Using the Multidimensional Students' Life Satisfaction Scale—MSLSS (Huebner, 1994), we compared life satisfaction reports of 80 high school students with mild mental disabilities with a matched sample of 80 typically achieving students. The results provided preliminary support for the use of the MSLSS in research with secondary school students with mild mental disabilities. Comparisons of mean levels of general and domain-specific satisfaction suggested that the students with mild mental disabilities reported comparable positive levels with two exceptions. They reported lower satisfaction with their friendships and higher satisfaction with school experiences than did their typically achieving counterparts. The school satisfaction reports of students with mild mental disabilities varied as a function of differences in placement in special education.
Editor in charge: Steven J. Taylor
In a guest editorial in Mental Retardation, Landesman (1986) noted that quality of life and personal life satisfaction were the new “buzz words” in the field of mental disabilities. Although quality of life constructs continue to impact research and practice related to individuals with mental disabilities, researchers have argued that quality of life has yet to be defined consistently, concisely, or comprehensively in applications to individuals with mental disabilities (e g., Hawkins, Kim, & Eklund, 1995).
Two major approaches have been used in defining and assessing quality of life: objective and subjective indicators (Diener & Suh, 1997; Schalock, 1990). Examples of commonly studied objective indicators include measures of income, health, integrated activities, friendship networks, quality of residential environment, and access to health care. The subjective indicators approach focuses on individual's internal evaluations of various life experiences, including satisfaction with life as a whole and satisfaction with specific life circumstances (e.g., family life, school experiences). Both approaches offer unique and complementary information; however, subjective indicators have been particularly neglected in research concerning the well-being of children and youth in general (Ash & Huebner, 1998).
Although considerable research has been devoted to the assessment of the nature, correlates, and consequences of individual differences in subjective quality of life among adults, little research has been conducted with children and youth, particularly children and youth with disabilities. Research with children and youth with disabilities (e.g., mental disabilities, emotional difficulties) is crucial to understanding their perceptions of the quality of their lives, including their potential unique needs and resources.
As a prerequisite to studying the levels as well as determinants and consequences of individual differences in subjective quality of life of individuals with disabilities, research is needed that establishes the reliability and validity of quality of life measures for youth with particular disabilities. Numerous authors have voiced concerns over the use of measurement instruments with individuals who have disabilities, which have been developed and validated with individuals who do not have disabilities (e.g., Prout & Strohmer, 1991). The psychometric properties of subjective quality of life scales cannot be assumed to generalize to various groups of children and youth with disabilities. For example, research by Harter and her colleagues (e.g., Renick & Harter, 1989; Silon & Harter, 1985) has revealed differences between the responses of children with learning disabilities and children with mental disabilities to her multidimensional self-concept scale (Harter, 1982). Such findings suggest the possible need for modifications to self-report measures as a function of handicapping conditions. Specifically with respect to children with mental disabilities, issues in application have included item content (e.g., use of concrete versus abstract terms; length of questions); item structure (e.g., acquiescence problems); and item difficulty related to comprehension and/or reading levels (e.g., see Hawkins, Kim, & Eklund, l995).
The possibility of differential reliability and validity of subjective quality of life scales is supported by Griffin and Huebner's (2000) results. They found that the Multidimensional Students' Life Satisfaction Scale—MSLSS (Huebner, 1994) had to be revised for use with middle school students who had serious emotional disabilities Because reliability coefficients were significantly lower than those reported for middle schoolers without disabilities, they had to eliminate five items, all of which were reverse scored, to achieve acceptable reliability estimates. Griffin and Huebner hypothesized that the students with serious emotional disabilities may have experienced difficulties with the reverse-scored items because of lower cognitive abilities, similar to those reported for young children (cf. Marsh, 1986).
