Factors that Influence the Daily Living Skills of Autistic Adults: The Importance of Opportunity

Abstract

Purpose.

While existing literature has demonstrated that Daily Living Skills (DLS) performance of autistic individuals is lower than what is expected of their age and cognitive abilities, limited studies have examined DLS in autistic adults. This study aimed to understand the influence of intellectual function (IQ) and contextual factors (i.e., provision of opportunities) on autistic individuals’ DLS performance.

Methods.

Participants included 33 autistic individuals ranging in age from 16 to 35 years. Their caregivers were administered the Vineland Adaptive Behavior Scales, 3^rd edition’s (Vineland-3) caregiver interview form. A novel coding system was developed to capture the frequency of reasons for participants’ non-performance of DLS tasks, based on caregiver’s report. “Target” scores reflecting expected possible score if reasons for nonperformance could be addressed were computed.

Results.

Qualitative analysis of parental responses indicated that, for adults with average or higher IQ, lack of opportunity to learn and/or implement the skill was the most frequent reason for not performing DLS. Lack of opportunity was also the second most common reason provided for adults with NVIQ below 85, following cognitive ability. Taking into account reasons for nonperformance, “Target” scores were, on average, 7.65 points higher for the NVIQ ≥85 group.

Conclusion.

These findings highlight a need for multi-dimensional assessment to go beyond individual strengths and difficulties to also include contextual factors that may influence adults’ skill acquisition and performance. It is essential that clinicians ensure that adequate opportunities for learning and performance are available to promote acquisition of important DLS.

Prior research has focused on broadly defined outcomes of autistic adults, often finding that autistic adults struggle to achieve traditional markers of independence (Howlin et al. 2004, Farley et al., 2009; Matson et al., 2009). More research is needed on potentially malleable factors that can inform development of supports for autistic adults. One such area explored in the extant literature has been daily living skills (DLS; Bal et al., 2015; Matthews et al., 2017; Perry et al. 2009; Pugliese et al., 2015; Tillman et al., 2019). DLS are a component of overall adaptive functioning, which include activities such as personal care, hygiene, household management (e.g., cooking, maintenance), occupational skills, skills needed for community navigation (e.g., awareness of safety rules, transportation) and basic financial management. Research suggests that DLS are strongly correlated with independent living, employment, educational attainments, community participation and social life in autistic adults (Chan et al., 2021; Chan et al., 2018; Farley et al., 2009; Pillay et al., 2022). Better DLS are also associated with the well-being of the autistic adults (Bishop-Fitzpatrick et al., 2016) and their parents (Marsack-Topolewski et al., 2021; Marsack-Topolewski, 2022).

Longitudinal studies suggest that autistic adolescents showed an increase in DLS before leaving the school system, but a plateau or decline after high school exit. DLS trajectories predicted participation in postsecondary education, even after controlling for IQ (Clarke et al., 2020). Profiles of DLS subdomains (e.g., personal, domestic, community) have not been widely studied in autistic adolescents or adults, though there is some evidence for slight difference in trajectories (Bal et al., 2015). In a sample of 151 autistic adolescents, community skills were identified as a strength relative to domestic and personal skills, though still significantly lower than chronological expectations (Glover et al., 2022). Their microanalysis of items identified domestic skills as an area with the fewest mastered or acquired skills, possibly suggesting differences in factors affecting acquisition and/or mastery of skills across subdomains. It may, for example, reflect strengths or abilities in personal and community subdomains, or be indicative of other factors (e.g., difficulty self-initiating or few opportunities to learn and implement skills) having a different influence on domestic skills.

Considering their importance to independent functioning and quality of life, many interventions for adults with Intellectual Disability (ID), with and without autism, focus on DLS (Wong et al., 2015). For example, evidence-based strategies have been used to target community skills (e.g., street crossing, Harriage et al. 2016; navigating public transport, Lubin & Feeley, 2016; Price et al. 2018) and domestic skills (e.g., food preparation; Tekin-Iftar & Birkan, 2010; Chazin et al. 2017). In a recent pilot of Surviving and Thriving in the Real World (STRW; Duncan et al., 2018), a group designed to help autistic adolescents acquire and master DLS, the STRW group showed greater increases in personal care (e.g., hygiene) than the waitlist control group (Duncan et al., 2022). The authors speculated that may have been due to STRW targeting hygiene routines; however, it may also suggest that personal skills are more amenable to brief interventions and more complex domestic or community skills may take longer to obtain mastery. Other factors may also account for differences in skill acquisition and/or mastery across subdomains.

