Skip to main content
NCBI home page
International Journal of Developmental Disabilities logoLink to International Journal of Developmental Disabilities
. 2019 Mar 27;66(4):317–326. doi: 10.1080/20473869.2019.1587939

Assessment of adaptive behaviour in children with intellectual disability in Oman: an examination of ABAS-3 factor structure and validation in the Arab context

Mahmoud Mohamed Emam 1,2,✉, Humaira Al-Sulaimani 1, Ehab Omara 1, Ragaa Al-Nabhany 3
PMCID: PMC7942765  PMID: 34141394

Abstract

In Oman and elsewhere in Arab countries school professionals rely on measures that were developed in Western countries. Measures that assess adaptive behaviour (AB) that have been developed in Western cultures are argued to be culturally bound. Nevertheless, these measures are used elsewhere despite a paucity of data that examine their cross-national transportability. The theme of test adaptation and transportability of psychological measures is an important issue for school/educational psychologists, clinicians and educational professionals. The present study describes the adaptation process of the U.S.–developed Adaptive Behaviour Assessment System, Third Edition Teacher Form (ABAS-3-TF) for ages 5 to 14 in the Sultanate of Oman, one of the Gulf Cooperation Council (GCC) countries. Procedures implemented to help ensure a valid translation and cultural adaptation process are described. Data on 410 children with intellectual disability are examined using the Arabic version of ABAS-3-TF, focusing on testing its factor structure and gender invariance. Data from the Arabic version of ABAS-3-TF provide evidence for the adequacy of a one general factor model. Scale reliability is high for the Arabic version. The finding of invariance across males and females indicate that the scale’s factor structure is similar for the two groups. The Arabic version of ABAS-3-TF is adequate for use by clinicians, school psychologists and educational professionals in Oman.

Keywords: intellectual disability, adaptive behavior assessment system, Arab context

Introduction

Adaptive behaviour (AB) refers to an individual’s independent display of behaviours associated with meeting one’s daily personal and social needs, including behaviours expected in domestic and social environments (Oakland and Harrison 2008). The American Association on Intellectual and Development Disabilities (AAIDD) defines AB more formally as “the collection of conceptual, social, and practical skills that have been learned and are performed by people in their everyday lives” (American Association on Intellectual and Developmental Disabilities 2002, p. 43). AAIDD’s further discussion of AB promotes the belief that “adaptive skills should be documented within the context of community and cultural environments typical of a person’s age peers” (Schalock et al. 2010, p. 45). Thus, the assessment of AB should be sensitive to the cultural contexts in which a person lives, including possible sociocultural differences together with opportunities, expectations, and standards for the development of AB. Expected age differences, especially among infants and children, also need to be considered (Schalock et al. 2010). According to the AAIDD 1992 definition, there are 10 adaptive skills including communication, community use, functional academics, leisure, health and safety, school living, self-care, self-direction, social skills, and work skills (Luckasson et al. 1992). In 2002, the AAIDD highlighted three domains within which these skills can be embedded, including conceptual, social, and practical (Luckasson et al. 2002). The conceptual domain skills involve receptive and expressive language, reading and writing skills, and handling money. The social domain skills involve friendships, interactions with others, social participation, social reasoning, comprehension, and reasoning. The practical domain include skills that involve household chores, dressing, bathing, preparing food, and washing dishes (Tassé et al. 2012).

In general, AB measures have been developed and standardised in western countries particularly in the United States (Oakland and Daley 2013, Oakland and Harrison 2008). It has been argued that there is no ideal instrument for assessing AB (The British Psychological Society (BPS) 2001, Ross et al. 2018). In a recent review of AB assessment models, Price et al. (2018) argued that the only common agreement among all AB assessment instruments is that AB is what AB scales measure, not to mention the fact that they are not cross-culturally oriented. Currently, four comprehensive individualized, standardized AB scales are available for use and were normed on a representative U.S. sample of the general population. These scales were developed specifically for ruling in or out a diagnosis of ID. Examples include Scales of Independent Behaviour-Revised (SIB-R; Bruininks et al. 1996), Adaptive Behaviour Scale-School, Second Edition (ABS-S:2) (Lambert et al. 1993) Vineland Adaptive Behaviour Scales-Second Edition (Vineland II) (Sparrow et al. 2005), and the Adaptive Behaviour Assessment System-Second Edition (ABAS-II) (Harrison and Oakland 2003). Recently, the third edition of the ABAS-II and Vineland-II were released by the publisher.

A number of professionals and clinicians in Arab countries and elsewhere may indulge in using these measures without any cautions whereas some others may be wary to use them, as they believe that AB measures are relevant only within the culture in which the test was normed and standardised (Oakland et al. 2013). The view held by the opponents of using adaptive scales in other cultures could be true given the difference in community and broader cultural environments between countries, which may result in differences in characteristics that affect the development and display of essential adaptive skills. Alternatively, there is empirical evidence from cross-national assessment of AB that supports the similarity between the original form and the validated forms (Oakland et al. 2013).

Nevertheless, professionals' caution is expected when there is lack of information on the cross-national validity of AB measures. Furthermore, the transportability of AB measures in similar cultures such as those of the Western world could be argued to be easier than in largely different cultures, as in the case of African or Arab countries. Measures of AB developed in the United States are thought to not reflect opportunities, expectations, and standards influencing AB in the Arab region as well as in other regions in which large cultural differences exist. In this regard, a number of scholars argue that the decision of adapting an existing measure developed in another culture rather than developing a locally new measure is driven by a sense of security and marketing issues. An evidence based and empirically supported AB measure may be seen more favourably than developing a new one even though it may be culturally bound (Tassé and Craig 1999, Hambleton 2005). This belief has lead to the adaptation of the ABAS-II in different cultures (Oakland et al. 2013) and it is this belief that has guided the current study.