There is a paucity of subjective quality of life measures that (a) are suited for children and youth, (b) assess multiple dimensions of life circumstances, and (c) have been subjected to rigorous empirical scrutiny (see Gilman & Huebner, in press). One exception is the previously mentioned MSLSS (Huebner, 1994), which was designed for students in Grades 3 through 12. Although additional research is needed, studies have provided promising evidence of reliability and validity (see also Gilman, Huebner, & Laughlin, 2000; Huebner, Laughlin, Ash, & Gilman, 1998). In addition to studies with students in the United States, evidence of validity has been provided with children in Canada (Greenspoon & Sasklofske, 1997), Korea (Park, 2000), and Spain (Casas, Alsinet, Rosich, Huebner, & Laughlin, 2001). Also, the MSLSS has been used in research with students who have emotional difficulties (Griffin & Huebner, 2000) and gifted students (Ash & Huebner, 1998).
To date, no researchers have reported on the use of the MSLSS with students who have mental disabilities. Thus, one of our major goals in the current study was to explore the reliability and validity of the MSLSS with adolescents who have mild mental disabilities. Specifically, we sought to determine (a) the internal consistency reliability of the MSLSS domain and total scores with this population; (b) the separability (i.e., multidimensionality) of the MSLSS domain scores (i.e., the ability of the students to differentiate the MSLSS domains); and (c) criterion-related validity of the MSLSS (i.e., convergence of the scale's scores with measures of related constructs (i.e., global self-worth and perceived social support). Specifically, we anticipated that the MSLSS Self domain would correlate significantly with a measure of global self-worth, whereas the MSLSS Family, School, and Friends domains would correlate significantly with measures of parent, teacher, and friend support, respectively.
Pending evidence of support of the reliability and validity of the MSLSS with adolescent students who have mild mental disabilities, there were two additional purposes of the study. First, we sought to compare the global and domain-specific life satisfaction reports of adolescents with mild mental disabilities and typically achieving adolescents without an identified disability. Given the limited theory and empirical work investigating the perceived life satisfaction of adolescents in general as well as adolescents with mild mental disabilities in particular, we did not formulate specific hypotheses regarding mean level comparisons across groups.
Second, we examined the relationships between global and domain-specific life satisfaction reports and types of placements in special education settings for the students with mild mental disabilities. Findings to date concerning the social–emotional correlates of special education placements have been equivocal (see Freenan & Alkin, 2000, for a review of this literature). For example, some investigators have suggested that special class placement of students with mild mental disabilities is associated with negative social–emotional outcomes (e.g., reduced self-esteem, perceived social competence and acceptance, academic efficacy), whereas others have reported contradictory findings. Given the separability of life satisfaction reports from traditional mental health indices (e.g., Huebner, Funk, & Gilman, 2000; Huebner, Gilman, & Laughlin, 1999), we thought that multidimensional life satisfaction reports would provide a unique and useful perspective for evaluating correlates of special education programs. Again, given the largely exploratory nature of this study, as well as the equivocal research, we did not develop specific hypotheses related to the relationship between life satisfaction reports and types of special education placements of students with mild mental disabilities.
The participants included 160 students in Grades 9 through 12, 80 of whom had received a special education classification of mild mental disability by a multidisciplinary team based on Individuals With Disabilities Education Act (IDEA) criteria. The students were drawn from schools in one urban and one rural school district in South Carolina and one nearby rural district in Georgia. A comparison group of 80 typically achieving regular education students was drawn from one urban and one rural school in South Carolina and was matched with the mild mental disabilities group on race and gender. Previous research has demonstrated that age, unlike race and gender, is unrelated to MSLSS scores; therefore, students were not matched on age (Huebner, 1994; Huebner et al., 1997).
The sample of students with mild mental disabilities consisted of 56 African Americans, 23 Caucasians, and one student classified as “other.” Females comprised 59% of the sample. The mean age of this sample was 15.8 years (standard deviation [SD] = 1.07).
The sample of typically achieving students consisted of 56 African Americans, 23 Caucasians, and one student classified as “other.” Females constituted 59% of the sample. Their mean age was 16.5 years (SD = 1.76).
The students with mild mental disabilities were classified further based on the amount of time spent in regular education classes per day, ranging from no hours to 5 hours (M = 1.4). None of the students who were classified as having mild mental disabilities attended all regular education classes. Their placements ranged from self-contained to resource programs. Data on socioeconomic status were not available due to the refusal of the school districts to grant permission to collect this information.