Importantly, many measures of DLS, such as the Vineland Adaptive Behavior Scales (Sparrow et al., 2016), emphasize distinguishing between one’s regular performance of task and their actual ability to carry out those tasks. In other words, measuring whether one does exhibit the skill consistently when required, without prompts or supports, rather than whether one can perform the skill. For example, one may know how to put clean laundry away, but does not do so without repeated prompts. Additionally, information about environmental expectations and availability of opportunities to perform skills are important to contextualize both skill levels and performance. Using the laundry example, a parent may do laundry for the entire household and therefore there is not an expectation that the person put away their own laundry, or even an opportunity to do so. Differentiation between factors affecting regular performance (i.e., ability vs. other factors) is essential to identify targets for intervention. The current study aimed to better understand factors that affect regular performance of DLS.

On the Vineland-3, items are scored from 0 to 2, based on whether the person Never, Sometimes or Usually performs the behavior without help or prompting. Despite the emphasis on daily performance, rather than ability, the scoring does not distinguish between inability to perform or the myriad of other reasons that may affect whether someone performs DLS tasks. For example, the Vineland authors highlight how other individual factors may impede daily performance, such as not being physically able to perform the behaviors (Sparrow et al., 2016). Cognitive impairment is widely known to be associated with lower levels of DLS performance in individuals with (Bal et al., 2015; Klin et al., 2007; Smith et al., 2012; Tillman et al., 2019; Clarke et al., 2021) and without autism (Klin et al., 2007; Perry et al., 2005). In autistic individuals, however, there is often a discrepancy between IQ and DLS performance. For example, in a sample of 417 autistic adolescents without cognitive impairment, Duncan and Bishop (2015) observed that mean IQ was at least one standard deviation higher than Vineland DLS scores. Beyond cognitive level, difficulties in executive function (i.e., organize, plan and execute tasks) are also associated with lower DLS in children, adolescents and adults with autism, perhaps explaining some of the DLS-IQ discrepancy (Ashwood et al. 2015; Granader et al., 2014; Pugliese et al., 2015; Wallace et al., 2016; Williams et al., 2014).

In a recent qualitative study, parents and autistic adults reported that the autistic adult knew how to complete daily living activities (e.g., laundry, cleaning), but a lack of intrinsic motivation to engage in these activities and executive function difficulties, particularly self-initiation or self-management of tasks, were obstacles to their completion (Matthews et al., 2021). Alternatively, what may be perceived by parents as a lack of motivation may actually reflect limited awareness of expectations or need to perform behaviors (Sparrow et al., 2016), including not anticipating responsibility of a task in the absence of explicit instructions (i.e., initiation skills, Hume et al., 2009, 2014) or, not realizing that expectations may generalize beyond a specific situation or timeframe (Duncan & Bishop, 2015; Sabat et al., 2020). Sparrow et al., (2016) also suggested that not being explicitly taught, or not being expected or allowed to perform tasks, may also affect performance of DLS. Each of these possibilities highlight the less-explored effect of opportunity on learning or performing daily living activities.

As teenagers and adults get older, there may be a shift in caregivers’ expectations of performance (Bal et al., 2015; Kirby, 2016), which could result in fewer opportunities for the individual to learn, practice or perform the skill. Consistent with this, studies have found that autistic teenagers reported wanting more independence and took pride in having agency rather than supports (Humphrey & Lewis, 2008; Rossetti et al., 2008). Cheak-Zamora et al. (2015) reported that most autistic adolescents agreed that their caregivers controlled most of their activities and schedules. This may apply for autistic young adults who often live at home with their caregivers (Farley et al., 2009; Wehman et al., 2014), as it is likely that at least some of their daily living needs (e.g., house chores, financial responsibilities) may be taken care of by parents or shared with other household members. Again, we see where environmental expectations (i.e., what role the adult is viewed as having in the household), as well as the number and regularity of opportunities, could limit both learning and performance of daily living skills.

In sum, there are many factors that may influence DLS performance. While some factors have been studied in autistic adults (i.e., IQ and executive function), the present study aims to fill a gap in understanding whether caregivers view self-initiation and provision of opportunities as factors affecting DLS performance in autistic young adults. Caregiver responses during the Vineland-3 interview were coded to evaluate whether ability, difficulty self-initiating or opportunity affected participants’ DLS performance. Coded responses were assigned scores to generate a “target” DLS score for participants in order to gain insight into whether supports and/or opportunity accounted for a meaningful difference in performance. The relationship between the codes and nonverbal IQ (NVIQ) was also examined. Finally, the gap between cognitive level and DLS was examined for both the original and the “target” DLS scores.

