|
Since 2002, 4 million visitors plus:
| |||||||
|
|
The Scale for Assessing Emotional Disturbance (SAED) (Epstein & Cullinan, 1998) This review was published in the NASP Communiqué and is reproduced here with permission (all rights reserved0 Dumont, R., & Rauch, M. (2000) Test Review: Scale for Assessing Emotional Disturbance (SAED), Communiqué, Vol. 28, 8, 24-25 The Scale for Assessing Emotional Disturbance (SAED) (Epstein & Cullinan, 1998) is a rating scale designed to assist identifying students who may be experiencing emotional and/or behavioral difficulties within the educational setting. It is also reported to be useful as a screening device a tool for research and a method to measure a student's progress. The scale comprises 52 items which encompass Seven subscales and a single item that highlights overall educational performance. These seven subscales include Inability to Learn (e.g., "Homework skills are poor") (8 items), Relationship problems (e.g., “Has few or no friends") (6 items), Inappropriate Behavior (e.g. "Cruel to peers") (10 items), Unhappiness or Depression (e.g., “Lacks self-confidence") (7 items), Physical Symptoms or Fears (e.g., Anxious, worried, tense") (8 items), Socially Maladjusted (e.g., "Runs away from home") ( 6 items), and Overall Competence (7 items). The single item highlighting overall academic performance is termed Adversely Affects Educational Performance. The actual rating form is broken down into three domains that include Student Competence Characteristics, Student Emotional and Behavioral Problems, and Adversely Affect Educational Performance. A psychologist, teacher, parent, caregiver, or any other individual knowledgeable about the individual and his or her behaviors can complete the SAED. Professionals who have appropriate knowledge and background about the SAED specifically and psychometric properties and psychological interpretation in general should complete scoring. Respondents to the SAED are asked to rate the child on each statement using a Likert-type scale. Items within the domain of Student Emotional and Behavioral Problems are scored on a four-point scale, the Student Competence Characteristics on a five-point scale and the Adversely Affects Educational Performance a six-point scale. There are also eight open-ended questions at the conclusion of the scale, which offer ratings related to the student's athletic, academic, social, family and community strengths. Once the SAED scale is completed, the scores for the subscales are summed, and these raw scores can be converted into percentile ranks and scaled scores (M = 10, SD= 3) .The scaled scores from the five subscales are combined to obtain an overall indication of emotional and behavioral functioning, termed the SAED Quotient (M = 100, SD= 15). Norms The SAED was normed on a nationally representative sample of students without emotional disturbance (NonED) and a national sample of students with emotional disturbance (ED). The non-emotionally disturbed sample included 2,266 students ranging in age from 5:0 to 18:11, while the emotionally disturbed sample included 1,371 students ranging in age from 5:0 to 18:11. No explanation whatsoever is given about how these 1,371 children came to be classified as ED. Given the extreme variability in identification of emotional disturbance between and within states, the inclusion of the ED sample without some clear description of how the children were identified in the first place leaves one to wonder about the validity of this sample. Are these children those with specifically diagnosed conditions of emotional disturbance as indicated in DSM-IV or are they the "typical" school-identified emotionally disturbed children? ReliabilityThe SAED was examined for four types of reliability—content sampling, time sampling, rater and scorer. The content sampling of the SAED yielded coefficients that all exceed .75. Two studies were completed to examine whether the results of the SAED were stable over time. Two groups of children identified as Emotionally Disturbed were rated twice over a two-week period. The results of the two studies yielded correlation coefficients that range from .84 to .94. Interrater reliability was examined with the rating of 44 ED students by six pairs of special education teachers. Resulting correlations ranged from .51 to .84. Most reliability coefficients were near or in the .80’s. Two subscales, Physical Symptoms or Fear and Unhappiness or Depression, had the lowest reliabilities, .51 and .61 respectively. Correlation coefficients between scorers (as opposed to raters) for all scales were .99. These results indicate that the SAED is a highly reliable scale. ValidityEpstein and Cullinan examined the content, criterion-related and construct validity of the SAED. Median item discrimination coefficients were calculated for each subscale for both the NonED and ED groups, and the results were reported by age. For the SAED all coefficients were above .39 with a majority in the .6 and .7 range. These results indicate a high degree of content validity To examine concurrent validity, the SAED was compared to the Teacher Report Form (TRF, Achenbach, 1991) as well as to the Revised Problem Behavior Checklist (RPBC, Quay & Peterson, 1996). Hypotheses regarding both the positive and negative correlational relationships between the scales are described in the manual. The correlations that were found were all both statistically and relatively high enough to support each of the predicted relationships made by the authors. The predictive validity of the SAED has not yet been explored. Finally, construct validity for the SAED was examined. Analysis of group differentiation ability using eight t-tests and the Bonferroni procedure to control for multiple comparisons indicates that the SAED did differentiate between the standardization NonED and ED groups effectively with the NonED group obtaining a mean SAED Quotient of 100 compared to 122 for the ED group. Examination of interrelationships between subscale and total scale scores revealed moderate to strong relationships both between individual subscales and the SAED quotient. According to Epstein and Cullinan, the SAED appears to have strong validity in all areas indicating that it is a valid measure of emotional disturbance. Discussion Despite the reliability and validity evidence provided in the SAED manual, these reviewers have a number of concerns regarding the use of the SAED. One of our major concerns regards the conceptualization of the SAED and its relevance to diagnosing children with "Serious emotional disturbance" as defined by the Federal Register. The stated focus of the SAED is on "the five qualifying conditions on which identification of emotional disturbance is based as well as other key features of the federal definition" (p. l ). Unfortunately, from these reviewers' perspective, the SAED avoids the more specific aspect of the federal definition which highlights "the term means a condition exhibiting one or more of the following characteristics..." (Emphasis added). The SAED appears to base a determination