The Sultanate of Oman is a country in Southwest Asia, on the southeast coast of the Arabian Peninsula. It borders the United Arab Emirates in the northwest, Saudi Arabia in the west and Yemen in the southwest. It is a member of the United Nations, the Arab League and the Gulf Cooperation Council (GCC). Oman, whose culture is collective as opposed to individualistic cultures of Western countries, has made a stride in expanding access to education, care, and rehabilitation services to individuals with disability by issuing The Omani Children with Disabilities Care and Rehabilitation Act in 1996 which was reauthorized in 2008. Children with ID, in particular, have been the target of a number of policies and initiatives that supported the provision of psychoeducational assessments, appropriate educational provision and early intervention services as part of Oman’s national scheme to work towards achieving the goals outlined by the UN convention on the Rights of Persons with Disabilities (Al-Balushi et al. 2011, Mohamed Emam 2016).

Research and practice on ID in Oman has been informed by scholarship from Western countries. For example, the definition and procedures followed by clinicians and practitioners and approved by the government are similar to those applied in the United States. Currently, there is no AB measure available for use in the Omani context by. Practitioners and clinicians either conduct an interview with parents or use their knowledge to observe children and determine if they have AB deficits. Arab countries in general and GCC countries in particular share several historical, cultural, linguistic, geographic, and social characteristics. Therefore, data collected on the ABAS-3 in the Sultanate of Oman could reflect a different structure from the one obtained for the US-based sample.

Measurement of AB

AB is defined as “the level of everyday performance of tasks that is required for a person to fulfil typical roles in society, including maintaining independence and meeting cultural expectations regarding personal and social responsibility” (VandenBos 2006, p. 18). Educational professionals engaged in assessment, diagnosis, special education eligibility determination, and treatment of children and adolescents with a variety of educational and behavioural conditions have often given priority to the measurement and evaluation of AB (Floyd et al. 2015, Oakland and Daley 2013). For example, according to the American Association on Intellectual and Developmental Disabilities (Schalock et al. 2010), diagnosis of intellectual disability (ID) requires consideration of AB in the following domains: conceptual, social, and practical. Similarly, the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (American Psychiatric Association 2000) recommended consideration of AB skill areas, and the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5; American Psychiatric Association 2013) requires consideration of conceptual, social, and practical skills as in the AAIDD guidelines.

The research field in the area of AB abounds with a large number of AB scales (Arias et al. 2013, Schalock 1999). According to research reviews (McNicholas et al. 2018, Floyd et al. 2015), four AB scales, out of several others, are based on a 3-factor structure of AB construct (i.e. conceptual, social, and practical skills) with adequate evidences of reliability and validity, and have been standardised on individuals with and without ID. These include: (1) the Vineland Adaptive Behaviour Scales-2nd edition (Sparrow et al. 2005); (2) the Adaptive Behaviour Assessment System-II (Harrison and Oakland 2003); (3) the Scales of Independent Behaviour-Revised (Bruininks et al. 1996) and (4) the Adaptive Behaviour Scale-School Version (Lambert et al. 1993).

The Adaptive Behaviour Assessment System-Third Edition

The Adaptive Behaviour Assessment System–3 (ABAS-3; Harrison and Oakland 2015), an updated version of the BAS-II (Harrison and Oakland 2003), is a comprehensive, norm-referenced assessment of adaptive skills needed to effectively and independently care for oneself, respond to others, and meet environmental demands at school. ABAS-3 is consistent with AAIDD’s 1992 and 2002 definitions. It includes an assessment of the aforementioned 10 skills that are understood further in light of the three adaptive domains together with a general adaptive composite (GAC) score. The ABAS-3 scales produce social, practical, and conceptual skill domains that are aligned with AAIDD (Schalock et al. 2010) guidelines and DSM-5 criteria for the assessment of ID (American Psychiatric Association 2013). There are teacher and parent forms for ABAS-3. It has been argued that there are differences between teachers and parents when rating children with ID. Teachers may have better opportunities to observe skills related to conceptual and social domains whereas parents have better opportunities to observe skills related to the practical domain. However, parents tend to be biased for their children when they rate their behaviours and abilities (Noland and McCallum 2000). The reason we chose to collect data using the TF was twofold: first the collective culture which characterized Oman stigmatizes parents when they disclose information about their children, second, teachers spend more time with children with ID which would affect the accuracy of the collected data.

No prior study was conducted on the robustness of the Arabic version of ABAS-3-TF. Information on the structural equivalence of the ABAS-3 could help to answer questions to which no clear conclusions were reached by previous studies. For example, previous research on the factor structure of AB measures including the ABAS-II structure provided inconclusive findings on whether they are uni- or multi-dimensional structure (Oakland et al. 2013, Aricak and Oakland 2010). A number of studies on ABAS-II provided support for an underlying single factor of AB (Oakland et al. 2013, Aricak and Oakland 2010, Oakland and Algina 2011, Aricak and Oakland 2009, Wei et al. 2008). These findings are consistent with research by McGrew and Bruininks (1989), who found that most AB instruments measure a general factor. This, however, is in contrast with the general position of the AAIDD advocates of a three-factor model, namely: Conceptual, Social, and Practical domains as separate factors.

The main purpose of the current study was to describe the adaptation process of the Arabic Omani version of the ABAS-3 and to answer the following two questions:

  1. Which of the two factor structures (one factor model vs. three-factor model) is supported by the validation data in the Arab context?

