Excerpted from Temperament and Atypical Behavior Scale (TABS) Manual: Early Childhood Indicators of Developmental Dysfunction, by John T. Neisworth, Stephen J. Bagnato, Ed.D., N.C.S.P., Ph.D., John J. Salvia, D.Ed., & Frances M. Hunt, Ph.D.

Copyright © 1999 by Paul H. Brookes Publishing Co. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.

Validity

Validity refers to the "appropriateness, meaningfulness, and usefulness of the specific inferences" one draws from a child's test performance (American Educational Research Association [AERA]), 1985, p. 9). We intend two types of inference to be drawn from TABS ratings: 1) inferences about the presence of temperament and regulatory problems in young children and 2) inferences about the degree of such problems, if they are found. The first type of inference closely parallels what is commonly called content validity, or the extent to which TABS items are proper indicators of temperament and regulatory problems. The second type of inference closely parallels what is commonly called construct validity, or the extent to which TABS scores conform to hypotheses derived from the theory on which TABS was developed.

Does TABS Assess Temperament and Regulatory Problems?

Support for the content validity of TABS comes from two sources. First, the individual items were developed from extensive reviews of the theoretical and descriptive literature on various disorders of infancy and early childhood. The behaviors characteristically associated with these disorders became TABS items. Thus, the items that compose TABS are either characteristic of or highly associated with a variety of serious disorders of infancy and early childhood.

Second, four theoretically meaningful factors underlie TABS. The statistical methods and findings for establishing these factors have been provided in detail in the sections describing the development of TABS. Here we repeat that all four factors have Eigenvalues greater than 1 and that each factor accounts for at least 5% of the rotated variance. Most compelling, however, is that the factor structure of TABS is the same as that theorized independently by ZERO TO THREE: National Center for Infants, Toddlers, and Families in 1994.

Given the thoroughness of item development and item selection, we believe that TABS does assess temperament and regulatory problems in young children.

Do TABS Scores Act as One Would Expect?

Examination of TABS raw scores indicate that the subtest and total scores act as expected. First, given the nature of the behavior sampled, one should expect highly skewed distributions. Most children should earn scores of 0 or 1, with few children earning high scores. Indeed, in the sample of children who have not been identified as having disabilities, the distributions of TABS TRI scores and subtest scores are highly skewed.

Second, we would expect no relationship between TABS scores and chronological age, The behaviors sampled by TABS are not developmental, they are aberrant. We would not expect children to mature out of these behavior problems, although the problems can be treated with intervention. TABS scores are indeed unrelated to age. For that reason, separate norm tables are not provided for children of different ages.

Third, we would not expect a relationship between sex and aberrant behavior in children without disabilities. We found no differences in TABS scores between boys and girls without disabilities; however, we would expect to find sex differences within the population with disabilities because boys are diagnosed as having disabilities with much greater frequency than girls (U.S. Department of Education, 1992). We did find sex differences when the sex of the child was reported. We found that there were more boys than girls identified as having disabilities for every age group except 1-year-olds. Moreover, we found the TABS scores of boys were significantly (p < .005) higher (i.e., more deviant) than those of girls for TRI, Detached, and Hypersensitive/active; boys' scores were higher (but not significantly [p < .05]) than girls' scores for Underreactive or Dysregulated.

Fourth, because TABS behaviors are aberrant, we would expect children with disabilities to earn much higher scores on TABS than children who have not been identified as having disabilities. Indeed, children with disabilities do earn significantly and meaningfully greater scores on TABS than do children without disabilities. No children without disabilities receive TRI scores that equal or exceed the mean of the children with disabilities. On the TABS subtests, 3%-7% of the children not classified as having disabilities earn scores that equal or exceed the mean of children with disabilities.

Fifth, in a study of 42 children, we found that the aberrant behavior assessed by TABS generalizes across contexts. That sample contained 12 girls and 24 boys (sex data were not available for 6 children); 16 of the children were not identified as having disabilities, and 26 children were classified as having disabilities. The mean age of the children was 44.5 months (standard deviation = 14.5 months). The raters observed the children in different contexts. Parents observed their children in all of the usual nonschool environments: home, neighborhood, family gatherings, and so forth. Teachers or clinicians observed the children in a more constrained environments: school, clinic, or assessment settings. The raters each completed a TABS protocol, usually on the same day. TABS and factor scores were moderately correlated. This finding suggests that the behaviors are exhibited across contexts.

A Final Caution About TABS and Validity

The validity of an instrument and the inferences that can be based on an instrument are never established. Rather, one accepts the validity of an instrument when it has not been shown to be invalid. We have assessed TABS in what we believe to be the major areas that would indicate a lack of validity. We found nothing to suggest that TABS is not valid. Future research can only add to our knowledge. We urge TABS users to send us the results of any research they might conduct with this scale.

Validity of the TABS Screener

The purpose of the TABS Screener is to identify children who are likely to be at risk or to have disabilities and who, therefore, should be assessed with the full TABS Assessment Tool. Thus, the only validity in which we are interested is the accuracy of the Screener in identifying children who do and children who do not require further evaluation, which is illustrated in Figure 5.1. When a child is predicted by the TABS Screener not to be at risk and is identified as being at risk (or to have a disability) by the Assessment Tool as not at risk, the prediction is correct. Similarly, when a child is predicted to be at risk (or having a disability) by the Assessment Tool, the prediction is accurate. However, when a child is predicted not be at risk by the Screener and is identified by the Assessment Tool as being at risk (or as having a disability), the prediction is incorrect; that child is called a false negative. For screening devices, false negatives are the most serious error because those children do not receive follow-up evaluation. If a child is predicted to be at risk (or to have a (or to have a disability) but is actually not at risk, the prediction is again inaccurate; this child is called a false positive. For screening devices, false positives decrease efficiency because children who do not really require a follow-up evaluation (because they are not at risk or do not have a disability) receive one.

Professional development on this title is available through Brookes On Location!