The adaptive behavior construct has gained prominent attention in human services over the last several years in Portugal, and its measurement has become an integral part of the assessment of populations with intellectual disability. In Portugal, diagnosis remains exclusively based on IQ measures, although some attention recently has been given to the adaptive behavior concept. In this article, we explain the adaptation and validation process of the Portuguese version of the Adaptive Behavior Scale (PABS) on a sample of 1,875 people with and without intellectual disability. Results of the study are discussed in terms of the reliability and validity of PABS on the sample. The PABS appears to be a valid and reliable assessment of adaptive behavior in individuals in Portugal with intellectual disability.
The construct of adaptive behavior has become increasingly important in the newest definitions and classification systems of intellectual disability (Luckasson et al, 2002; Schalock et al, 2007a; Schalock et al, 2010), which require that limitations in both intellectual functioning and in adaptive behavior (expressed in conceptual, social, and practical adaptive skills) be present for a diagnosis of intellectual disability (American Psychiatric Association, 1994; Luckasson et al, 2002; World Health Organization, 1994). Nevertheless, in Portugal, overreliance on intelligence quotient (IQ) testing continues, and only recently has attention been given to the concept of adaptive behavior. Only since the 1990s has Portugal begun to include an evaluation of adaptive behavior to describe the clients' competencies in community settings, providing a basis for programs and intervention design to be more focused in everyday life (Santos & Morato, 2002; Santos & Morato, 2012; Tassé et al., 2012). In Portugal, the measurement of adaptive behavior is just now emerging (Santos & Morato, 2010).
The modern ecological framework has challenged Portuguese research and practice, emphasizing the role played by the individual's ability to cope with the demands of his or her environment (i.e., independent functioning). It was necessary to have a measure of such behaviors and skills appropriate to the Portuguese context, because persons with intellectual disability were (and are still) segregated and weren't taught to meet society's expectations successfully. Furthermore, adaptive behavior measurement may offer information to develop habilitation training programs and goals. On the other hand, for the first time, in 2007, the Portuguese secretary of state of rehabilitation declared the need to use adapted evaluation scales such as the Adaptive Behavior Scale-Residential and Community (ABS-RC) adapted to Portuguese populations (Moniz, 2007), because of the practical and pedagogic usefulness at both the research and intervention levels.
In Portugal, there is still a tradition of using non-Portuguese standardized tests or scales, frequently mere translations of tools that are essentially North American in origin, which presents several repercussions to be avoided. Furthermore, Portuguese policy toward disability is changing and becoming more focused on adaptive behavior, supports, and quality of life, rather than exclusively focused on academic-scholar content. The modern rethinking of intellectual disability emphasizes the importance of the quality of interactions between individual and everyday environmental demands, not remaining focused solely on “individual deficits.” For Portugal, it was essential to have such an instrument that is valid and reflects the context of Portuguese society. As Schalock et al (2007b, p.13) stated, adaptive behavior limitations must be established through the use of standardized measures.
The Adaptive Behavior Scales (ABS; Lambert, Nihira, & Leland, 1993) are an instrument designed to measure the level of adaptive behavior that emphasizes the observation of spontaneous behavior in day-to-day situations, providing data about the actual/typical behavior (and not potential behavior). Therefore, the ABS-RC (Nihira, Leland, & Lambert, 1993) was selected for study, based not only on a review of the literature (psychometric properties and utility), but also to align with the Portuguese policy framework within the Occupational Activities Centre's certification process (Moniz, 2007). The ABS-RC is one of the most respected and frequently used instruments in research studies (Harries, Guscia, Kirby, Nettelbeck, & Taplin, 2005; Hatton et al., 2001; Cooper as cited in Heal & Tassé, 1999; Tassé et al., 2012); is one of the most comprehensive instruments of its type (Roszkowski, 1982); involves extra domains such as domestic and sexual activity (when compared with the scholar version of ABS), considered by the Portuguese as important in all persons, regardless of age; allows for good flexibility (it could be used as a whole or using specific domains); and its items are pertinent to evaluation of adaptive behavior (Nihira et al., 1993).