  2. Is there a gender invariance for the Arabic version of ABAS-3-TF?

Method

Participants

The study sample included 410 children with intellectual disability from both genders (54.1% boys, 45.9% girls). Participants, aged between 5 and 14 years old (M = 10.32; SD = 2.46) for boys and (M = 10.06; SD = 2.44) for girls, were selected from three main geographical areas in the Sultanate of Oman, and were enrolled in elementary and intermediate inclusive public schools in addition to Al-Tarbia Al-Fekriya school (special education school for children with intellectual disability). Educational placements of children with ID depend parents’ preference and fulfilment of the formal requirements that include: (1) acceptable verbal communication, (2) no history of emotional and behavioural difficulties, (3) receiving a formal diagnosis of ID from a clinician at a public hospital. All participants received a formal diagnosis of intellectual disability by a developmental paediatrician and/or a behavioural medicine professional at public hospitals as part of the formal procedures by the Ministries of Social Development and Education to place those children in educational settings or care and rehabilitation centres. In Oman, children receive a formal diagnosis of ID if they met the following criteria: (a) an IQ score on the Arabic Wechsler Intelligence Scale for Children (WISC) or any other intelligence test of approximately two standard deviations below average, thereby denoting a significant cognitive deficit, (b) a record of continued poor daily living skills as rated by a formal interview with the parents, and (c) absence of any comorbid condition such as attention deficit hyperactivity disorder, autism spectrum disorder or any other condition.

Instrument

The ABAS-3 consists of five forms: parent and caregiver forms for ages 0 to 5, parent and teacher forms for ages 5 to 21, and an adult form for ages 16 to 89. This study utilizes only the teacher form and does not cover the whole age range. We collected data on children aged between 5 and 14. The reason for confining the data collection to this age range is twofold: first Oman does not have a formal preschool education since the formal education system begins from grade 1 at the age of 5. Second, we could not collect data on children with ID above the age of 14 because most of them leave school at that age to join vocational training centres. ABAS-3 Teacher Form (TF) assesses ten adaptive skills: communication (C), community use (CU), functional academics (FA), school living (SL), health and safety (H&S), leisure (L), self-care (SC), self-direction (SD), social and work skills (SW). Data from these 10 skill areas contribute to one of the three domains: conceptual, social, or practical. The domain scores are used to obtain the general adaptive composite (GAC).

For the purpose of this study, work skills were not included because the study sample targeted younger population. Validation data on the Teacher’s Form were acquired from classroom teachers by indicating for each item if the student was able to independently perform an activity (when needed) and, if so, how often (never, sometimes, or always).

Translation and adaptation of the Arabic version of ABAS-3-TF

ABAS-3 was first purchased from the publisher. The test adaptation process in Oman was guided by the following principles: (1) Using terminology commonly understood in the country, (2) examining the cultural appropriateness of each item carefully, (3) attending to the item’s implied meaning, not a literal translation, in order to retain the item’s psychological meaning, not merely its semantic meaning, (4) maintain the equivalent meaning of the original item, (5) strivin to maintain the difficulty of the original item when it was substituted, (6) maintaining the test administration, format, and scoring designs, (7) Attempting to maintain the factor structure of the original test, (8) promoting item comprehension, and (9) ensuring persons understand the methods used to complete the scale. Applying these principles results in the Arabic version of ABAS-3-TF that included the following number of items: communication (22), community use (15), functional academics (22), school living (22), health and safety (15), leisure (16), self-care (19), self-direction (21), social (22). The items are distributed across three areas: conceptual (C, FA, SD, 65), social (L, SW, 38), practical (CU, SL, H&S, SC, 71), and General Adaptive Composite (GAC, 174).

A committee of five professionals (i.e., the authors: one faculty member with a background in special education and training in a western country, two faculty members with a background in measurement and evaluation, a senior special educator with a 10 year experience in working with children with ID, and an applied linguist who has good command of both Arabic and English), translated the ABAS-3 items from English to Arabic. Each translator worked independently on the Arabic version of the ABAS-3. This method, called a parallel approach (Schoua-Glusberg 1992) or team translation (Acquadro et al. 1996), involves several translators working independently. The prior professional experiences of this team are consistent with Van de Vijver and Hambleton’s (1996) recommendation for committee diversity. The forward translation method is considered superior to the popular back translation method, which “puts a premium on literal translation” (van de Vijver and Leung 1997, p. 39). In the forward translation, delivery of the meaning in the target language is prioritized, whereas the back-translation method tends to overlook serious problems in the translated version because the back-translator compensates for any errors by “recovering” the original phrase into the source language (Hambleton 2005, van de Vijver and Leung 1997).

Following the translation process, the translators found that the items designate widely recognised daily behaviours expressed in simple and understandable words. The work done by the applied linguist and the translation team was guided by the explicit requirement to adhere to the best linguistic equivalence for the items. Every item was examined to ensure it measured a behaviour typical of Omani culture for children and youth and whether the behaviour described by each item was likely to be exhibited by Omani children and youth. The five resulting translations were compared and the final version of every item was decided on by consensus when possible. When consensus was not possible, decisions were made as to the best way to reconcile discrepancies. This draft translation was back-translated from Arabic to English. The back-translation was analysed during two review meetings between the authors and the applied linguist. This process lead to the final translation of the Arabic version of ABAS-3-TF that then was used to collect the data for research purposes. No item was deleted or added as a result of the translation procedure.

Data collection

Prior to data collection ethical approval was obtained from the University Research Ethical Board (UREB). Additionally, approval was obtained from the Ministry of Education to conduct the study at the selected schools. Consent was obtained from the schools as well as from the families of students who would be rated by their teachers on the Arabic version of ABAS-3-TF. Families gave assent by signing a consent form prior to their children's participation in the study, indicating that they were willing to have their children included in the study in order to be assessed with regard to the aspects relating to their adaptive skills. Families and teachers were informed that participation in the study was voluntary and that confidentiality of the data obtained would continue even after the study results are disseminated. The authors ensured that the teachers who rated the children were familiar with them and worked with them at least for one year. Teachers were given three months to fill out the Arabic version of ABAS-3-TF, thereby not intervening with their daily schedule. The ABAS-3-TF administration time ranges from 15-20 minutes per rating form. The fourth author moderated the data collection phase and provided assistance with regard to administration instruction and other queries that were raised by the teachers.

