Premium
Moderator Effects on Attitude Scale Construction 1
Author(s) -
Warren Richard D.,
Mulford Charles L.,
Winkelpleck Judy M.
Publication year - 1973
Publication title -
family and consumer sciences research journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.372
H-Index - 31
eISSN - 1552-3934
pISSN - 1077-727X
DOI - 10.1177/1077727x7300100404
Subject(s) - moderation , scale (ratio) , psychology , reliability (semiconductor) , statistics , construct (python library) , cluster (spacecraft) , level of measurement , residence , econometrics , social psychology , mathematics , computer science , geography , demography , cartography , power (physics) , physics , quantum mechanics , sociology , programming language
Data from a study on smoking and health are used to examine three methodological questions: (1) Does the method of scale construction influence the items in the resultant scales? (2) Do alternative subgroup delineations influence the formulation of scales? (3) If different scales are formed, what is the influence on measurement criteria, such as reliability, inter‐item and item‐total correlations, and statistical testing? Attitude scales were formed for subgroups based on smoking behavior, sex, and area of residence. Cluster analysis and theory were used to formulate the scales. The data provide empirical support that: (1) the method used to construct scales and (2) the moderators or subgroups do influence items in the scales, measurement criteria, and statistical testing.