Methods

Participants

Between July 2019 and February 2021, caregivers of 33 participants completed the Vineland-3 as part of clinical diagnostic evaluations (n=9), admission to a vocational support program (n=6) or research (n=18). All eligible participants had a diagnosis of autism spectrum disorder, confirmed by direct assessment supervised by a licensed clinical psychologist, including a minimum of ADOS-2 (Lord et al., 2012), cognitive assessment, and the Vineland-3. See Table 1 for participants characteristics.

Table 1.

Participants Demographics and Characteristics

	All
N	33
Age (M, SD)	22.2 (4.3)
Female (%)	12
Race (%)  African-American/Black  Asian/Pacific Islander  White	6.06 21.2 72.7
Ever lived alone (%) NVIQ (M, SD)	6.06 88.9 (29)
VIQ (M, SD) Intellectual Disability (%)	89.8 (31.3) 9 (27%)

Note. M= mean; SD= standard deviation. NVIQ= nonverbal IQ. VIQ= verbal IQ. Intellectual Disability reflects clinical diagnosis taking into account cognitive and Vineland-3 scores.

Measures

DLS

The Vineland-3 Comprehensive Interview form was used. It includes three domains (Communication, DLS and Socialization). The DLS domain includes the Personal (PERS; self-care skills ranging from basic feeding skills to elements of dressing and hygiene and health management), Domestic (DOM; e.g., household chores and safety) and Community (COM; skills that are important to navigating situations outside of home, such as travelling and managing money) subdomains. Vineland-3 items are scored on a 3-point scale of 2, 1 or 0 point. Raw scores for each subdomain are summed and converted to a standard v-scale (VS) score (M=15, SD=3) and age equivalent (AE); the three subdomains are combined and converted to yield a DLS domain standard score (M=100, SD=15).

To examine factors for nonperformance of DLS, for each DLS item scored 0, the examiner followed-up with additional questions to identify the reasons that the participant did not exhibit the skill. The follow-up questions began with open-ended questions (e.g., Tell me more, or Can you explain why they are not doing it?), which tended to elicit responses that were sufficient for coding purposes. If needed, additional close-ended questions were asked to help distinguish between ability, difficulty self-initiating and opportunity (e.g., Would they do it correctly if you asked them to?). The impact of factors was considered in the computation of a “Target” score described below.

Cognitive Functioning

NVIQ scores were derived using the following hierarchy of measures: the Wechsler Adult Intelligence Test, Fourth Edition (Wechsler, 2008), Wechsler Abbreviated Scale of Intelligence, Second Edition (Wechsler, 2011), Differential Abilities Scale, Second Edition (Elliott, 2007), Mullen Scales of Early Learning (Mullen, 1995). Instruments were selected according to the ability of the participant. As participants were 18 years or older (except one 16-year-old), when the Mullen or DAS-II was used out of standardized age range, ratio IQs were computed from derived age equivalents (Bishop, Farmer & Thurm; 2015). Participants were split into two IQ groups, those with NVIQ at or above 85 (NVIQ ≥85; n=23) and those with NVIQ under 85 (NVIQ <85; n=10; n=3 participants had NVIQ between 70–80 and all had FSIQ below 70).

Coding Categories and Procedures

Development of Coding Scheme

Taking a directed content analysis approach (Hsieh & Shannon, 2005), a coding scheme was developed by the first author to capture the reason that respondents may endorse a rating of zero, i.e., “Never or seldom”. Categories were based on the reasons that the Vineland-3 authors (Sparrow et al., 2016) suggested to capture if the examinee (a) has not learned the behaviors (e.g., how to use the stove for cooking), (b) is not physically able to perform the behaviors (e.g., wash dishes), (c) is not expected or allowed to perform the behavior (e.g., cooking or using tools), (d) can perform the behavior, but does not perform (e.g., puts laundry away) and (e) is unaware of the need to perform the behavior. Using one participant interview, the first author trained a research assistant (RA) on the coding procedures with a combination of written instructions, in-vivo coding and discussions. For the second participant interview, both author and RA independently coded one item at a time and compared their codes immediately. Feedback from the RA on both interviews was used to modify reason codes and refine definition of codes. The coding scheme was then finalized. The final coding scheme comprised of three summary codes and five reason codes (see Table 2).

Table 2.