Multidimensional Student's Life Satisfaction Scale. The MSLSS is a 40-item self-report scale designed to measure life satisfaction in children and adolescents. Differing from unidimensional scales that measure only global life satisfaction, the MSLSS is used to examine five separate domains that are thought to be important in a child's life (Friends, Family, School, Self, and Living Environment), along with general (Total) life satisfaction. Examples of items from each domain include: Family (“My family gets along well together”), Friends (“I wish I had different friends”), School (“I look forward to going to school”), Self (“I like myself”), and Living Environment (“There are lots of fun things to do where I live”). Students rated their satisfaction across each area using four options (never, sometimes, often, almost always). To determine the domain scores on the MSLSS, we summed the ratings within each domain and divided by the number of items in that domain. The total score was calculated by summing all the item scores and dividing by the total number of MSLSS items. Therefore, a high score indicates a high degree of satisfaction.
Research has indicated acceptable psychometric properties for the MSLSS (see Bender, 1997, for a comprehensive review). Coefficient alphas in the .80s for the five domain scores have been reported by Huebner (1994). Convergent validity has been supported by predicted correlations with measures of different, but related constructs (Gilman et al., 2000; Huebner, Laughlin, Ash, & Gilman, 1998). Support for the construct validity of the measure has also been shown through exploratory and confirmatory factor analyses (Greenspoon & Saklofske, 1997; Huebner et al., 1998). The readability of the MSLSS is estimated to be at the 1.5 grade level.
Self-Perception Profile for Adolescents (Harter, 1988). The five-item Global Self-Concept subscale of the Self-Perception Profile for Adolescents was used as a criterion measure for the MSLSS Self domain. Wichstrom's (1995) revision of the Self-Perception Profile for Adolescents, in which each item was presented as a self-description (e.g., “I like the kind of person I am”), was used to increase the likelihood that the students with mild mental disabilities could respond effectively. As with the original version, the revised scale had four options: 1 (describes me very poorly), 2 (describes me quite poorly), 3 (describes me quite well), and 4 (describes me very well). The results of Wichstrom's study involving adolescents in Norway indicated that the revised version, applying the more simplistic format, proved to have better reliability and validity than the original version.
Preliminary reliability analysis of the Global Self-Worth domain with the current mild mental disabilities sample indicated an internal consistency (alpha) estimate of .65. Because of its low reliability, which was considered unacceptable for a criterion measure, no further analyses with this measure were conducted.
Social Support Scale for Children (Harter, 1985). The Social Support Scale for Children is a 24-item multidimensional, self-report scale designed to measure perceived social support in children. The scale contains four domains: Parents, Classmates, Teachers, and Friends. The response format is identical to the Self-Perception Profile for Adolescents (Harter, 1988); thus, the original version was modified in the same manner as the Self-Perception Profile for Adolescents (see above). As evidence of criterion-related validity, the Parent Support domain was to be compared to the MSLSS Family domain; the Teacher Support domain to the MSLSS School domain; and the Friends Support domain to the MSLSS Friend domain. The Classmate subscale was not administered because it did not have a related scale on the MSLSS.
Preliminary reliability analysis of the Parents, Teacher, and Friends domains with this mild mental disabilities sample indicated alpha coefficients of .53, .62. and .69, respectively—much lower than those for typically achieving students. Again, no further validity comparisons with the Social Support Scale for Children were conducted because of the low reliability coefficients.
The first author met with all students to explain the purpose of the study and to hand out consent/assent forms. Students were required to return both a completed parent consent and student assent form to be eligible to participate in the study. After the forms were collected, both samples of students completed the MSLSS. In addition, the mild mental disabilities sample completed the Social Support Scale for Children and Self-Perception Profile for Adolescents questionnaires.