Data analysis

Confirmatory factor analysis was conducted on the whole sample to test the two competing factor models (the one-factor vs. the three-factors) of the Arabic version of ABAS-3-TF. The results of the CFAs were compared to determine which factorial model strongly fits with the collected data. Model fit was evaluated using a combination of absolute and incremental goodness-of-fit indices, including Comparative fit index (CFI), Tucker-Lewis index (TLI), root mean square error of approximation (RMSEA), standardized Root Mean Square Residual (SRMR), Akaike Information Criterion (AIC), chi-square statistics (χ2). In addition, chi-square/df index (χ2/df) was used given that the magnitude of χ2 is dependent on sample size, which may cause plausible models to be rejected (Schermelleh-Engel et al. 2003). CFI and TLI values in the range of 0.90 to 0.95 may reflect an acceptable model fit, whereas values greater than greater than 0.95 indicate an excellent fit (Hu and Bentler 1999). RMSEA values of 0.05 or below with 90% confidence intervals whose lower bound contains, or very close to, zero and whose upper bound is not more than 0.08, are thought to indicate a close fit, 0.05–0.08 a fair fit, and 0.08–0.10 a marginal fit by one standard (Cheung and Rensvold 2002). SRMR values as high as 0.08 are acceptable (Hu and Bentler 1999). AIC value is used to determine the most parsimonious model that has the smallest value, and because AIC is not normed to 0–1 scale, it is difficult to specify cut-off criteria for the most superior model (Kline 2005). Finally, multi-group CFA for Arabic version of ABAS-3-TF was conducted to examine different degrees of factorial invariance across gender. This includes configural invariance, weak invariance, and strong invariance. Non-significant chi-squared difference and values of ΔTLI and ΔCFI smaller than or equal to 0.01 are used as criteria for model comparisons (Cheung and Rensvold 2002).

Results

Descriptive statistics

Table 1 shows the means, standard deviation, standard error, skewness and kurtosis of each of the nine subscales. It can be noted that none of the subscales has a severe skewness or kurtosis as all values are between −2 and +2, which indicates the normality of all subscales scores (George and Mallery 2010). The correlations among the nine subscales of Arabic version of ABAS-3-TF (Table 1) were significant (p < .001) and ranged from 0.43 (Functional Academics and Self-Care) to 0.83 (School Living & Self-Direction, Health and Safety & Social). These results indicate that none of the nine subscales were overlapped (all correlations were less than one). Table 1 also reports the Cronbach's alpha Coefficients that are derived from raw scores. The coefficients for the nine subscales were greater than 0.90, indicating good reliability.

Table 1.

Means, standard deviations, standard errors, skewness, Kurtosis, and the inter-correlations among the nine components of ABAS-3

  M SD SE Sk Kur 1 2 3 4 5 6 7 8 9
Communication 37.34 13.79 0.68 −0.37 −0.29 (0.94)                
Community use 23.16 9.53 0.47 −0.21 −0.59 .76 (0.91)              
Functional academics 20.17 13.46 0.66 0.56 −0.27 .69 .72 (0.95)            
School living 44.33 14.05 0.69 −0.56 0.00 .65 .71 .60 (0.93)          
Health and safety 29.92 10.51 0.52 −0.79 0.11 .71 .76 .60 .81 (0.93)        
Leisure 29.46 10.54 0.52 −0.52 0.00 .71 .67 .59 .77 .76 (0.92)      
Self-care 44.19 10.52 0.52 −1.28 1.59 .51 .55 .43 .67 .63 .58 (0.94)    
Self-direction 36.61 13.82 0.68 −0.39 −0.18 .70 .74 .67 .83 .82 .80 .67 (0.94)  
Social 42.81 14.40 0.71 −0.67 −0.01 .76 .72 .63 .76 .83 .81 .65 .82 (0.95
Open in a new tab

Note: M = mean, SD = standard deviation, SE = standard error, Sk = skewness, and Kur = Kurtosis.

All correlations are based on row scores and were statistically significant (p < 0.001). Also, Cronbach's alpha Coefficients were included between brackets in the diagonal (Alpha).

One-factor model vs. three-factor model

In order to determine which of the two models fit with the validation data for the Arabic version of ABAS-3-TF, each model was tested separately. Table 2 shows the goodness of fit indices for both models. Overall, all standardised factor loadings had values greater than 0.7 and were significant (p < 0.01) for both models. To test the one-factor structure of the Arabic version of ABAS-3-TF, a confirmatory factor analysis was conducted to evaluate the structure. It was hypothesised that the nine subscales would load on a general factor. The goodness-of-fit indices for the model were: χ2(21) = 64.54, χ2/df = 3.07, CFI = 0.99, TLI = 0.98; AIC = 112.54, SRMR = 0.02 and RMSEA = 0.07 with 90% CI (0.05–0.09). The CFI and TL, fulfilled the criteria as recommended by Hu and Bentler (1999), indicating a reasonable fit to the data. Whereas χ2 value was significant (p < 0.01), χ2/df was slightly more than 3, which is considered acceptable (Schreiber et al. 2006). SRMR value indicates a good fit and RMSEA value represents a fair fit. Overall, the fit indices indicated acceptable fit of the model to the data.

Table 2.

Goodness of fit statistics and standardized factor loading of the one-Factor vs. three-Factor Confirmatory Factor Analysis of ABAS-3

Subscales One-factor model Three-factors model
Conceptual Practical Social
Communication 0.80 0.81    
Functional academics 0.68 0.71    
Self-direction 0.91 0.91    
Self-care 0.71   0.70  
Community use 0.79   0.83  
School living 0.90   0.88  
Health and safety 0.90   0.90  
Leisure 0.86     0.88
Social 0.92     0.92
Goodness-of-fit indices        
Chi-square 64.54* 219.48*
df 21 24
Chi-square/df 3.07 9.15
CFI 0.99 0.95
TLI 0.98 0.92
RMSEA (90% CI) 0.07 (0.05–0.09) 0.14 (0.12–0.16)
SRMR 0.02 0.04
AIC 112.54 261.48
Open in a new tab

Note: All factor loadings are significant (p < 0.01). Chi-Square value is significant (p < 0.01)

To test the three-factor model which include three correlated factors: conceptual, practical, and social, a CFA was performed on the whole sample. The goodness-of-fit indices for this model were: χ2(24) = 219.48, χ2/df = 9.15, CFI = 0.95, TLI = 0.92, AIC = 261.48, SRMR = 0.04 and RMSEA = 0.14 with 90% CI (0.12–0.16). The magnitude of χ2 was high and χ2/df was not at an acceptable level (> 3). In addition, the lower value of the 90% confidence interval of RMSEA was greater than 0.1 and TLI was slightly smaller than 0.95. The comparative measure of fit AIC for the three-factor model was higher than double its figure for the one-factor model. Overall, fit indices of the three-factor model indicated poor fit to the data compared to the one-factor model.

