05	−0.07
p-value	0.41	0.97	0.80	0.74
No Opportunity to Comply	−0.37	0.34	−0.42	−0.56
p-value	0.03	0.04	0.01	0.002
Compliance		−0.43	0.45	0.38
p-value		0.01	0.01	0.05

3.3. Predictors of Child Compliance

Because there were no differences in compliance across tasks, the following analyses used compliance averaged across tasks within each participant. As shown in Table 5, adolescent compliance was moderately associated with maternal responsivity such that higher maternal responsivity tracked with higher adolescent compliance; however, this association approached but did not reach significance (p=0.08). Maternal request for behavioral compliance frequency was also significantly correlated with adolescent compliance, such that adolescent compliance decreased when requests were more frequent. Adolescent compliance was also associated with autism symptoms (CARS score) and with receptive language (PPVT raw score). The first step in determining predictors of adolescent compliance was to examine maternal requesting behaviors and then test adolescent-level characteristics (i.e., autism symptoms and receptive language). As shown in Table 6, the only significant maternal effect was the frequency of requests for behavioral compliance; the adolescent-level effects were not significantly predictive above and beyond the effect of maternal request frequency. Although all models accounted for roughly 38% of the variance in adolescent compliance, the most parsimonious model was the one which only included maternal request frequency. As illustrated in Figure 5, the adolescents were less compliant when mothers used a higher proportion of requests for behavioral compliance compared to their other communication acts. However, it is important to note the distribution of compliance and request frequency across autism groups, as adolescents who were less compliant and had mothers who used more requests for behavioral compliance tended to be in the FXS+ASD group rather than the FXS-O group.

Table 6:

Regression Predicting Child Compliance

Regressions	Est.	SE	t-value	p-value	Total R2
Step 1:					0.3812
Intercept	106.26	16.99	6.26	<.0001
Responsivity	0.003	0.036	0.09	0.93
Request Frequency	−1.04	0.31	−3.31	0.002
Request Directness	−0.09	0.23	−0.41	0.69
Step 2:					0.3844
Intercept	111.93	19.79	5.66	<.0001
Request Frequency	−1.05	0.4	−2.61	0.01
CARS Score	−0.29	0.42	−0.7	0.5
PPVT Raw Score	−0.04	0.1	−0.38	0.71
Step 3:					0.3775
Intercept	101.86	6.75	15.09	<.0001
Request Frequency	−1.08	0.24	−4.41	0.0001

Figure 5:

Association between Maternal Request Frequency and Child Compliance by Autism Group

4. Discussion

This series of analyses examined the transactional effects between maternal responsivity and requesting, adolescent compliance, and task demands in mother-son dyads with FXS. Broadly, maternal responsivity and request frequency differed slightly by task with responsivity highest during the puzzle task and request frequency highest during the snack task. Maternal request directness was significantly higher in the iPad activity than puzzle or snack. Adolescent compliance did not differ by task, but instead was predicted by maternal request frequency and associated with maternal responsivity. In a study by O’Nions et al. (2020), parents of children with ASD reported adjusting their own behavior to accommodate their child’s needs and preferences to reduce familial stress and child problem behavior. Specifically, they endorsed adjusting their expectations, planning activities around the child, and taking over parts of the activity that the child was struggling with (O’Nions et al., 2020, p. 393)—parental behaviors which are intended to reduce the likelihood of child problem behaviors. Parental accommodation strategies were associated with parent-reported child noncompliance with routine demands and during social situations, such that parents were more accommodating for children who had higher noncompliance (O’Nions et al., 2020). Similarly, the mothers in our sample may have also adjusted their behavior based on the child’s behavior and support needs. Our findings suggest that mothers are attuned to their child’s abilities and needs as they relate to child behavior and task demands.

4.1. Effects of Task Demands

In this study, mother-son dyads were instructed to complete three different tasks together to examine whether task demands influenced maternal requesting behavior or adolescent compliance. Our intent was that the tasks would vary in familiarity for the dyad and in reward for the adolescent. Although for many adolescents, playing on the iPad (or similar tablet) is a common activity, in this procedure the dyad was instructed to play together, the iPad did not have data or internet access, and the only applications that the dyad could access were pre-installed and pre-selected by the research team. That is, the activities on the iPad only included games that required two players (e.g., Air Hockey, Foosball). Thus, the dyads’ expectations going into the iPad task were unlikely to be met, which could have been distressing to the adolescent. Specifically, our research team noticed that many of our adolescent males with FXS were displeased with the limited capacity of the iPad and were frustrated that they could not connect to the internet to play their preferred games or interact with media by themselves. We believe that this letdown may be related to the slight decrease in compliance between iPad and other tasks. Along with decreased compliance in the iPad activity, mothers were more direct with their requests, perhaps indicating their frustration with their sons’ behavior, or the nature of playing video games.

