Date of Award

Spring 1996

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Program/Concentration

Business Administration-Marketing

Committee Director

C. P. Rao

Committee Member

John B. Ford

Committee Member

Wilbur W. Stanton

Abstract

The objective of this research was to understand the effect of culture on Extreme Response Style (ERS) for cross-cultural marketing efforts employing Likert type scales. Growth in quantitative cross-cultural marketing research has brought alternative methodological considerations, like ERS, to the forefront of attention, since, traditional inferential statistics used in group comparisons become confounded by this systematic bias. Marketers comparing groups with differing degrees of response style may incorrectly make claims based on observed differences that are strictly attributable to cultural ERS. Therefore, the specific objective of this dissertation was to relate ERS to culture, Likert response formats and to demonstrate the statistical adjustment required to remove the inherent bias ERS imposes on marketing scales. Finally, the study investigated a culturally based explication of this response style.

Data were collected via a translated questionnaire distributed to students in Australia, France, Mexico and the U.S. Hypotheses were investigated through ANOVA and correlational analyses. Response style adjustment was demonstrated using a post hoc attitude/behavior adjustment. Reliabilities of scales are reported through Cronbach's Alpha and split-halves statistics.

The hypothesis that culturally distinct countries will have different levels of ERS was supported. Support was found for the hypothesized relationship between ERS and cultural value. ERS was found to be related to the agreement response style, but, not to differ with social desirability. Increasing the number of response formats was found to decrease ERS within cultural groups, but, not to reduce between group differences beyond the 7-point format. Inconsequential ERS difference was found between odd and even numbered scales. Higher numbers of response intervals did not result in a convergence of between culture ERS difference. ERS was found to bias the outcomes of the CETSCALE in a cross-cultural setting. When post hoc adjustments were invoked, the statistical outcomes varied from the unadjusted set.

Recommendations for cross-cultural research include: (1) the continued reporting of response styles, (2) the use of 5, 6 or 7-point scales to minimize between group differences, (3) removing the ERS bias through a post hoc procedure prior to making across group comparisons. Discussion of study limitations and possible future research are included.

DOI

10.25777/9gv1-1z98

Share

COinS