Consumer Preference Analysis: Three Brands of Vodka
Home - Table of Contents - Executive Summary - Introduction
Methodology - Analysis - Conclusions - Summary - Appendix A - Appendix B
Conclusions
Conclusions of the different Consumer Preference Data Analyses:
The means, Brand Index Score, and Ad Index Scores presented in the analyses were calculated using constant sum brand preference fields to analyze the relationship between pre/post ad exposure, ten Likert items for each brand, attribute checkboxes, and agree/disagree questions for each of the three vodka brands: Finlandia, Ketel One, and Stolichnaya. The results of the tests suggest that the respondents overall slightly preferred Ketel One to Finlandia, but the other Brand Index Score comparisons between two of these three brands were too small to be statistically significant. The data collected reflected that most respondents viewed the Finlandia ad positively, causing their opinion of the product to increase.
A linear regression analysis was performed for three vodka brands to determine if there was a correlation between how the 76 respondents rated the brand based on the 10 brand attribute questions (Likert items), and the change in score from the pre-to-post print advertisement exposure. For Finlandia, the attributes ‘Good’, and ‘Something special’ are significant independent variables that explain the variance in pre-to-post ad exposure change scores. For Ketel One, it was the attributes 'Smooth' and 'Something special'; and for Stolichnaya, ‘Good’, ‘For me,' and ‘Good taste’ were statistically significant.
This discriminant analysis is based on one of the three brands of vodka in the survey, Finlandia. The purpose of the analysis was to see how well respondents could be divided into groups of up- or down-movers based on their responses to Likert questions that composed the Brand Index Score. The classification matrix indicated that only 64.5% of the grouped cases (Up-movers and Down/Non-movers) were correctly predicted by the discriminant analysis. This percentage was quite low, suggesting a lot of unpredictability in classification.
The purpose of this one-way analysis of variance (ANOVA) was to compare the means of two factors (change scores for up movers, non movers, and down movers in the pre-to-post print advertisement exposure, and the respondents' age) with one dependent variable(the attribute "Finlandia is a good vodka"). In these descriptive statistics, the mean value attributes did not vary dramatically. The mean scores of the “Finlandia is a good vodka’ attribute of the up movers, non movers, and down movers were all greater for the over 25 age group. These mean scores were all very similar, each within the standard deviation of one another.
A factor analysis was performed for each brand of vodka to place related brand attributes into independent groups called factors. The analysis revealed those Likert items that were evaluative in nature and related to the most ambiguous Likert item: "BRAND X is a good vodka". The results of this test could be projected to the population only in comparing the attitude score between Ketel One and Stolichnaya. Comparisons between any other two attitudes could not statistically be expected to project to the population.
A cluster analysis was run on the gathered data to determine if respondents could be grouped into clusters based on their responses to the ten brand attitude Likert items for Finlandia. The clusters were then analyzed to determine if the clusters were differentiated on the basis of level of education. If these percentages differed greatly from the 48.7% and 51.3% distribution of the education level samples, we would expect that cluster position is affected by the level of education. However, with all of the resulting statistics coming in between 40% and 60%, it suggests that in this survey, the education level of the respondents didn’t mean much regarding which cluster they fell into.