During the snack task, the dyad was supplied with a list of ingredients and specific step-by-step instructions. As such, it was believed that this was the most structured task and that mothers may have been inclined to provide more guidance and direction to their sons. If mothers perceived that the snack task was harder for their adolescent to complete independently, they may have provided more support and adjusted their requesting behaviors. Similarly, if mothers felt that the social context of making a snack together necessitated a certain degree of behavioral control, they may have adjusted accordingly. In this study, mothers provided more guidance (i.e., gave more requests for behavioral compliance) during the snack task than during the iPad or puzzle tasks, but this difference did not reach statistical significance.

Based on definitions by Warren, Brady and colleagues (Brady et al., 2014; Warren et al., 2010), maternal responsivity in each task was captured and quantified. Responsivity was highest in the puzzle task. Given that responsivity included requests for verbal compliance, comments, and mother restatements of child utterances, it can be hypothesized that responsivity represents the mother’s use of language intended to create a social interaction, rather than to request behavioral compliance. In contrast to the frustration that dyads felt with the iPad and to the relative rigidity of the snack instructions, the puzzle task may have been an activity during which mothers and sons were able to have a casual social interaction marked by reciprocity and contingent responses by the mother. Of course, this is speculative and post-task surveys along with more in-depth analyses of adolescent behavior during the task would be more informative. Nonetheless, subtle differences in maternal directness were observed and may be related to maternal perception of her son’s behavior and task-related focus.

It is important to recognize that the clinical applicability of our findings is limited, because the tasks have little ecological relevance to tasks that adolescents with FXS may encounter in their daily activities across home, educational, and vocational settings. Rather, the findings here are an important step towards understanding the transactional relationships between task demands, maternal requesting and responsivity, and adolescent compliance. Future studies may investigate other tasks and consider individual preferences for various tasks as a factor affecting compliance.

4.2. Maternal Factors

Although adolescent compliance did not vary greatly by task, it was predicted by maternal request frequency and associated with maternal responsivity. Maternal request frequency had a strong inverse relationship with adolescent compliance across tasks. Because we examined correlations, rather than causative relationships, it is unclear whether maternal request frequency led to decreased compliance or whether adolescents who had worse compliance had mothers who, in turn, increased their number of requests for behavioral compliance. If the adolescent was generally compliant, the mother may not have needed to issue very many requests, whereas if an adolescent was not complying, the mother may have given more requests as an attempt to reorient him to the contextual demands and task at hand. The inverse relationship between adolescent compliance and maternal request frequency may also be explained by adolescent temperament and tendencies. Adolescence is marked by increasing autonomy and the desire for independence, which may explain why adolescents who were given more frequent requests for behavioral compliance were less likely to comply. Perhaps, by being overly controlling or commanding, mothers were giving an impression of being overbearing and overdemanding, and adolescents thus responded with more defiance than compliance. It is possible that adolescents who are given the opportunity to try activities independently, or with minimal direction, may be more inclined to comply when necessary. Future examinations of the association between maternal requesting and child compliance would benefit from noting whether the child’s noncompliance was linked with defiance and refusal to complete the action, or whether noncompliance was likely due to another source, such as disagreement with task expectations or failure to comprehend the request.

Maternal responsivity is believed to not only reflect contingent responsiveness and warmth, but also flexibility and willingness to follow the child’s lead. Children with DS are more likely to comply if the maternal request follows the child’s lead (Landry et al., 1994), suggesting the autonomy of the child is an important factor in compliance. Thus, adolescents who are provided with opportunities to develop autonomy and who are provided with age and developmental level appropriate supports that follow their lead, may be more inclined to comply with maternal requests for behavioral compliance. There was a moderate correlation between the total amount of maternal responsivity and child compliance, such that adolescents were more compliant when mothers were more responsive. However, this association was not statistically significant, potentially reflecting issues with sample size. In a follow-up exploratory analysis, we found that proportion of responsive communication turns (i.e., the proportion of comments, recodes, and requests for verbal compliance relative to all communication turns) was significantly, positively associated with child compliance (r=0.57, p=0.0004), indicating that mothers who use a higher proportion of responsive communication turns have children with higher compliance. This finding also suggests that the relative frequency of maternal responsivity, rather than the total count of maternal responsivity actions, is important for child behavioral compliance.