Categorization and Definition for Summary and Reason Codes

Summary Codes	Reason Codes	Abbreviation
1) Does not do	(a) Reason Unknown/ Unclassified OR Confounding reasons provided	Unspecified
	(b) Unless reminded OR Does not perform even when reminded	Able- reminded
2) Not Able to do	(c) Not able to do—cognitive (intellectual, conceptual)	Unable- cognitive
	(d) Not able to do— non-cognitive	Unable- other
3) No Chance	(e) Never had chance to assess performance OR Never had regular chance to perform skill	No Chance

Summary and Reason Codes

Three summary codes were created to capture reasons for non-performance, namely Does not do, Not Able to do and No chance. The Does Not Do summary code indicates that the participant has had opportunity to learn and perform the skill, but did not independently perform the skill regularly, or whenever required (i.e., capturing those who did not self-initiate). This category is further split into two reason codes: (a) Reasons Unknown OR Unclassified, in which the informant’s response was insufficient for coding purposes (e.g., response does not provide reasons for not performing task regularly, or follow-up questions not asked), and (b) Unless reminded OR Does not perform even when reminded, in which the participant has demonstrated clear ability in performing task, but does not do so unless prompted, or even when prompted. This reason code was specifically allocated to responses indicating that the participant was able to perform the skill, but required prompts to evoke the target behavior.

The Not Able to Do summary code indicates that the participant has had opportunity to learn or perform the skill, but has not mastered the skill (i.e., capturing those whose caregivers reported ability level as a reason for non-performance). This category also applies to participants reported as not having the pre-requisite skills to learn a DLS task and hence, there was no attempt to teach or probe target skill (e.g., participant not taught to use a stove given their difficulty with learning to use the microwave).This category is further split into two reason codes: (c) Not Able to Do - Cognitive Reasons, reflecting parents’ perceptions that the participant is not able to master the skill due to level of intellectual disability, and (d) Not Able to Do – Non-Cognitive Reasons, indicating the participant fails to master skills due to other non-cognitive factors (e.g., weak muscle tone, sensory sensitivity, food aversion, consistent difficulty following multi-step instructions, anxiety and medical conditions).

The No Chance summary code functions as its own reason code. No Chance indicates that the participant either did not have any opportunity to learn or perform the skill (i.e., ability to perform skill is unknown), or was not provided with consistent opportunities or requirements to perform skill (i.e., participant had shown ability to perform skill on at least one occasion, but the task was not made a responsibility of the participant, or there were limited chances to perform it because someone else did it for them). This code was also used to capture if the participant had not been exposed, or required to learn about certain concepts (e.g., rights to medical records).

Coding Procedures

Parent responses for all 0-scored DLS items before the Vineland-3 ceiling were transcribed and double-coded by the first author and the second author, who was blind to the participants’ diagnoses, IQ and ages, to reduce potential for rater bias. After co-coding the first two participants for training, coders coded 2–4 transcripts independently and calculated exact agreement after each set. Exact agreement for each interview ranged from 85.7%- 100% for summary codes and 75%−100% for reason codes. Across interviews, interrater reliability (% agreement and Cohen’s kappa coefficient, k, was highest for Personal subdomain (summary codes: 82.1%, k=.69; reason codes: 76.2%, k=.80) and lowest for the Community subdomain (summary codes: 91.1%, k =.84; reason codes: 87.4%, k=.89). All disagreements were discussed between the coders and consensus codes were used for analyses.

Target Scores

To evaluate the potential impact of different factors on DLS assessment, the five Reason codes were assigned point values, based on the estimated ability to perform the task if circumstances for nonperformance were taken into account (see Table 3). For each subdomain, items scored 0 based on standard interview procedures were rescored (e.g., if the reason was Able-reminded, the item was given 2 points; if the reason was No Chance, the item was given 1 point, etc.). The recoded items were then used to recompute the raw subdomain score and derive Target VS, AE and DLS scores. These Target scores were intended to reflect the expected possible score a participant could achieve if addressable reasons for nonperformance could be addressed through supports. In this way, the Target score may provide a measurable treatment target and outcome for future DLS interventions

Table 3.

Assignment of additional points to subdomain raw score based on estimated mastery ability

Reason Code	Estimated ability to perform skill without assistance	Points added
Able- reminded	Clear ability to perform when asked to	2
Unable- other No Chance	Might be able to perform skill if asked to or might be able to master skill if taught	1
Unable- cognitive Unspecified	Skill mastery is perceived to be unlikely OR information collected is insufficient to determine possibility of mastering skill	0

Note. Points added refers to the number of raw score points added to the subdomain raw total for each 0-score item with the corresponding reason code.

Analyses

Descriptive statistics were used to describe the frequency with which Reason codes are reported for each DLS subdomain and the proportion of 0-scored items that each code accounted for in each subdomain; each are presented separately for individuals with NVIQ ≥85 vs <85 (Table 4). The number of items coded for each reason was computed for each participant (Table 5) and used to compute percent of overall 0-score items (i.e., number of items coded for each reason divided by the number of 0-score items; see Table 6). Independent Samples T-tests (i.e., NVIQ ≥85 vs <85) were used to compare the average number of items assigned each reason code for each domain, as well as the percent of items assigned each reason code per participant. Paired t-tests were used to evaluate the difference between standard and Target scores within IQ groups and independent samples t-tests used to compare IQ group differences in standard and Target scores.

