Multitrait-Multimethod Analysis

Multitrait-multimethod analysis is a methodology used to assess the construct validity of an instrument. The technique requires that subjects be measured on at least two variables with at least two methods of instrumentation; for example, reading and mathematics each measured with multiple-choice items and open-response items. A trait correlation is the correlation coefficient between the same variable, measured with two different methods (e.g., reading multiple-choice scores correlated with reading open-response scores). A method correlation is the correlation coefficient between the two different variables, measured with the same method (e.g., reading multiple-choice scores correlated with mathematics multiple-choice scores). A relatively high trait correlation, coupled with a relatively low method correlation, is evidence of the construct validity of a variable.