Construct Validation of Anxiety Measures Using Multitrait-Multimethod Matrix

This study sought to establish the construct validity for an instrument for measuring anxiety. The researchers used a four-point questionnaire and a seven-point Osgood semantic differential scale on depression to ascertain the convergent validity while two instruments measuring aggression were employed to establish the divergent trait with anxiety using multitrait-multimethod matrix. The face validity was carried out by experts in Educational Measurement and Evaluation. Cronbach Alpha reliability estimates for internal consistency of the items yielded 0.76, 0.98 for anxiety measures; 0.74, 0.85 for depression measures and 0.63, 0.79 for aggression measures respectively. The PPMC coefficient was used to test the hypotheses. Samples of thirty Senior Secondary III students of University of Nigeria Demonstration Secondary School were purposively selected for the study. The results demonstrated moderate convergence (r = 0.20, 0.49, 0.39 for measures of anxiety, depression and aggression respectively) between two different methods of the same trait. Measures assessing anxiety and depression could be distinguished from measures assessing aggression. In conclusion the rejection of the first hypothesis and the retention of the second and third hypotheses based on the correlation confirm the convergent and divergent validities of the instruments; therefore, the instruments for measuring anxiety were deemed valid and reliable.