In addition, according to Luckasson et al (2002), ABS-RC is one of the instruments with adequate psychometric qualities shown by several studies (Lambert et al., 1993; Nihira et al., 1993; Roszkowski, 1982). Its internal consistency, inter-rater reliability, and stability characteristics ranged from .80 to .99 (Nihira et al., 1993), with normative tables for the population with intellectual disability. These findings were supported by other cross-cultural studies that also reported good psychometric properties, such as in Korea (Shin, 2002), China (Wan, 2009), Iran (Sadrossadat, Moghaddami & Sadrossadat, 2010), and Finland (Saloviita, 2002), where most measures of internal consistency ranged from .53 (part two of the Finnish study) to .99. Tassé et al. (2012) also pointed out the scholar version of ABS as “one of the four comprehensive individualized, standardized, and psychometrically sound adaptive behavior scales available [that…] have been developed specifically for the purpose of ruling in or out a diagnosis of intellectual disability” (p. 293).
Therefore, our goal is to assess the psychometric properties of the Portuguese version of the ABS as a measurement instrument of adaptive skills of persons with intellectual disability in Portugal, because it is expected that scores should have an effective contribution for planning intervention purposes.
The normative sample comprised of evaluations of 1,800 persons with intellectual disability and 75 without intellectual disability (807 females, 1068 males). Their age ranged from 6 to 60 years (M = 29.94, SD = 9.66). The most common level was the moderate classification1 (42.5%), followed by severe/profound (16.4%), mild (14.5%), with multiple disabilities (13.9%), with Trisomy 212 (8.7%), and without disability (4.1%) categories. All adults with intellectual disability in this study were living in institutional settings. The majority was in Occupational Activities Centers (provided by institutions to increase the quality of life of persons with disabilities through useful social and leisure' activities with no productivity demands, 76.3%) vs. 2.1% at Vocational Training (i.e., institutional structure that teaches skills relevant for employment for persons with disability).
The structure of the Portuguese Adaptive Behavior Scale (PABS) is similar to the original scale and is divided in two parts (the factor structure is currently being analyzed). Part I focuses on personal independence and comprises 10 domains (independent functioning, physical development, economic activity, language development, numbers and time, domestic activity, prevocational activity, self-direction, responsibility, and socialization). Responses to items take two forms: rating the highest level of adaptive behavior or a checklist of yes/no responses that are summed to form the item score. Part II is concerned with challenging behaviors. The behaviors in Part Two are grouped in eight domains (social behavior, conformity, trustworthiness, stereotyped and hyperactive behavior, self-abusive behavior, sexual behavior, social engagement, and disturbing interpersonal behavior). Responses to items are made by rating them according to frequency (never, occasionally, or frequent). The behaviors within both parts are grouped in the same way as the original scale.
The goal of PABS is to describe adaptive behaviors of the Portuguese population with intellectual disability, allowing a comprehensive evaluation of how they deal with their environmental conditions through the analysis of behaviors, situations and daily activities.
Translation and Adaptation of PABS
Considerations of context, language, and culture are all highly relevant when adapting a scale, due to the societal expectations of each cultural group and the meaning/relevance of items and, thus, it requires more than a simple translation (Tassé & Craig, 1999). The process of translation, adaptation, validation, and administration of the PABS followed the guidelines of the International Test Commission (2010). Special attention was given to cultural and linguistic adaptation. At the beginning of the study, several focus groups were conducted across the country to clarify and explain the concepts and goals, allowing consensus recognition of relevant adaptive behaviors. The first step was the initial translation of ABS-RC into Portuguese (sequential approach), which was made by an expert professional committee (professors and PhD students) of all native Portuguese speakers. Each member conducted an initial translation and adaptation of the scale. Then committee members compared their individual translations in order to reach a consensus of a preliminary version. Any disagreements were solved through a face-to-face work session. Then a back translation was made to ensure that the Portuguese version reflected the same item content as the original scale.