We also used exploratory factor analysis (EFA) to examine if there is more appropriate structure than the two models that were examined using CFA (one factor vs. three factor model). The EFA with principal component extraction method was conducted for the subscales to explore if there are other factor structures describing participant’s responses. The results indicated that only one component with eigenvalue greater than one was extracted. The percentage of explained variance of the extracted component was 73.23%. The factor loadings of subscales ranged between (0.73 and 0.92). This result also supports the one factor model for ABAS-3-TF. Based on the results of testing the two models the well-fit and simpler model is preferable, which applies to the one factor model in our case. Thus, subsequent results will be reported for the one-factor model only.

Single-group confirmatory factor analysis and configural invariance

To test gender invariance we conducted CFA separately for boys and girls. As shown in Table 3, standardised factor loadings were greater than 0.7 for both groups. Although χ2values were significant due to their sensitivity to sample size (Chen 2007), the overall goodness-of-fit indices indicated that the model fit was adequate. Therefore, the one factor model was accepted as the baseline model in subsequent multi-group CFA. Then, we examined the configural invariance using multi-group CFA. As shown in Table 4, the goodness-of-fit indices of unrestricted model for the data support the consistency of the hypothesized construct of Arabic version of ABAS-3-TF as measuring one general factor. Thus, the construct of Arabic version of ABAS-3-TF was configurally invariant across gender.

Table 3.

One-factor confirmatory factor analysis of ABAS-3-AR for boys and girls

Subscales Standardized factor loadings
Boys Girls
Communication 0.77 0.83
Community use 0.80 0.78
Functional academics 0.65 0.72
School living 0.90 0.90
Health and safety 0.92 0.87
Leisure 0.86 0.87
Self-care 0.73 0.68
Self-direction 0.90 0.91
Social 0.94 0.90
Goodness-of-fitting indices    
Chi-square 57.11* 38.79*
df 21 21
Chi-square/df 2.72 1.85
CFI 0.98 0.99
TLI 0.97 0.98
RMSEA (90% CI) 0.09 (0.06–0.12) 0.07 (0.03–0.10)
SRMR 0.03 0.02
AIC 105.11 86.79
Open in a new tab

Note: All factor loadings are significant (p < 0.01). (*) p < 0.01

Table 4.

One-factor multi-group confirmatory factor analysis results of ABAS-3 for boys and girls

  Unrestricted factor loadings (M1)
 Equal factor loadings (M2)
 Equal intercepts (M3)
  Boys Girls Boys Girls Boys Girls
Communication 0.77 0.83 0.79 0.81 0.79 0.81
Community use 0.81 0.79 0.80 0.81 0.80 0.80
Functional academics 0.65 0.72 0.67 0.70 0.67 0.70
School living 0.90 0.90 0.90 0.90 0.90 0.90
Health and safety 0.92 0.88 0.92 0.88 0.92 0.88
Leisure 0.85 0.87 0.86 0.87 0.86 0.87
Self-care 0.73 0.68 0.73 0.69 0.73 0.69
Self-direction 0.90 0.91 0.91 0.91 0.91 0.91
Social 0.94 0.90 0.94 0.90 0.94 0.90
Goodness-of-fitting indices            
Chi-square 109.72* 121.61* 163.16*
df 44 52 61
Chi-square/df 2.49 2.34 2.68
CFI 0.98 0.98 0.97
TLI 0.97 0.97 0.97
RMSEA (90% CI) 0.06 (0.05–0.08) 0.06 (0.04–0.07) 0.06 (0.05–0.08)
SRMR 0.027 0.030 0.030
AIC 237.72 233.61 257.16
Model comparisons indices      
ΔChi-square – 11.89 53.44*
Δdf – 8 17
ΔTLI – <0.01 <0.01
ΔCFI – <0.01 <0.01
Open in a new tab

Note: Only standardized factor loadings are reported. (*) p < 0.01.

Weak factorial invariance

Following testing the configural invariance, we tested the weak invariance by conducting multi-group CFA with factor loadings constrained to be equal for boys and girls. As shown in Table 4, the results show that χ2(52) = 121.61, p < 0.01, and the other goodness-of-fit indices were: χ2/df, CFI, TLI, and RMSEA, indicating a possible fit. Additionally, the model comparison indices for comparing M1 and M2 were: χ2(8) = 11.89, p = 0.16, indicating insignificant differences in the factor loadings for both groups. These results support the weak factorial invariance for Arabic version of ABAS-3-TF across gender. Obtaining the weak invariance, however, is not enough to make comparisons of means across gender.

