Empirical Generalisations of Brand Extension Theory and the Role of General Linear Mixed Models

brand extensions;empirical generalisations;general linear mixed models;GLMM;hierarchical linear models;complimentarity;substitutability;transferability;empirical generalizations;close replications;sample size;power;type II error;cross-country study;original brand quality;difficulty;residual centering;ideal point evaluation

Mather, Damien William ; Lawson, Rob ; Holdsworth, David

University of Otago

关键词: brand extensions;    empirical generalisations;    general linear mixed models;    GLMM;    hierarchical linear models;    complimentarity;    substitutability;    transferability;    empirical generalizations;    close replications;    sample size;    power;    type II error;    cross-country study;    original brand quality;    difficulty;    residual centering;    ideal point evaluation;   

Others  :  https://ourarchive.otago.ac.nz/bitstream/10523/622/3/MatherDamienW2011PhD.pdf

美国|英语

来源: Otago University Research Archive

【 摘 要 】

The objective of this research thesis is to resolve the confusion that exists in the marketing literature around brand extension theory, focusing on models that include original brand quality and four brand extension fit attributes realized as complementarity (of the extension with the original branded product), substitutability (of the extension for the original), transferability (of the firm’s resources and capabilities from the original to the extension) and difficulty (perhaps best explained as the opposite of the triviality of the manufacture of the extension by the firm, given the firm’s perceived level of capability).The method used to achieve this objective is to pool the data from one original and eight ;;close replication’ experiments, and re-analyze the data using General Linear Mixed Models. Appropriate random effects specifications to model correlation and covariance amongst the predictor variables, as well as random study location and subject effect specifications are employed. A range of models for the fixed effects listed above will be compared, including models with (a) direct effects, (b) indirect or interaction effects amongst the original brand quality and effects of complementarity, substitutability, transferability and difficulty, and finally (c) second-order ;;ideal point’ model effects on extension evaluation as suggested by certain marketing academics.The implications for brand practitioners are that, contrary to the findings of some of the close replication studies, both direct and indirect effects of quality and fit on brand extension evaluation are supported, with the exception of the difficulty attribute. Natural limits to fixed practitioner brand extension preference strategies are discovered. These limits mean that additional choice modeling or similar is required to correctly explore the implications of proposed brand extensions. The implications for researchers are that seemingly unrelated research streams predicting ;;ideal point’ evaluation responses are shown to be inter-related with brand extension theory, and General Linear Mixed Models are highlighted as a useful inference framework for improving and expanding empirical generalisations in marketing research.