THE VALIDITY OF THE REVISED CHILDREN\u27S MANIFEST ANXIETY SCALE FOR CHILDREN AND ADOLESCENTS

The purpose of this study was to examine the construct validity of the recently revised Revised Children\u27s Manifest Anxiety Scale (Reynolds and Richmond (1978)). The convergent and discriminant validity of the RCMAS for children and adolescents was evaluated using the multitrait-multimethod (MTMM) procedure proposed by Campbell and Fiske (1959). The subjects included in this study were 110 children and adolescents (ages 10 to 16) who had been referred for a psychological evaluation due to either academic or behavioral problems. Ratings of the behavioral/trait domains of anxiety, depression and aggression were obtained from the student, parent and teacher. Convergent validity is confirmed when high and significant correlations are found between various measures of the same trait (i.e. anxiety) obtained by dissimilar methods (Hypothesis One). Discriminant validity is demonstrated when the test/trait (RCMAS) fails to correlate with different traits measured by both similar (Hypothesis Two) and dissimilar methods (Hypothesis Three). In addition, the test/trait (RCMAS) should not correlate with measures of intelligence (Hypothesis Four) or with a response tendency to describe oneself in a positive light (Hypothesis Five). The MTM matrix was evaluated by a traditional analysis of the correlation matrix, path analysis and stepwise multiple regression. The results disconfirmed Hypotheses One and Two as the monotrait-heteromethod (MH) correlations for anxiety were non-significant; MH correlations were lower than the heterotrait-monomethod (HM) correlations all of which were significant. Hypothesis Three was supported as only 28% of the heterotrait-heteromethod (HH) correlations were significant while 66% of the MH correlations were significant. The correlations between the RCMAS and IQ, and the RCMAS and the RCMAS-Lie Scale were non-significant providing support for Hypotheses Four and Five, respectively. The path analysis and stepwise multiple regression showed that the method (rater) contribution to the dependent variables was much larger than the trait contribution. The results cast doubt on the validity of the RCMAS as a measure of trait anxiety across situations. The issue of whether distinctive traits exist was not supported in this study