Strong factorial invariance

We tested the strong factorial invariance. As shown in Table 4, a multi-group CFA was conducted with factor loadings and factor intercepts constrained to be equal for boys and girls (M3). The values of ΔTLI and ΔCFI were less than 0.01, which still indicates insignificant differences in the factor intercepts for boys and girls. This supports the strong factorial invariance for Arabic version of ABAS-3-TF across gender. Thus, our results show that teachers view boys and girls, who are assessed on the general latent construct, equally on the subscales of the ABAS-3-TF. This means that there is not potential measurement bias in the Arabic version of the ABAS-3-TF subscale scores. Given that that this level of invariance is established in our study, practitioners who will use the Arabic version can make valid inferences about the differences in the ABAS-3-TF scores across gender

Discussion

The current study aimed to examine the factor structure of the Arabic version of ABAS-3-TF. Two competing models were tested (one factor model vs. three-factor model). In addition, we aimed to examine the measurement invariance of Arabic version of ABAS-3-TF across gender. A multi-group confirmatory factor analysis was used to test for configural, weak, and strong invariances across both genders. The baseline one-factor model was identified and was then simultaneously tested in both genders by means of four progressively restrictive models. The difference in the χ2 statistic, as well as differences in fit indices between the restricted and full models was used as a criterion to determine the measurement invariance. The results revealed that the data has an acceptable fit with the one-factor model in the whole sample. The χ2 values and fit indices ranged between acceptable to fair model fit for the observed data on the Omani sample of children with ID aged 5–14. When testing for measurement invariance across both genders, chi-squared difference and values of ΔTLI and ΔCFI showed configural, metric, and scalar, invariances. This indicates that AB was being measured in the same way in both males and females.

The results are consistent with previous studies. Several studies suggested that a child’s individual’s independent display of behaviours associated with meeting his or her daily personal and social needs, including behaviours expected in domestic and social environments can be interpreted by one single underlying general factor of AB (Oakland et al. 2013, Aricak and Oakland 2010, Oakland and Algina 2011, Aricak and Oakland 2009, Wei et al. 2008). Aricak and Oakland (2010) performed a multi-group confirmatory factor analysis of ABAS-II on 1,690 children aged 5–21, resulting in one factor model for the ABAS-II Teacher Form for the sample which was divided in two age groups (5–12 and 13–21). Oakland et al. (2013) tested the one and three factor models of ABAS-II Parent Form, in three different countries including U.S. The authors reviewed the data using the adjusted goodness-of-fit index (AGFI) and the root mean square error of approximations (RMSEA) and found a slightly better fit for the three-factor model for the U.S. version and a slightly better fit for the one-factor model for the Romanian and Taiwanese versions. The fit indices for all three versions were below the conservative threshold (RMSEA < .05 or RMSEA < .08) typically used in rigorous fitting standards. The study by Oakland and Algina (2011) on the ABAS-II parent form for children aged 0–5 similarly supported the one factor model.

Most importantly, AB was conceptualised in children with ID by Omani teachers in concurrence with previous findings by Aricak and Oakland (2009). The difference, however, lies in the sample which was selected in both studies. In our study, the data came from sample of children with ID whereas the data in Aricak and Oakland's study came from a sample of typically developing children. Harrison and Oakland (2015), however, argued that samples from clinical groups such as those with ID displayed distinct profiles on the nine adaptive skills. The authors stated that the dimensionality of a construct of AB depends on whether the data come from a normal population or from clinical samples. The number of dimensions the test intends to measure as well as the number of skill dimensions and levels on those dimensions exhibited by the examinees also may affect the tests factor structure (Reckase 1997).

Our findings, however, raises questions about the sample of our study. The participants included children who received their education in inclusive settings with typically developing children and others from special education school for children with ID. It could be argued that teachers' ratings of children in those two different contexts, particularly those in inclusive settings, could have affected the resulting dimensionality of the Arabic version of ABAS-3-TF. We did not include children with ID who receive their care and education in care and rehabilitation centres that are run by the Ministry of Social Development. These children, if included, could have affected the resulting factor structure of the ABAS-3. Children who participated in our study were in relatively similar contexts given that children in mainstream schools were in self-contained class and were given opportunities to interact with their typically developing peers in school recess and activity times.

Our results supporting the unidimensionality of the ABAS-3 have some connotations for professionals working with children with ID in Oman. First, the selection and clinical use of the ABAS-3 should consider the match between a test’s content and a child’s opportunities, expectations, and standards for AB development. Second, when using the ABAS-3, the Omani professionals are advised to consider a child’s opportunities and expectations governing the acquisition of AB. This advice is particularly warranted when assessing the AB of children with ID in school vs. care settings. For example, in school settings as compared to care settings, children with ID are exposed to more opportunities for interaction with typically developing peers who can act as role models for several communication and social skills. Third, Omani professionals in Oman may find using the adapted Arabic version of the ABAS-3 easier than developing a new scale particularly as our results showed that there are more similarities than differences in the children’s foundation daily behaviours that may characterize children’s growth and development in different cultures.

Limitations

The purpose of this article was to provide evidence on the construct validity of the Arabic version of ABAS-3-TF for use with young children with ID. Our findings supported the one factor solution over the three-factor structure of the ABAS-3. Given that we performed our analyses at the component levels, future studies may consider running the CFA at the item level with larger samples; and use item response theory to examine the item bias and differential item functioning. We did not examine other forms of validity (e.g., content, concurrent, predictive). Furthermore, knowledge of a measure’s construct validity does not necessarily inform test uses of a test’s validity at the subtest and item levels. The ABAS-3 Manual (Harrison and Oakland 2015) addresses these issues. We also examined gender invariance in our study but a future study that examines measurement invariance across different countries in the Arab region is needed. Although this study did not examine the development of AB longitudinally, the study’s cross-sectional design provides some information as to the stability of the general construct and its underlying structure for children with ID from both genders. Longitudinal studies that explore the structure of the AB across different ages are needed in the Arab context. Similar studies that examine metric invariance of data from parents of children and adults are also needed. Additionally, the results from the present study should be considered cautiously when generalizing to other samples as the data collection was confined only to children with ID. Future studies may include typically developing children and compare between both populations. Future research should also attempt to validate this widely used measure of AB across all of the Arab countries in larger samples across different age cohorts. Regardless of these limitations, the current study is the first one in the region to validate a widely used AB measure.