We also explored whether there was an association between compliance and maternal self-completion of the request for behavioral compliance. “No Opportunity” to respond was coded when mothers either completed the request themselves, used hand-over-hand or physical manipulation of the child to complete the task, or did not provide enough time for the child to complete the task before issuing another request for behavioral comply. Mothers with higher instances of “No Opportunity” had children who were less likely to comply. They also had children with lower language scores and higher autism symptom severity. Although our findings suggest an association between adolescent characteristics and maternal self-completion of requests, the direction of effects is unknown and untestable in the current study. It may be that children are less compliant when they are used to their mothers completing the requests themselves, or mothers are more likely to complete the request themselves when their children are generally more noncompliant. Similarly, maternal perception of child ability (language, social, or otherwise) may influence whether she gives the child an opportunity to independently complete the request rather than completing it herself.

The directness of maternal requests for behavior did not impact adolescent compliance. Indirect speech, or speech that is ambiguous and requires interpretation, was once thought to be too difficult for children with language delays to understand, as it can be overly complicated and unclear (Blue, 1981). Although this is logical from a treatment perspective, it has not borne out in recent years or in this study. In our data, compliance did not differ by request directness, suggesting that compliance and the necessary linguistic comprehension to understand and react to the request are not made more difficult in indirect request structures. Our findings are similar to those of Kissine and colleagues (2015) who found that children with ASD complied with indirect requests. However, Bryce and Jahromi (2013) found that children with ASD were less likely to comply with indirect requests than typically developing, language-matched peers. Again, these two studies differ in context, which may help to explain differences in findings. In their study, Gauvain and Perez (2008) reported that mothers of typically developing children who were less compliant used more directive instruction (i.e., direct requests rather than indirect) than mothers of children who were more compliant. When mothers used directives, rather than indirect requests, they often gave clear instructions for action instead of providing scaffolding and guidance as a learning tool for their children (Gauvain & Perez, 2008). As such, use of directives may be less supportive of learning, and may indicate the mother’s lack of trust in the child’s abilities. Examination of maternal perception of child abilities to complete the task at hand and its relationship with maternal requesting behaviors and child compliance may be informative.

4.3. Adolescent Characteristics

We examined the relationship between compliance and autism symptoms and receptive language abilities in male adolescents with FXS and found that both autism symptoms and receptive language ability were significantly, moderately associated with compliance. Given that individuals with FXS who have high autism symptomology or co-morbid ASD diagnoses are more likely to have elevated rates of behavioral problems, the association between increased autism symptoms and reduced compliance is logical.

The association between receptive language and compliance bears further considerations. First, individuals with lower receptive language ability may have difficulty comprehending and interpreting requests for behavioral compliance. This would be supported if our data suggested that compliance was lower following indirect requests (as they are ambiguous and often hard to interpret). Alternatively, individuals with FXS+ASD tend to have greater impairments in receptive language than those with FXS-O (Lewis et al., 2006; Philofsky et al., 2004), suggesting that the association between language and compliance may be somewhat complicated by the association between autism symptomology and compliance.

In their study of typically developing children, Gauvain and Perez (2008) divided a group of 40 mother-child dyads into a compliant and noncompliant group and found that mothers of noncompliant children gave more requests for compliance than mothers of compliant children. As addressed in our results, adolescents who had higher autism symptom scores received more frequent instruction from their mothers than those with lower autism symptoms. Although we cannot determine in our analysis whether higher frequency of maternal requests is due to noncompliance associated with autism symptoms, we suspect that this may be the case. Additionally, autism status, as reported in the figures, was determined through both the CARS and ADOS-2 scores, so correlations between CARS scores and other variables do not completely align with autism status. Yet, in ours and others’ studies, mothers seem to modify the frequency of their requests based, to some degree, on their child’s compliance (Gauvain & Perez, 2008). Importantly, however, the type of instruction that mothers provide may also be linked to child compliance, with mothers providing requests that are more sensitive to the child’s needs following compliance but not following noncompliance (Gauvain & Perez, 2008). In this way, child compliance may serve to regulate maternal instruction. Future investigation of contingent responses following maternal requests and child behavior might clarify how mothers of children with FXS adapt their instruction to meet their child’s unique learning needs.