Table 4.

Per participant frequency of 0-score items and reason codes across DLS subdomains

		DLS	PERS	DOM	COM
		M (SD)	M (SD)	M (SD)	M (SD)
NVIQ < 85 (n=10)	Number items with 0-score	17.30 (6.65)	4.70 (3.30)	4.60 (2.72)	8.00 (3.23)d
	[Range]	[9–29]	[0–13]	[0–9]	[3–14]
	Able - reminded	0.80 (1.32)	0.30 (0.48)	0.30 (0.67)	0.20 (0.42)
	No chance	6.10 (4.48)	1.20 (1.75)	1.70 (1.34)	3.20 (2.25)
	Unable - other	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)
	Unable - cognitive	7.80 (7.2)a	2.60 (3.50)b	1.70 (1.70)c	3.50 (3.24)e
	Unspecified	2.40 (2.67)	0.60 (1.08)	0.80 (1.23)	1.00 (1.25)
NVIQ ≥ 85 (n=23)	Number items with 0-score	13.17 (6.07)	3.61 (1.92)	4.35 (3.28)	5.22 (2.70)d
	[Range]	[4–25]	[0–9]	[0–14]	[0–11]
	Able – reminded	2.17 (2.64)	0.96 (1.43)	1.00 (1.68)	0.22 (0.42)
	No chance	7.48 (4.04)	1.52 (1.24)	2.52 (1.83)	3.43 (2.66)
	Unable – other	0.48 (0.99)	0.30 (0.76)	0.04 (0.21)	0.13 (0.46)
	Unable – cognitive	0.87 (1.36)a	0.13 (0.46)b	0.17 (0.39)c	0.57 (0.95)e
	Unspecified	2.09 (2.19)	0.70 (1.11)	0.52 (0.67)	0.87 (1.1)

Note. Means reflect the average number of 0-score items and reason codes per participant. PERS= Personal subdomain, DOM= Domestic Subdomain, COM= Community Subdomain. Same-lettered superscripts indicate differences between NVIQ <85 vs ≥ 85

^{a, c, e, =}p≤.001

^b=p ≤.01

^d=p≤.05

Table 5.

Gains in DLS domain and subdomain scores after allocation of points

	Mean Gain in points (SD)	DLS	PERS	DOM	COM
NVIQ < 85 (n=10)	Raw	-	1.80 (1.68)	2.20 (2.20)	3.50 (2.37)
	V-Scale	-	0.80 (1.13)	0.20 (0.42)b	0.30 (0.48)d
	Age Equivalent (in years)	-	1.83 (2.79)	0.54 (0.65)c	0.58 (0.60)
	DLS SS	1.90 (2.73)a	-	-	-
NVIQ ≥ 85 (n=23)	Raw	-	3.74 (2.82)	4.65 (4.44)	4.13 (2.97)
	V-Scale	-	1.61 (1.12)	1.48 (1.34)b	0.96 (0.71)d
	Age Equivalent (in years)	-	3.12 (2.39)	1.94 (1.28)c	1.10 (0.72)
	DLS SS	7.65 (3.72)a

Note. PERS= Personal subdomain, DOM= Domestic Subdomain, COM= Community Subdomain. Same-lettered superscripts indicate differences between NVIQ <85 vs ≥ 85

^{a, c =}p≤.005

^{b, d =}p≤.05

Table 6.

The proportion of reason codes (per participant) by NVIQ

		NVIQ (N=33)
		<85 (n=10)	≥85 (n=23)
		M% (SD)	M% (SD)	d
Reason Codes	Able - reminded	3.3 (5.5)	14.3 (16.0)	.79**
	No Chance	39.9 (28.6)	59.0 (21.7)	.80*
	Unable -other	0.0 (0.0)	3.0 (6.0)	.59*
	Unable -cognitive	43.4 (30.4)	4.8 (8.4)	−2.17**
	Unspecified	13.4 (18.3)	19.1 (16.0)	.34

Note.

^**p≤.01, two-tailed.

^*p≤.05, two-tailed. NVIQ = Nonverbal IQ;

Community involvement

The focus of this study was informed by the authors’ clinical experiences, in which parents and autistic adults shared their views on the importance of creating opportunities to learn and practice skills in adulthood. Community members were not directly involved in the design or execution of this study.