Implications for practice

The Arabic version of ABAS-3-TF can be a promising measure for assessing adaptive skills in children with ID. Our findings provided evidence supporting the one factor structure of the scale and the measurement equivalence in both males and females. Because the ABAS-3 is a widely used instrument and reflects the recent conceptualization of AB by professional organizations, it allows teachers to contribute to the diagnosis of children with ID in the Oman and other Arabic contexts. Given the fact that Oman shares the same culture with other Arabic speaking countries, the findings of our study can be argued to generalize to other Arab countries. This could help professionals and practitioners use the ABAS-3 Arabic version to obtain information that can benefit in planning appropriate intervention. Additionally, professionals can use the information to help parent understand the strengths and difficulties of their children. This is the first study to provide empirical data on the validation of the ABAS-3 in the Arabic region. Several clinicians, school professionals complain of the lack of adequate and evidence based assessment instrument of AB; they can now use the Arabic version of the ABAS-3-TF, for early identification and monitoring of the of AB in students with ID in school settings.

Disclosure statement

No potential conflict of interest was reported by the authors.

References

  1. Acquadro, C., Jambon, B., Ellis, D. and Marquis, P. 1996. Language and translation issues. In: Spilker B., ed. Quality of life and pharmacoeconomics in clinical trials. 2nd ed. Philadelphia, PA: Lippincott-Raven, pp. 575–586. [Google Scholar]
  2. Al-Balushi, T., Al-Badi, A. H. and Ali, S. 2011. Prevalence of disability in Oman: statistics and challenges. Canadian Journal of Applied Sciences, 1(3), 81–96. [Google Scholar]
  3. American Association on Intellectual and Developmental Disabilities . 2002. Mental retardation: Definition, classification, and systems of support. Washington, DC: American Association on Intellectual and Developmental Disabilities. [Google Scholar]
  4. American Psychiatric Association . 2000. Diagnostic and statistical manual of mental disorders. 4th ed., text rev. Washington, DC: American Psychiatric Association. [Google Scholar]
  5. American Psychiatric Association . 2013. Diagnostic and statistical manual of mental disorders. 5th ed. Washington, DC: American Psychiatric Association. [Google Scholar]
  6. Arias, B., Verdugo, M. Á., Navas, P. and Gómez, L. E. 2013. Factor structure of the construct of adaptive behaviour in children with and without intellectual disability. International Journal of Clinical and Health Psychology, 13(2), 155–166. [Google Scholar]
  7. Aricak, O. T. and Oakland, T. 2009. Multigroup confirmatory factor analysis for the teacher form, ages 5–21, of the Adaptive Behaviour Assessment System-II. Journal of Psychoeducational Assessment, 28, 1–7. [Google Scholar]
  8. Aricak, O. T. and Oakland, T. 2010. Multigroup confirmatory factor analysis for the teacher form, ages 5 to 21, of the Adaptive Behaviour Assessment System-II. Journal of Psychoeducational Assessment, 28(6), 578–584. [Google Scholar]
  9. Bruininks, R., Woodcock, R., Weatherman, R. and Hill, B. 1996. Scales of independent behaviour-revised. Rolling Meadows, IL: Riverside. [Google Scholar]
  10. Chen, F. F. 2007. Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling, 14(3), 464–504. [Google Scholar]
  11. Cheung, G. W. and Rensvold, R. B. 2002. Evaluating goodness-of-fit indexes for testing measurement invariance. structural equation modeling. A Multidisciplinary Journal, 9(2), 233–255. [Google Scholar]
  12. Floyd, R. G., Shands, E. I., Alfonso, V. C., Phillips, J. F., Autry, B. K., Mosteller, J. A., … and Irby, S. 2015. A systematic review and psychometric evaluation of adaptive behaviour scales and recommendations for practice. Journal of Applied School Psychology, 31(1), 83–113. [Google Scholar]
  13. George, D. and Mallery, M. 2010. SPSS for windows step by step: a simple guide and reference, 17.0 update. 10th. ed. Boston, MA: Pearson. [Google Scholar]
  14. Hambleton, R. K. 2005. Issues, designs and technical guidelines for adapting tests into multiple languages and cultures. In: Hambleton R. K., Merenda P. F. and Spielberger C. D., eds., Adapting educational and psychological tests for cross-cultural assessment. Mahwah, NJ: Lawrence Erlbaum, pp. 3–38. [Google Scholar]
  15. Harrison, P. and Oakland, T. 2003. Adaptive behaviour assessment system-(ABAS-II). San Antonio, TX: The Psychological Corporation. [Google Scholar]
  16. Harrison, P. and Oakland, T. 2015. Adaptive behaviour assessment system. 3rd ed. Torrance, CA: Western Psychological Services. [Google Scholar]
  17. Hu, L. T. and Bentler, P. M. 1999. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. [Google Scholar]
  18. Kline, R.B. 2005, Principles and practice of structural equation modelling. 2nd ed. New York: The Guilford Press. [Google Scholar]
  19. Lambert, N., Nihira, K. and Leland, H. 1993. Adaptive Behaviour Scale—School, 2nd ed. Austin, TX: PRO-ED. [Google Scholar]
  20. Luckasson, R., Borthwick-Duffy, S., Buntinx, W. H., Coulter, D. L., Craig, E. M. P., Reeve, A. , Schalock, R. L., Snell, M. E., Spitalnik, D. M., Spreat, S. and Tasse, M. J. 2002. Mental retardation: Definition, classification, and systems of supports. Washington, DC: American Association on Mental Retardation. [Google Scholar]
  21. Luckasson, R., Borthwick-Duffy, S., Buntinx, W. H., Coulter, D. L., Craig, E. M. P., Reeve, A., Schalock, R. L., Snell, M. E., Spitalnik, D. M., Spreat, S. and Tasse, M. J. 1992. Mental retardation: Definition, classification, and systems of supports. Washington, DC: American Association on Mental Retardation. [Google Scholar]