The language differences and cultural context of Portugal led to the development of a Portuguese version with some modifications of grammatical construction of sentences, addition of new items (e.g., “eats compulsively,” “blows up,” and “brushes her/his hair,” as recommended by expert practitioners); clarification of some concepts and terms (e.g., hyperactivity); use of euros (€) in the Economic Activity section; minimization of sentences beginning with “no”; and division of some original items into two (e.g., “uses table knife for cutting or spreading” was converted into “uses table knife for cutting” and “uses table knife to spread”; “opens zippers and unbuttons” was converted into two different item) due to social values, which led to introduction of “more” scores and some reordering by complexity level. Also an item was added in the Independent Functioning section—“does not participate in laundry”—because it is usual that females (e.g., mothers, sisters) take care of this kind of task, rather than males. Further, in the Physical Development domain, items were added for evaluating the right/left recognition, body awareness, and motor fine/gross skills (e.g., draw, grab a pencil, cuts with a scissor, kicks a ball, among others). The economic activity domain also required adaptations simplifying some tasks, because Portuguese people with intellectual disability typically do not use money.
Next, the new version was analyzed by 10 experts selected based on their experience within populations with intellectual disability and with academic experience (PhD) in assessment and intervention (i.e., research methodology, tests validation and standardization, intervention) in the special education field, particularly with persons with intellectual disability. A consensus version was determined and field-tested by being applied to a small sample of persons with intellectual disability (n = 28) to test the understandability of the instructions, acceptability of time limits, wording of the items, consistency of responses, and so on, ensuring that this new adapted version still retained its equivalence in an applied situation. Lastly, a final version was determined based on the results of the field test.
Part I measures adaptive behavior and Part II focuses on challenging behavior, and it was hypothesized that PABS values should intercorrelate significantly within Parts I and II, but that Part I should intercorrelate modestly and negatively with Part II. This was considered a test of construct validity because both types of behaviors do not relate to each other. Secondly, PABS should differentiate people with and without intellectual disability in adaptive and challenging behaviors. Finally, items within each PABS domain and factor measure similar traits and, thus, items should relate highly to the scale's subscores.
Consistent with the research ethic approval, informed parent/caregiver consent was obtained for each participant. Each individual was administered all items in accordance with the guidelines of the original scale. The PABS was administered, in the majority of cases, through an interview. Each item was completed using another person's knowledge (e.g., parents, caregiver, or a key worker who knew the participant well—at least 2 to 3 years) about the individual being evaluated through a questionnaire. The majority of scales were completed by practitioners (n = 1795) rather than by parents (n = 100). Statistical analyses were carried out using the Statistical Package for the Social Sciences (SPSS version 17.0).
A two-stage process for estimating content validity was performed: a development stage based on literature reviews, interviews, and focus group (including persons with intellectual disability, experts, parents, and practitioners), and a judgment stage where a panel of experts reviewed the purpose and objectives of the study. Each one received a copy of the scale and was asked to rate each item's relevance to the domain of content using a Likert-type rating scale (1 = not relevant, 2 = somewhat relevant, 3 = quite relevant, and 4 = very relevant). Experts were also asked to share qualitative comments about items and the overall scale as well as to rate the clarity of items and the comprehensiveness of the measure (Wynd, Schmidt, & Schaefer, 2003). All raters agreed that all domains were “content relevant,” and the few disagreements were about the need to clarify some concepts or to hierarchically reorganize some items or change the beginning of sentences (e.g., sentences do not begin with “no” in Portuguese).