Results

Reason code distributions across DLS subdomains

The item-level distribution of reason codes for each subdomain is provided in Tables S1–S3. The number of 0-score items ranged from 4 to 29 across participants, with an average of 14.4 (SD=6.4) items per participant. Table 4 shows the breakdown of codes, by NVIQ group. Unable – cognitive was the most frequent reason code for the NVIQ < 85 group, representing an average of 7.8 items for the overall DLS domain, most commonly explaining 0-scores on the Community subdomain (M=3.5 items). No Chance was the most frequent reason code for the NVIQ >85 group, accounting for 7.48 items across the DLS domain (M= 1.52–3.43 items across subdomains). No Chance was also the second most frequently reported reason for non-performance in the NVIQ <85 group (M = 6.1 items in the DLS domain). All other reasons were coded relatively infrequently, accounting for an average of less than 1 item per participant for each of the subdomains.

Target DLS scores

The relationship between number of items with 0-score (Table 5) and the gain in raw points was not linear, given that the assignment of scores based on the reason codes varied from 0 to 2 (Table 3). As shown in Table 5, across subdomains, addition of points resulted in an increase of approximately 1 to 3 raw scores points for the NVIQ<85 group and 3 to 4 raw score points for the NVIQ ≥ 85 group, which translated into increases, on average of 0.20 – 1.61 VS points and 0.54 to 3.12 years in age equivalents. Notably, despite comparable raw score point gains across NVIQ groups, the > 85 group showed significantly larger VS and AE gains. As shown in Figure 1, after assigning raw points based on reason codes, seven participants (21.2%) did not show differences in their standard or Target DLS domain scores. After point applications, the NVIQ≥85 group gained an average of 7.65 DLS points (t(22)=9.85, p<.001, d=3.73), while the NVIQ<85 gained an average of 1.90 (t(9)=2.20, p=0.06, d=2.73).

Figure 1.

Differences in Standard Scores between DLS and Targeted DLS.

As shown in Table 6, the NVIQ≥85 group had a significantly higher percentage of items coded as No Chance, Able-reminded and Unable-other, compared to those with NVIQ<85. In contrast, the NVIQ<85 group had a higher percentage of items coded as Unable-cognitive. No Chance accounted for the largest proportion of reason codes per participant in the NVIQ ≥85 group (M=59.0% of items), whereas Unable-cognitive was the most frequent for those with NVIQ<85 (M=43.4% of items).

Discrepancies between NVIQ and DLS

Participants in both NVIQ groups showed a clear discrepancy between their NVIQ and DLS standard scores. As shown in Figure 2, DLS performance was 34.4 points lower than NVIQ in the NVIQ≥85 group (p<.001, d=1.63). In contrast, the discrepancy was only 12.2 points in the NVIQ<85 group (p<.001, d=1.60). After assigning points according to reason codes, the NVIQ-Target DLS gap was somewhat attenuated, but remained large for both NVIQ≥85 (M= 26.8, p<.001, d=1.33) and NVIQ<85 (M= 10.1, p=.04, d=.78) groups.

Figure 2.

Discrepancies between NVIQ vs. VABS-3 DLS and Targeted DLS. Note. *** p<.001, two-tailed. * p<.05, two tailed.

Discussion

Previous studies identified multiple factors that are associated with adaptive skills in autistic individuals, including intellectual ability, age and executive function (Bal et al., 2015; Duncan & Bishop, 2015; Kanne et al., 2011; Ashwood et al. 2015; Granader et al., 2014; Hume et al., 2009; Pugliese et al., 2016; Wallace et al., 2016; Williams et al., 2014). However, limited studies have examined contextual factors, such as the availability of opportunities to learn or perform daily activities (Cheak- Zamora et al., 2015; Humphrey & Lewis, 2008; McCollum et. al, 2016; Rossetti et al., 2008; Wehman et al., 2014). In the current study, qualitative content coding indicated that No Chance was the most common reason for non-performance of DLS across subdomains reflecting limited opportunities for learning or performing DLS. No Chance was more commonly endorsed in individuals with average NVIQ (59% of participants’ 0-score items), compared to adults with lower IQ (40%).

Why autistic adults with higher cognitive appeared to have fewer opportunities to learn or regularly perform these skills may be explained in multiple ways. First, due to the Vineland’s basal and ceiling rules the types of skills that parents are being asked often vary by IQ. For example, for adults with ID, caregivers may be more proactive in creating opportunities to learn and gain independence in daily hygiene routines (e.g., hand washing, showering) assessed on the Personal subdomain. In contrast, for adults who are independent in hygiene, parents may be less inclined to teach or consider more advanced tasks captured in the subdomain, particularly those that require multiple EF skills that may be more difficult for the adult, such as monitoring medication or going to a doctor when needed.