  22. McGrew, K. S. and Bruininks, R. H. 1989. The factor structure of adaptive behaviour. School Psychology Review, 18, 64–81. [Google Scholar]
  23. McNicholas, P. J., Floyd, R. G., Woods, I. L., Jr., Singh, L. J., Manguno, M. S. and Maki, K. E. 2018. State special education criteria for identifying intellectual disability: A review following revised diagnostic criteria and Rosa's Law. School Psychology Quarterly, 33(1), 75–82. [DOI] [PubMed] [Google Scholar]
  24. Mohamed Emam, M. 2016. Management of inclusive education in Oman: A framework for Action. Support for Learning, 31(4), 296–312.
  25. Noland, R. M. and McCallum, R. S. 2000. A comparison of parent and teacher ratings of adaptive behaviour using a universal measure. Journal of Psychoeducational Assessment, 18(1), 39–48. [Google Scholar]
  26. Oakland, T. and Algina, J. 2011. Adaptive Behaviour Assessment System-II parent/primary caregiver form: Ages 0–5: Its factor structure and other implications for practice. Journal of Applied School Psychology, 27(2), 103–117. [Google Scholar]
  27. Oakland, T. and Daley, M. 2013. Adaptive behaviour: Its history, concepts, assessment, and applications. In: Geisinger K. F., ed., APA handbook of testing and assessment in psychology, Vol. 3: Testing and assessment in school psychology and education. Washington, DC: American Psychological Association, pp. 183–212. [Google Scholar]
  28. Oakland, T. and Harrison, P. 2008. Some historical background for adaptive behaviour and its use. In: Oakland T. and Harrison P., eds., Adaptive Behaviour Assessment System–II: Clinical uses and interpretations. New York: Elsevier, pp. 4–20. [Google Scholar]
  29. Oakland, T., Iliescu, D., Chen, H. Y. and Chen, J. H. 2013. Cross-national assessment of adaptive behaviour in three countries. Journal of Psychoeducational Assessment, 31(5), 435–447. [Google Scholar]
  30. Price, J., Morris, Z. and Costello, S. 2018. The application of adaptive behaviour models: a systematic review. Behavioral Sciences, 8(11), 1–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Reckase, M. D. 1997. A linear logistic multidimensional model for dichotomous item response data. In: van der Linden W. J. and Hambleton R. K., eds., Handbook of modern item response theory.New York: Springer-Verlag, pp. 271–286. [Google Scholar]
  32. Ross, G. E., Hocken, K. and Auty, J. M. 2018. The reliability and validity of the Adaptive Functioning Assessment Tool in UK custodial settings. Journal of Intellectual Disabilities, [DOI] [PubMed] [Google Scholar]
  33. Schalock, R. L. 1999. Adaptive behaviour and its measurement: Implications for the field of mental retardation. Washington, DC: American Association on Mental Retardation. [Google Scholar]
  34. Schalock, R. L., Buntinx, W. H. E., Borthwick-Duffy, S., Bradley, V., Craig, E. M., Coulter, D. L., … Yeager, M. H. 2010. Intellectual disability: Definition, classification, and system of supports, 11th ed. Washington, DC: American Association on Intellectual and Developmental Disabilities. [Google Scholar]
  35. Schermelleh-Engel, K., Moosbrugger, H. and Müller, H. 2003. Evaluating the fit of structural equation models: Tests of significance and descriptive goodness-of-fit measures. Methods of Psychological Research Online, 8(2), 23–74. [Google Scholar]
  36. Schoua-Glusberg, A. S. 1992. Report on the translation of the questionnaire for the National Treatment Improvement Evaluation Study. Chicago, IL: National Opinion Research Center. [Google Scholar]
  37. Schreiber, J. B., Nora, A., Stage, F. K., Barlow, E. A. and King, J. 2006. Reporting structural equation modeling and confirmatory factor analysis results: A review. The Journal of Educational Research, 99(6), 323–338. [Google Scholar]
  38. Sparrow, S. S., Cicchetti, D. V. and Balla, D. A. 2005. Vineland Adaptive Behaviour Scales, Second edition: survey forms manual. Circles Pines, MN: American Guidance Service. [Google Scholar]
  39. Tassé, M. J. and Craig, E. M. 1999. Critical issues in the cross-cultural assessment of adaptive behaviour. In: Schalock R. L., ed., Adaptive behaviour and its measurement: Implications for the field of mental retardation. Washington, DC: American Association on Mental Retardation, pp. 161–184. [Google Scholar]
  40. Tassé, M. J., Schalock, R. L., Balboni, G., Bersani, H. Jr, Borthwick-Duffy, S. A., Spreat, S., Thissen, D., Widaman, K. F. and Zhang, D. 2012. The construct of adaptive behaviour: Its conceptualization, measurement, and use in the field of intellectual disability. American Journal on Intellectual and Developmental disabilities, 117(4), 291–303. [DOI] [PubMed] [Google Scholar]
  41. The British Psychological Society (BPS) . 2001. Learning disability: Definitions and contexts. Leicester: The British Psychological Society. [Google Scholar]
  42. Van de Vijver, F. J. R. and Hambleton, R. K. 1996. Translating tests: Some practical guidelines. European Psychologist, 1, 89–99. [Google Scholar]
  43. van de Vijver, F. J. R. and Leung, K. 1997. Methods and data analysis of comparative research. In: Berry J. W., Poortinga Y. H. and Pandey J., eds., Handbook of cross-cultural psychology, 2nd ed. Boston, MA: Allyn & Bacon, pp. 257–300. [Google Scholar]
  44. VandenBos, G. R. 2006. APA dictionary of psychology. New York, NY: American Psychological Association. [Google Scholar]
  45. Wei, Y., Oakland, T. and Algina, J. 2008. Multigroup confirmatory factor analysis for the parent form, ages 5–21, of the Adaptive Behaviour Assessment System-II. American Journal on Mental Retardation, 113, 178–186. [DOI] [PubMed] [Google Scholar]

Articles from International Journal of Developmental Disabilities are provided here courtesy of The British Society of Developmental Disabilities

RESOURCES