Mean values and standard deviation of adaptive behavior scores of the two groups (with and without intellectual disability) are presented in Table 1 and show some significant differences between them (p<.01). The internal consistency reliability (using only data from participants with intellectual disability) of PABS domains (all items included) was investigated using the coefficient alpha procedure. Reliabilities above .80 are considered good, and very good/excellent above .90 (Hill & Hill, 2009, p. 149; Pestana & Gageiro, 2005, p. 528). Nunnally (1978) emphasized that, for research purposes, reliabilities above .70 are adequate, but when it comes to important clinical decisions, reliability should be at least .90. As can be seen in Table 1, the instrument demonstrated good to excellent internal consistency (all domain reliabilities exceeding .80, except the Self-Abusive Behavior domain with a Cronbach's α = .79). The other domains showed high scores of Cronbach's α, ranging all above .80 to .97 (Independent Functioning). Stability was measured using a test-retest reliability. The retest sample consisted of 28 adults (16 females and 12 males) with an intellectual disability, who ranged in age from 21 to 51 years (M = 34.54; SD = 9.594). Two test administrations of the PABS, separated by approximately a 2- to 3-week period were obtained. Both interviewer and respondent pairs were the same. Retest reliabilities were calculated from the Pearson's correlation coefficients. Part I domains ranged from .40 to .81 and Part II domains yielded low correlations (p<.51). Findings reported in the original ABS study of Nihira and colleagues (1993) were higher (ranging from .81 to .99) in both parts.
The structure of PABS was investigated with an exploratory factor analysis conducted using a principal components extraction method. A varimax rotation was employed to simplify the factor structure. The analyses were applied in the total sample. The Kaiser-Meyer-Olkin index and Bartlett's test of sphericity were used to evaluate sampling adequacy for a satisfactory factor analysis. Criteria for extraction included eigenvalues greater than 1.0, a minimum of 5% explained variance per component, components loadings of .40 and above, and a difference of at least .10 in cross-loadings across factors. Results of analysis revealed that three components of Part I had eigenvalues beyond 1.00, accounting for 56% of the variance. Domains of Part II revealed a three-component structure accounting for 41% of the overall variance. The factor loadings are presented in Table 2 and seem to support the five-factor model proposed by the original ABS authors (Lambert et al., 1993). Internal consistency of these five factors ranged from .90 (Personal Adjustment) to .98 (Community Self-Sufficiency)
Analysis of the intercorrelations showed that the domains of Part I are strongly correlated (with the exception of Physical Development) and the domains of Part II are moderately to negatively correlated (see Table 3). Data were consistent with the hypothesized directions: items were significantly intercorrelated within domain, but they show a modest and negative correlation between each other (Table 3). The PABS differentiates people with and without intellectual disability across domains (Table 1) and, items of each PABS domain and factor relate highly to scale subscores (Table 3)
According to Lambert and colleagues (1993, p. 50) “the results of adaptive behavior scale should clearly distinguish between people with and without intellectual disability.” As shown in Table 1, mean scores of the group with intellectual disability are sufficiently below the average scores of the group without intellectual disability, and differences in the mean scores of the two groups are statistically significant (p<.01). In general, there is sufficient variability within both groups. These data support Lambert and colleagues' findings.
To our knowledge, this study is the first that addresses the study of adaptive behavior measurement and the application of an adaptive instrument in Portugal among a sample of persons with intellectual disability. Although the definition of intellectual disability demands the need of a multidimensional system of evaluation involving intelligence and adaptive skills measures, Portugal is still using, exclusively, IQ measures for an intellectual disability diagnosis. This fact stresses the relevance of having an adaptive behavior instrument in Portugal. Based on findings, it seems that the PABS scale could be used for assessing and screening the adaptive behavior of persons with intellectual disability in Portugal, appearing to be useful in identifying educational or training-related goals (Nihira, 1999).
Findings of the present study suggest that the PABS has an overall good reliability, although further studies should be conducted in the field. Reliability analyses demonstrated the satisfactory test-retest reliability and internal consistency coefficients, ranging from .92 to .99. The reliability data reported confirm the findings of earlier studies for Part I (Godfrey et al., 1986; Hatton et al., 2001; Nihira et al., 1993; Spreat, 1982; Wan, 2009; Watkins, Ravert & Crosby, 2002). There was a consistently high correlation between Part I domains, except for Physical Development, that showed correlations coefficients considered as weak (below .50) with Economic Activity and Numbers/Time.