Second, most participants had never lived alone at the time of interview (93.9%). In contrast, the U.S. Census Bureau’s data on America’s Families and Living Arrangement reported that in 2012, 44.1% of young adults ages 18 to 34 moved out of their parents’ homes (Vespa et al., 2013). Living with parents very likely affected opportunities for autistic adults in our sample. Many continued to rely on caregivers for transportation, meal preparation, and monitoring financial and medical needs (which resulted in 0-scores on related items), despite reportedly having the abilities to learn these skills. Caregivers frequently cited convenience as a factor against delegating tasks to their adult children. For example, caregivers tended to cook for the entire family regularly, rendering limited opportunities for the adults to learn, practice or exhibit skills assessed on the Domestic subdomain, such as preparing ingredients, using the stove and cooking full meals. It is possible that families intentionally complete these activities for individuals with EF difficulties to limit daily EF demands (e.g., cooking after work when already tired) or allow them more time to focus on other activities seen as higher priority (e.g., school work). In such cases, well-intentioned (and even necessary) daily supports may be inadvertently limiting learning opportunities and alternative opportunities for teaching that fit with individual and family needs should be identified.

While one could argue that the autistic adults’ lower adaptive functioning may be a contributor, not the outcome, of their living with their families, interventions targeting DLS require careful consideration of if and how contextual factors may affect DLS learning and performance. It is important to note that reduced opportunity in the home should not be interpreted as always suggesting the onus should also be placed on the family to teach or create additional opportunities. “Convenience” may sometimes reflect situations where someone in the family performs an activity (e.g., meal preparation or laundry) for everyone in the household. It may, however, also signify a need for greater levels of support to help the adult complete that activity, which is not realistic to embed in daily family life given other situational constraints (e.g., parental work schedules, supervision needs, availability of specific resources). Thus, exploring with the family realistic options for learning and developing skills will be important. It may be that clarifying realistic expectations (e.g., helping with one meal on the weekend or preparing an afternoon snack) may be especially helpful to mitigate practical concerns. There may also be opportunities outside the home to foster skill development, such as taking a community cooking class accompanied by an aide or requesting they work on these skills in their day program. Reframing these options as opportunities for skill and relationship building may provide additional encouragement (e.g., having the family prepare a weekend meal or attend a cooking class with a neighbor provides both skill building and social opportunity).

It is notable that several caregivers remarked during the interviews that the DLS items highlighted skills that they should perhaps consider teaching and creating opportunities for their adult child to perform more often to promote independence. This suggests a possible need for clinicians who work with autistic adolescents and transition-age adults and their families to provide psychoeducation on the skills that may promote daily independence as the individual gets older. This could be done in school, such as in the context of home economics-type classes. As such classes are often offered as electives, care may be warranted to ensure that autistic students are considering more practical classes and that their electives are not subsumed by other courses or services aimed at supporting academic development.

It is important that those interested in teaching autistic adolescents and adults these skills be aware that there are a range of evidence-based strategies (e.g., video modeling and visual supports, task analysis, chaining, reinforcement) and that have been found to support direct teaching and acquisition of daily living skills in adolescents and adults (National Autism Center, 2015; Wong et al., 2015). There is also emerging evidence for new strategies, such as incorporating technology (Hrabal et al., 2022) and programs, such as Surviving and Thriving in the Real World (Duncan et al., 2018, 2021). If self-initiation or other aspects of executive function are affecting adaptive performance, interventions could range from implementing simple reward-systems or clearly laid out expectations for different DLS to therapeutic supports, such as organizational skills training (OST; Gallagher et al., 2014), behavioral activation, or family therapy if the amotivation is maintained by parent-child conflicts (Malti & Buchmann, 2010; Marmorstein et al., 2004). Notably, low levels of self-initiation should not be assumed to reflect the autistic adults’ motivation or avoidance of non-preferred tasks; exploration of possible factors impeding self-initiation (e.g., executive functioning difficulties, lack of clear expectations, depression, etc.) are critical to inform appropriate supports.

The addition of points based on reason codes resulted in mean DLS standard score gain of 7.65 points for those in the NVIQ ≥85 group, and only 1.9 points for adult with NVIQ <85. This gain resulted in the NVIQ ≥85 group moving from a Mean DLS domain score of 70 (on the cusp of the range consistent with mild intellectual disability) to 77, which falls in the borderline range of function. This suggests that lack of opportunities to learn, practice and implement DLS may account for some of the discrepancy between adaptive function and cognitive ability in adults with average or higher cognitive abilities. Although adding points to scores (based on assumptions of performance given opportunity) did not fully close the NVIQ-DLS gap, it may be that focusing on the gap between “can do” and “does do” is a useful first step in intervention to promote engagement and self-efficacy. Indeed, we propose that the so-called “Target scores” may serve as a potential goal in the measurement of treatment-related outcomes. In other words, when using subdomain scores or age equivalents as pre- and post-treatment measures, clinicians may choose to monitor improvement on individual items and consider the Target score (i.e., expected ability if given adequate opportunity) as a marker for treatment-related success, rather than striving for a score in the “adequate” range. Notably, these score adjustments would not be appropriate in the absence of actual intervention to test whether the presumed reason for nonperformance (e.g., lack of opportunity, difficulty self-initiating) resulted in skill acquisition and application.