of an emotional disturbance on the identification of one or more of the "characteristics" listed in the federal definition rather than on identifying the condition of emotional disturbance from which the characteristics stem. The federal definition of Serious Emotional Disturbance also includes limiting criteria that seem poorly operationalized by the SAED. For example, the federal definition requires that the characteristic evidenced must have been present "over a long period of time." In its description of this specific criterion, the SAED manual notes that during standardization, all raters had known the individual children for at least two months. This raised concern for these reviewers in two ways. First, the period of time for which a characteristic is being rated is based upon the length of contact the racer has with the individual and not on how long the child has exhibited the characteristic. Second, the minimum two-month period is fairly arbitrary and may not conform to individual state or district policies. Another concern relates to the ratings for "adversely effects educational performance." This area is scored using a single rating of a single question. There are no definitions or criteria given on the protocol or in the manual to indicate how one should judge the "adverse effect" or what the terms "slightly," "moderately," or "considerably" mean in relation to scoring this category. Although the SAED reports a large (1,371) sample of children labeled ED, the manual provides absolutely no information about how these children came to be identified as ED, no breakdown of what were the different emotional “conditions" that led to identification and no explanation of the criteria used for inclusion in the emotionally disturbed group. The need for a scale like the SAED is predicated upon the fact that there is such extreme variability in tile diagnosis and classification of ED. To validate a scale on an undefined sample seems problematic. There is also some confusion caused by differing numbers in the SAED tables. Table 4.1 provides the demographic characteristics of the normative NonEd sample. The table reports that, for ages 5 through 11, 12 through 14, and 15 through 18, there were 1,287; 535; and 44 children respectively. In Table 4.2 the number of children in the age ranges is reported as 1,333; 581; and 352 respectively. Beside these inconsistencies, the three age categories used to convert raw scores to standard scores are presented in what seem to be fairly large bands. The manual (p. 15) refers to the age ranges as "elementary," "middle" and "high school" age ranges. There is no empirical justification presented for dividing the normative data in that manner. The SAED does not include Social Maladjustment in any of the norms for children below the age of 12 “because many of the items are about behaviors in which elementary children would not typically engage (e.g., exhibits precocious sexual behavior)" (p. 18). This seems naive and a bit arbitrary. Examination of the SAED items found close similarities between those on the Inappropriate Behaviors scale and those on the Socially Maladjusted scale. For example, on the Inappropriate Behavior scale one finds: “Cheats, Lies, Steals"; "Destroys or ruins things"; and "Uses obscene, profane, or sexually oriented language." In comparison, on the Socially Maladjusted scale, one finds the items: “Steals in the community or at home," "Vandalizes property in the community," and "Exhibits precocious sexual behavior." The only distinction between these items seems be that those described on the "Inappropriate Behaviors" scale occur in the home and in the schools, while those on the "Socially Maladjusted" scale happens solely in the community. These reviewers completed the SAED on eight students, aged 10 to 14, classified as emotionally disturbed by individual school districts and placed in a school specializing in the treatment of children with emotional problems. Of the eight students, only three seemed accurately assessed by the SAED. For only three of these children did the SAED scales indicate characteristics of SED. For the other five students, none of the SAED scales reached a level considered significant, suggesting that they did not have characteristics of SED, although they were already classified as such. This same concern (scores of students identified as having emotional disturbance obtaining average scores on the SAED) is again raised when one examines some of the scores obtained by the various samples used in the SAED validity studies. Table 6.5 provides the subscale and SAED quotients for the NonED and ED standardization groups. While the NonED group has an SAED Quotient of 100.00, the ED group's SAED quotient is, as would be expected, much higher, with a mean of 122.66. However, in Tables 5.6 and 5.7 the various ED groups have mean SAED Quotients of 94.00, 100.32, and 98.27 respectively. Why are the mean quotients for these selected smaller samples of ED children so different from the ED norming sampling? They are well below what would be expected and in fact are either “average" or somewhat "below average" in overall rating. The SAED possesses reportedly strong psychometric properties and might possibly be useful in gaining a deeper and clearer understanding of a child and his or her functioning status. These reviewers, though, do have some skepticism regarding the usefulness of, and need for, such a scale. Although the SAED may be useful at times in differentiating between Emotionally Disturbed and non-Emotionally Disturbed children, it, like many other similar scales, does not provide any information for remediation of identified problems. The SAED is a deficit-oriented scale, but it would be useful for professionals working with children to have information regarding strengths. Such an approach (strength-based) might be useful in writing individual education plans and creating treatment methodology. Examiners must be careful to use SAED and similar scales only for their narrow purposes. As noted above, the SAED does not diagnose the underlying psychological condition that is presumed to cause one or more of the five "characteristics." That diagnosis must still be done by the examiner by additional means. The SAED does not rule out social maladjustment as an alternative cause of the "characteristics." It is up to the examiner and the Team to determine by other means whether a student is emotionally disturbed, socially maladjusted, or both. The "period of time" assessed by the SAED extends back, at most, only to the rater's first contact with the student and, really, only as far back as the length of contact the rater has in mind while completing the form. References Achenbach, T. M. ( 1991 ). Manual for the Teacher Report Form and 1991 profile. Burlington, VT: University of Vermont. Department of Psychiatry. Epstein, M.H., & Cullinan, D. ( 1998). The Scale for Assessing Emotional Disturbance. Austin, TX: PRO-ED. Quay, H. & Peterson (1996) Revised Problem Behavior Checklist Professional Manual. Odessa FL: Psychological Assessment Resources. |