4.4. Limitations

The results of this study suggest that there are relationships between task and maternal requesting behavior and between maternal requesting behavior and adolescent compliance. However, there are several notable limitations to this study. First, these findings are limited to a small sample that does not represent the racial and ethnic diversity of the general US population. Rather, this sample is predominantly white and highly-educated, with incomes generally > $100,000 annually. Given that maternal education is linked with child language development (Dollaghan et al., 1999), a larger sample with more socioeconomic and educational diversity may yield different results. For instance, mothers who have lower educational attainment may use different types and frequencies of requests which may impact child compliance. In addition, much of the existing literature on mother-child interactions and child compliance is centered on children, rather than adolescents. However, the adolescents with FXS in this study had marked language delays, with expressive and receptive language age equivalency scores around 5.5 years old which is commensurate with other studies of compliance (Bryce & Jahromi, 2013; Ledbetter & Dent, 1988; Rimac, 1985). An additional limitation could be the selection of tasks that mothers and sons completed. We aimed to provide them with tasks that were fun, interactive, and achievable, and with tasks that might reflect typical interactions between the mother and her adolescent son. However, these activities may have been ones that are not normally encountered in daily life, which could lead to different behaviors than usual. Finally, the fixed order of tasks may have led to fatigue effects between tasks.

5. Implications

The findings discussed here may have important ramifications for therapy and treatment of behavioral disorders in adolescents with neurodevelopmental disabilities such as FXS or ASD. A multi-dimensional therapy approach in which parents receive training in how to appropriately guide and teach their children along with identification of the child’s understanding of task demands may yield levels of compliance that are developmentally appropriate and lead to skill acquisition. Indeed, parent training has demonstrated efficacy in reducing disruptive and noncompliant behaviors in children with ASD (Bearss et al., 2015). Additionally, adults may be able to modify their requesting behaviors and/or the environment/task to improve the child’s likelihood of complying. For instance, in educational settings, high probability command sequences have been shown to improve compliance in children with disabilities (Austin & Agar, 2005). Identification of antecedents to noncompliant behaviors may also be a promising therapeutic target to improve compliance (Majdalany et al., 2017). Compliance with maternal requests is a valuable learning tool, as mothers teach their children about social expectations and norms through this interaction. However, compliance is also important in the mother-child relationship because child noncompliance and presence of child challenging behaviors have been associated with poorer maternal mental health in FXS and other neurodevelopmental disabilities (Baker et al., 2012; Fielding-Gebhardt et al., 2020; Kochanska & Aksan, 1995; Smith et al., 2008). Thus, it is important to identify factors that promote compliance and determine what therapeutic and intervention approaches could target compliance.

What this paper adds?

This study examined the transactional relationships between maternal request frequency, request directness, and responsivity, child compliance, language, and autism symptoms, and task demands in adolescent males with fragile X syndrome (FXS). This paper contributes to several bodies of existing literature. First, while previous studies have focused on autism or developmental language disorder, this study explores maternal requesting and its effect on child compliance in FXS. Importantly, when mothers used requests more often, their adolescent sons were less likely to comply. Second, this study considers the frequency and effectiveness of indirect and direct requests for eliciting compliance adding to a body of literature on speech acts and the role of indirect speech in social communication. Children with FXS were compliant with maternal requests regardless of the directness of the request structure. Finally, this study considers different task demands during mother-child interactions and how this might impact maternal requesting behaviors and child compliance, with the conclusion that certain tasks may predispose mothers to alter their requesting behaviors, but task demands do not seem to influence child behavior. Together, these findings provide caregivers and interventionists with knowledge that may be useful in their interactions with adolescent males with FXS, which could result in more effective teaching strategies for this population.

Highlights:

• Mothers of adolescent males with FXS and higher ASD symptoms were less responsive and gave more requests for behavioral compliance.

• Frequency of maternal requests predicted adolescent compliance, with more requests yielding lower compliance.