Limitations and Future Directions

The present study was restricted to a small convenience sample of caregivers completing the Vineland-3. While subgroups analyses must particularly be interpreted with caution considering the very small subgroups (e.g., n=10 in the NVIQ < 85 group), these data nonetheless highlighted opportunity as an important factor affecting DLS scores which has practical clinical implications. Notably, a subset of interviews occurred during the COVID-19 pandemic. While the pandemic’s impact on each participant’s performance was not systematically recorded, the majority post-March 2020 were conducted after October 2020. Although opportunities to learn some skills (e.g., navigating the community) may have been limited by lockdown circumstances, given the age of this sample (18+, except one participant), the pandemic is unlikely to completely account for the high rates of “no chance,” as participants would have been old enough to be taught many of these skills prior to the pandemic. A few interviews seem to suggest that some adults had more opportunities (e.g., for cooking), due to their being home more often with less time pressure in their daily routine. Thus, pandemic circumstances likely served to both limit some and enhance other opportunities for a small number of individuals in this sample.

Despite its significance suggested by the Vineland authors (Sparrow et al., 2016), to our knowledge this is the first study to approximate the caregiver’s estimate of adult’s potential, using further querying to derive “Target” Vineland-3 scores. Considering that Vineland scores are standardized in non-clinical samples, it may be that these point assignments are “over-correcting.” In other words, the examined factors, particularly lack of opportunities, may also affect DLS performance in neurotypical adults, which would be accounted for in the standardization process. It would be informative for future studies to directly explore how household expectations and opportunities compare for neurotypical and autistic adolescents and adults, particularly those living with parents or other caregivers, to inform potential supports to promote DLS acquisition.

Future studies are also needed to compare other assessments of adaptive behavior, such as the Adaptive Behavior Assessment System (ABAS-3; Harrison & Oakland, 2015) to the Vineland-3. The ABAS-3 has a 4-point scale, which captures if a person is unable to perform the behavior (i.e., 0) and how often they perform behaviors they are able to do (i.e., 1–3, (Never or almost never) to Always (or almost always), potentially offering an easier way to identify “can do” vs. “does do.” Comparisons of the ABAS-3 and VABS-3 in autistic adolescents suggest the VABS-3 scores may be lower than ABAS-3 on comparable domains (Tamm et al., 2022), but research is needed to explore what might explain these differences and whether similar patterns are observed in adults. Better understanding of ABAS-3 and VABS-3 would also expand the options for direct inclusion of adults in the assessment of their daily living skills, as the ABAS-3 is also designed for use as a self-report. Indeed, regardless of instrument used, while not a part of the standard administration of the Vineland-3 (and not included in the present study), the adult should be consulted in the evaluative process, particularly when considering potential barriers to skill learning and developing supports to foster development.

It is also important to note that the Vineland-3 was developed and standardized in the United States; scores may also be affected by cultural factors that affect environmental expectations and opportunities, such as differences in roles or routines. Within the United States, the use of the Vineland-3 to inform individualized intervention planning should contextualize family’s cultures, practices and preferences, instead of relying solely on the tasks in Vineland-3. Many countries have adapted the instrument and updated standardized scores for earlier Vineland versions (e.g., Touil et al., 2021; Katsiana et al., 2022) and the Vineland-3 items have more general wording that allows more flexible interpretation of items to account for cultural differences (e.g., responds politely, rather than responds with “thank you”). Nonetheless, use of the Vineland-3 with different cultural groups within America typically relies on US-based norms and thus should be used with cultural humility.

Conclusion

Research suggests that better DLS are associated with increased independence in adulthood. Consistent with prior studies, NVIQ was associated with DLS performance. Regarding contextual factors affecting DLS, lack of opportunity emerged as the most common influence across all DLS subdomains for adults with average or higher NVIQ. Opportunity was also the second most common parent-reported factor affecting DLS performance of adults with NVIQ below 85, following cognitive ability. Better understanding of how these factors intersect will be critical to development of interventions aimed at promoting independence and fostering skill development for autistic adults. Intervention plans should start with providing opportunities to tease apart the reasons for nonperformance of DLS, which could improve the specificity of interventions and treatment targets.