For the mild mental disabilities sample, administration was conducted in small groups of 5 to 10 students to ensure that each student understood the questions and formats of the instruments. The first author read all instruments to the students. The total administration time ranged from 30 to 45 minutes. All instruments were completed in a confidential manner.
The comparison students were from randomly chosen English classes. Students with any identifiable disability were excluded from participation. The MSLSS was administered to all students in the classes who had returned parent consent and child assent forms. Instructions for completing the MSLSS were read to the students, and the first author remained in the room to answer questions. The total administration time ranged from 15 to 20 minutes. Like the mild mental disabilities sample, the comparison group completed the MSLSS in a confidential manner. Following completion of the study, all students were thanked for their participation and provided with a small tangible reward (i.e., soft drink and candy bar).
Internal Consistency Reliability of the MSLSS
Coefficient alphas were obtained for the MSLSS total score as well as for each domain. Coefficient alphas of .83 and .92 were found for the MSLSS total score for the mild mental disabilities and typically achieving samples, respectively, suggesting good overall internal consistency. Alpha coefficients for the Life Satisfaction domains were initially computed using all component items from the five domains of the MSLSS. Resulting alpha coefficients yielded acceptable internal consistency estimates for the mild mental disabilities group in the Family (.76), Friends (.76), and School (.71) domains. However, the resulting alpha coeffcients yielded lower internal consistency reliabilities for the mild mental disabilities group in the Living Environment (.63) and Self (.65) domains. Alphas for the domains for the typically achieving group were all acceptable, ranging from .73 (Self) to .85 (School).
To improve upon initial internal consistency estimates for the mild mental disabilities group, we selectively deleted items with the weakest item-total correlations in their respective domains (<.30) and recalculated coefficient alphas. The recalculated alpha coefficients for the adolescents with mild mental disabilities are provided in Table 1. Specifically, three items (“There are lots of fun things to do where I live,” “This town is filled with mean people,” and “My family's house is nice.”) were omitted from the Living Environment domain, resulting in no improvement in the alpha coeffcient for the mild mental disabilities (.63 vs. .63) or typically achieving (.84 vs. .82) groups. In the Self domain, deletion of a single item (i.e., “Most people like me.”) resulted in little or no improvement for the mild mental disabilities group (.65 vs. .66) and the typically achieving group (.73 vs. .66). Given the lower reliability coeffcients in the Living Environment domain for students with mild mental disabilities and in the Self domain for both groups, scores on these domains should be interpreted cautiously.
Intercorrelations Among MSLSS Domain Scores
Intercorrelations among the five MSLSS domains for the mild mental disabilities and typically achieving groups are provided in Table 1. In the mild mental disabilities group, correlations ranged from −.04 to .44, with a median correlation of .28. In the typically achieving group, correlations ranged from .26 to .55, with a median correlation of .45. The modest to moderate correlations indicated that students with mild mental disabilities and typically achieving students rated the MSLSS domains as related, yet separable areas of life satisfaction (Tabachnick & Fidell, 1996), thus providing preliminary support for the five-dimensional structure of this scale with both groups.
Comparisons of the magnitude of the domain intercorrelations across the mild mental disabilities and typically achieving groups were made using Fischer's Z transformation. With a Bonferroni-corrected alpha of .005, 3 of the 10 comparisons were significant. In all three cases (Friends/School, Friends/Self, School/Living Environment), the coefficients were lower for the students with mild mental disabilities. Their lower domain intercorrelations may suggest that students with mild mental disabilities make greater distinctions among specific satisfaction domains.
