Premium
Perceived product instrumentality: the process of scale purification and validation
Author(s) -
Triki Abdelfattah,
Wesson David
Publication year - 2002
Publication title -
journal of consumer behaviour
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.811
H-Index - 43
eISSN - 1479-1838
pISSN - 1472-0817
DOI - 10.1002/cb.82
Subject(s) - exploratory factor analysis , psychology , confirmatory factor analysis , scale (ratio) , sample (material) , product (mathematics) , reliability (semiconductor) , social psychology , latent variable , purchasing , set (abstract data type) , construct (python library) , construct validity , structural equation modeling , statistics , psychometrics , developmental psychology , marketing , computer science , mathematics , power (physics) , physics , chemistry , geometry , chromatography , quantum mechanics , business , programming language
Perceived product instrumentality (PPI) is a new construct that is proposed as a key process component of a general model of family purchasing behaviour. PPI reflects the degree to which consumers, apprehended as actors of social roles, deem a product to be helpful, facilitative of role performance, compatible with role identity and congruent with the self‐concept. The objective of this paper is threefold: (1) assess the PPI unidimensionality and reliability; (2) purify the PPI scale, and (3) assess its validity. First, a pilot survey was administered to a convenience sample of men and women, who filled in four identical lists of 33 items tapping their attitudes towards durables, and exploratory factor analysis was conducted on each set to explore the overall pattern of the items relationships. Five try‐out pools of different sizes (33, 28, 15, 13 and 9 items) were involved in the analysis. The 15‐item scale was retained. Secondly, a large‐scale survey was administered to 500 couples as part of an extensive research involving comprehensive model testing. Exploratory factor analysis was conducted on the whole sample for reliability and unidimensionality assessment. At times, the analysis is done on men's and women's sub‐samples separately in order to account for eventual differences among both populations. Thirdly, confirmatory factor was conducted, splitting the sample into two random halves: the generation sample and the validation sample. The first half served for the PPI scales purification. In this case, PPI was posited as the latent variable and the scale items were posited as the manifest ones. The second served to validate the PPI theory in a system's framework: PPI was posited as a latent dependent variable while other role orientations variables were posited as latent independent variables. Copyright © 2002 Henry Stewart Publications.