The correlations in Part II domains were not as high as in Part I, but the findings of exploratory analysis indicate that they seem sufficiently reliable to justify the important clinical information provided by this documentation of problem behaviors. Content validity was based on expert agreement about the scale.
In general, there is sufficient variability within both groups in all the PABS domains (especially within Part I). Analyzing mean scores confirms group differentiation in independence and adaptive behaviors and, therefore, demonstrates potential practical-clinical significance, that is, PABS appears to be able to differentiate groups: people without intellectual disability score higher than persons with intellectual disability on adaptive behavior scores. Comparing the results of this study with the results from the study of Nihira and colleagues (1993) revealed that the results closely resemble each other. In general, we found that basic properties for PABS were the same as those of the original scale.
On the other hand, preliminary separate item analyses were conducted for both parts of PABS. In Part I, items showed moderate to high correlations with their own domain score and low scores with other domains, although further study is recommended for some items. Many of the items on PABS Part II that revealed a low degree of relationship to their own domain did so because they occurred occasionally or never.
The principal components analysis of the scale scores produces findings consistent with previous studies, but further studies should be conducted to prove the normative factor structure of PABS. We are now conducting the confirmatory analysis process to study the dimensional structure of the Portuguese version aiming to contribute to a more evidence-based use of the scale in this population, following Portuguese governmental recommendations (Moniz, 2007).
Test users should be cautious in the interpretation of data, especially when referring to norm comparison and, consequently, diagnosis. We suggest that test users use it in a more descriptive rather than quantitative way. The users of total scores must be aware of the meaning and limitations of final indexes (norms). Concerning diagnosis, great attention should be given to PABS administration and the development of (more) precise evaluation criteria and specific training for raters should be considered.
Only 4.1% of the persons evaluated by the PABS were in mainstream settings at the time of the study. With the introduction of new inclusive laws in 2008 that established that all children, regardless of their disability, should be placed in regular schools, it is recommended that a new study on the scholarly population with and without intellectual disability be conducted to analyze the possible differences in adaptive behaviors.
Several practical implications can be drawn from this study. The use of PABS to identify the adaptive behavior level provides guidance for the development of the individualized plan and person-centered planning strategies for people with intellectual disability. It also provides the opportunity to re-evaluate the principles and goals of education and interventions in Portugal, evaluating the effectiveness of programs and interventions through a comparison between adaptive behavior scores over time. Further, the results allow comparisons of an individual's behavior across settings, and it might offer a common language among all practitioners. According to the descriptive literature, comparisons between common models of intervention between the two countries suggest that education and intervention models in the United States may more frequently focus on community living adaptation factors, while the Portuguese tradition currently emphasizes more academic content, especially in the learning of mathematics, writing, and reading.
The future challenge will be re-evaluating the persons who were in the past diagnosed as having an intellectual disability based solely on IQ measures and now adding the adaptive behaviors results. This may lead to changes in the diagnosis of some individuals, and will certainly provide more information for individualized supports and interventions.
As Baker (1996) pointed out, the adaptive behavior concept, the paradigm of supports, and the conceptualization about quality of life “challenges us to look beyond our traditional perspectives.” We add that, in the case of Portugal, we are challenged to readapt the field of intellectual disability and we hope that reconsideration of the measurement of adaptive behavior can positively affect the concept of intellectual disability and professional practice in Portugal.
In Portugal, the classification system of mild, moderate, severe, and profound intellectual disability is still used, so our data were organized this way.
Because Trisomy 21 is “easily” diagnosed, most individuals with Trisomy 21 do not have the label by “level classification” but by etiology.
Sofia Santos, Human Kinetics Faculty, UIDEF, University of Lisbon, Portugal; Pedro Morato, Human Kinetics Faculty, UIDEF, University of Lisbon, Portugal; and Ruth Luckasson, Department of Special Education, University of New Mexico.
Address correspondence concerning this article to Sofia Santos, Human Kinetics Faculty, University of Lisbon, Estrada da Costa 1499-002, Cruz Quebrada, Lisbon 1499-002, Portugal (e-mail: firstname.lastname@example.org).