Mean life satisfaction scores were calculated for the total score and each of the five domains of the MSLSS by summing the item responses and dividing by the number of items in that domain. The mean total and domain scores for each group are provided in Table 2. Consistent with findings from other studies of adolescents (Gilman et al., 2000; Huebner et al., 1998), mean ratings for both groups of students fell in the positive range of life satisfaction for all domains. A multivariate analysis of variance (MANOVA) with group membership (mild mental disabilities vs. typically achieving) as the independent variable and the five MSLSS domains as the dependent variables was conducted to determine whether there was a significant overall multivariate effect. The results of the MANOVA were significant, Wilks' lambda (5,154) = .68, p < .0001. Using Bonferroni-corrected alpha levels (.05/5 = .01), we found that follow-up analyses of variance (ANOVAs) indicated that the Friends, F(1, 158) = 26.04, p < .0001, and School, F(1, 158) = 13.44, p < .0003, domains were significantly different for the two samples. Students with mild mental disabilities reported lower satisfaction with their friendships than did their typically achieving peers (Ms = 3.08 and 3.51, respectively). The opposite was found in the School domain, where students with mild mental disabilities indicated higher satisfaction with their school experiences than did their typically achieving peers (Ms = 2.86 and 2.53, respectively). In addition, an ANOVA comparing Total scores on the MSLSS for students with mild mental disabilities and typically achieving students did not reveal a significant difference between overall life satisfaction scores.
Educational Placement Differences
A MANOVA was performed to assess the relationship between three levels of time spent in regular versus special education placements and domain-specific life satisfaction reports of the mild mental disabilities sample. The students were divided into three groups based on the approximate number of hours they spent in regular education per day (1 = 0 hours; 2 = 1 to 2 hours; 3 = 3 to 5 hours). A significant overall placement effect was obtained, Wilks' lambda (5,74) = .74, p < .04. A follow-up Bonferroni-corrected ANOVA indicated a significant School domain effect, F(2, 77) = 5.15, p < .008. Scheffé‘s post hoc comparisons of the means revealed that the school satisfaction of students with mild mental disabilities who were in a self-contained special education setting was significantly higher than that of peers who spent 3 or more hours in a regular education setting (Ms = 3.14 and 2.6, respectively). Table 3 displays mean life satisfaction total and domain-based scores for special education placements. A separate ANOVA was also conducted using the total life satisfaction score as the dependent variable and the three placements as the independent variable. The results revealed no significant differences.
One major purpose of this study was to provide some preliminary data on the reliability and validity of the MSLSS for use in research with adolescent students who have mild mental disabilities. Following scale modifications (i.e., deletion of 4 MSLSS items), the obtained reliability coefficients for the MSLSS domains were acceptable for research purposes, with the exception of the Living Environment and Self domains. Although the reliability of these two scales approached acceptable levels (i.e., r = .70), the current findings suggest that the students with mild mental disabilities in this sample failed to conceptualize their responses to items in the Living Environment and Self domains in as consistent a manner as did their typically achieving counterparts.
Regarding validity, the magnitude of the intercorrelations among the five life satisfaction domains of the MSLSS provides promising, yet tentative support for the separability (i.e., multidimensionality) of the domains for students with mild mental disabilities. The lower domain intercorrelations for students with mild mental disabilities as compared to the typically achieving students suggest greater validity in that the students with mild mental disabilities in this sample made greater distinctions among specific satisfaction domains.
Taken together, these findings provide tentative support regarding the reliability and validity of the MSLSS for research purposes. Although this scale demonstrates promise as a research tool for use with students with mild mental disabilities, additional modifications (e.g., additional items for Living Environment and Self domains) may need to be considered. This study, therefore, serves as an important first step in the validation of a measure to assess perceived quality of life that is appropriate for adolescents with mild mental disabilities.
Although this study provides some support for using the MSLSS with high school students who have mild mental disabilities, several limitations should be addressed in future research. First, our samples were drawn from three school districts in two southeastern states. Studies of life satisfaction ratings from larger and more nationally representative samples of students with mild mental disabilities should enhance the external validity of the findings. Second, we did not explore the stability of life satisfaction ratings for students with mild mental disabilities. Investigations of short-term and longer-term reliability should help clarify the nature of the construct for students with mild mental disabilities. Third, the moderate sample size for the students with mild mental disabilities precluded more sophisticated analyses of the multidimensionality of the MSLSS. For example, future studies with larger sample sizes would enable the assessment of the dimensionality of the MSLSS through factor analytic procedures. Fourth, the lack of reliable criterion measures hampered the study of criterion-related validity. Future researchers need to identify suitable criterion measures for this population. Finally, although life satisfaction is important in and of itself, life satisfaction reports of students with mild mental disabilities should be compared to indices of adaptive functioning, such as school completion, academic achievement, discipline records, classroom behaviors, and social acceptance. Such analyses would allow for comparisons between satisfaction and important academic and behavioral outcomes, thus assessing the adaptive implications of global and domain-based life satisfaction for this group of students.
