Measurement in Marketing:A Study on Construct Validation with Special Reference to Multirait-Multimethod (MTMM) Matrix

Marketing being a Social Science has to investigate and offer theoretical explanations, backed by empirical evidence, for the behaviour of consumers. This requires development of measuring instruments which can unequivocally assess the abstract nature of constructs relevant to marketing e.g. personality, attitude, consumption innovativeness, materialism etc. Construct validation is a significant step towards formulation of justified measurement instruments for such abstract marketing conceptualisations. The marketing researchers have been operationalising abstract concepts through measurement instruments in the form of questionnaires carrying relevant statements. However, these instruments need to go through a statistically appropriate process to determine their ability to capture the construct in totality. The attempt should be towards determining if the measuring instrument is able to represent the construct, the whole of the construct, and nothing but the construct. The present study provides a detailed discussion on the conceptual underpinnings and ensuing indispensability of construct validity. This is followed by an in-depth review of the marketing research literature on the construct validation process, which involves determination of reliability, convergent validity and discriminant validity of the measuring instrument. The measures of reliability of a scale viz. Cronbach coefficient alpha and item-to-total correlation are discussed in the succeeding sections. The assessment of convergent and discriminant validity through Campbell and Fiske's (1959) Multitrait-multimethod (MTMM) matrix has been profoundly delved upon, followed by an exhaustive evaluation of a modified, but more practical, version of Multitrait-multimethod (MTMM) matrix, as put into application by Ruekert and Churchill (1984).