Implications for assessment practices are suggested by this research. First, information gained from multidimensional life satisfaction measures may offer researchers, and perhaps eventually practitioners, the opportunity to gain more comprehensive information concerning the well-being of adolescents. Such information should complement objective indices of quality of life for children with mild mental disabilities. Such information should also complement the more traditional subjective measures that focus on psychopathological symptoms. Life satisfaction judgments offer a unique perspective from which to view adolescents' well-being.
Nevertheless, one of the major implications of this study involves the need for researchers to exercise caution when using psychological well-being measurement instruments designed for typically achieving students with special populations. Potential users of self-report scales in particular should never assume that the reliability or validity of well-being measures will generalize to youth with special needs. Rather, these measures should be examined empirically because modifications may be necessary to meet psychometric requirements for research or clinical applications.
Comparisons of the mean levels of life satisfaction across the mild mental disabilities and typically achieving samples yielded several significant findings. First, both samples reported positive levels of overall life satisfaction. This finding is consistent with prior research demonstrating that global life satisfaction reports are typically positive for adults with and without disabilities (Diener & Diener, 1996). Thus, the current results extend this trend to adolescents with and without disabilities. Second, students with mild mental disabilities and typically achieving students did report significant differences on two of the five life satisfaction domains: Friends and School. Compared to their typically achieving counterparts, the students with mild mental disabilities reported lower satisfaction with friends, but greater school satisfaction. Third, we conducted additional analyses to investigate the reports of the life satisfaction of students with mild mental disabilities in relation to their special education placements. Analyses of differing amounts of time spent in regular versus special education placements revealed one significant difference. Students in self-contained special education placements were more satisfied with their school experiences.
The differences in both friendship and school satisfaction are difficult to interpret without further information. The meaning of the findings related to satisfaction with friends is somewhat unclear because the MSLSS does not require the participants to identify the particular friends considered in responding to the items. Thus, students with mild mental disabilities who are mainstreamed may be evaluating friends with mild mental disabilities, typically achieving friends, or both. Similarly, with respect to school satisfaction, the MSLSS cannot be used to assess the manner in which students with mild mental disabilities formulate their school satisfaction judgments (e.g., compared to school lives of friends with or without mild mental disabilities, “ideal” school, past school experiences). According to Silon and Harter (1985), a child with mild mental disabilities may differ from other students in terms of the psychological processes leading to self-evaluative judgments. Similar differences may also operate in formulating life satisfaction judgments. Future research should be directed toward the exploration of the bases upon which students with mild mental disabilities formulate their domain-based life satisfaction judgments (i.e., friends, school) so as to clarify the meaning of differences relative to other youth.
In conclusion, results of this study provide implications related to educational services for adolescent students with mild mental disabilities. These findings underscore the potential complexity of outcomes associated with various special education placements because they suggest that placements may be associated with multiple co-existing outcomes for adolescents with mild mental disabilities, including both positive and negative outcomes (e.g., greater dissatisfaction with friends and greater satisfaction with school). Although these results do not lead to a simplistic conclusion about the desirability or undesirability of special education placements, the results do encourage the further study of the relationship between special education services and multidimensional perceptions of quality of life of adolescents with mild mental disabilities.
Note: Intercorrelations Among Life Satisfaction Domains for Students with mild mental disabilities and Typically Achieving (NA) Students
Authors:Ashley Brantley, PhD, School Psychologist, Richland School District One, Columbia, SC. E. Scott Huebner, PhD ( email@example.com), Professor, and Director, School Psychology Program, Department of Psychology, and Richard J. Nagle, PhD, Department of Psychology, University of South Carolina, Columbia, SC 29208