CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

• Executive Summary

• Conclusions

• Introduction

• Summary

• Methodology

• Appendix A

• Analysis

• Appendix B

CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

The purpose of this research was to use copy testing as a means to examine brand preferences among three well-known brands of bottled water: Evian, Fiji, and Dasani. This study determined the initial brand preference, analyzed common attributes associated with each brand, and then sought to find the effect of print advertisement on brand preference. Respondents for the survey were solicited through email, and asked to fill out an online form. Data were collected via an online questionnaire completed by 62 respondents and later imported into SPSS for a variety of statistical analyses. Four different analyses were conducted in SPSS: basic statistics (independent samples and paired t-tests, frequencies, and correlations), multiple regression analysis, discriminant analysis, and factor analysis. The tests revealed a more positive brand attitude towards Evian over both Fiji and Dasani, when Evian's advertisement also had the highest favorability with the least decrease after the print ads were shown. Fiji had almost same respondents with improved and lowered brand perception after exposure to the ad, while Dasani had the least respondents with improved brand perception and the most respondents with lowered brand perception after exposure to the ad. The intial brand index score (pre-ad exposure) was in the following order: Dasani, Evian, and Fiji; however, these findings indicate that respondents ranked the brands in the following order: Evian, Fiji, and Dasani, in terms of both brand index score (post-ad exposure) and attitude score after viewing the ads. A more detailed report of the results can be found in the Statistical Analysis section.

CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

This study used copy testing to understand the effectiveness of print advertisements for three brands of bottled water. Copy testing of advertising research is a general class of tests that evaluate and diagnose the communication power of an advertisement such as the effects of advertisement on consumer perception/attitudes.

The purpose of this research is to examine brand preferences among three brands and the effect of the brands' print ads. For this study, three well-known, comparable brands (i.e., Evian, Fiji, and Dasani) of a certain product category in convenience goods (i.e., bottled water) were selected because they are low cost, and purchased with low consumer involvement. Consumer attitudes towards the brands were measured both pre and post ad exposure. Statistical significance was set at p < .15 in order to determine if in 85 or more samples drawn from the same population as this sample, the expected values would be the same as founded in this sample.

The survey (Appendix B) was taken online after friends, family and colleagues were sent a hyperlink via email. The 62 results were downloaded to a database and then analyzed using the statistical software SPSS. Average values for all of the measured variables are added to the survey in Appendix A.

This final report offer the results from this study and includes Executive Summary, Introduction, Terminology, Statistical Analysis, Methodology, Conclusions, Summary, and Appendix A and B (survey with the frequency output for each question and the original survey as viewed by respondents).

CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

• Structure of the Questionnaire

• Sampling

• Design

• Online Data Collection

CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

The questionnaire included the following sections: Introduction, Pre-ad Exposure, Advertisement Exposure, Post-ad Exposure, Brand Attitudes, Advertisement Attitudes, and Demographics. A description of each of these sections follows.

Introduction

The questionnaire begins with a brief message telling respondents about the educational purpose of the research and the strict confidentiality of their responses.

Pre-ad Exposure (Sections 1 and 2)

Respondents were asked to report their most recent purchase of a soft drink, a sports drink, and bottled water. Following that, the likelihood of purchasing each of three named brands of soft drinks, sports drinks and bottled water was reported. These constant sum values were reported by respondents using a 10 point total allocation among the three choices within each category. Only the bottled water scores were used for analysis and were collected immediately prior to and after viewing the ads. The difference between the initial score and the post-ad exposure score is referred to throughout as the "Change Score."

Advertisement Exposure (Section 3)

In this section there are three full-color, single page magazine advertisements, one for each brand of bottled water. Respondents were asked to read the ads and then to complete the remainder of the survey without looking back at the ads. Three print ads were selected from some lifestyle magazines. All three ads are in the same layout and in color. Below are the ads appeared to respondents:

Post-ad Exposure (Sections 4 and 5)

The same constant sum scale used to measure likelihood to buy before the advertising exposure was used for the same purpose after respondents had viewed the ads. Respondents were also asked several "yes" or "no" questions about their attitudes towards the brands after seeing the ads.

Brand Attitudes (Section 6)

This section asked ten Likert item questions about each brand. These questions asked respondents to indicate how strongly they agreed or disagreed with the statements by selecting "strongly agree," "agree," "neither agree nor disagree," "disagree," or "strongly disagree." These questions measured respondent attitudes by linking either a positive or negative statement with the 1-5 Likert scale. Agreement with the statement is indicated by a higher score than disagreement.

Advertisement Attitudes (Sections 7 and 8)

The next series of questions measured response to the advertisements in two formats. Section 7 allowed respondents to "check" a box that indicated agreement with a single word descriptor identified as "describing" the advertisement. Section 8 was a series of 10 statements to mark "yes" or "no" signifying either agreement or disagreement with the statement. The Ad Index Score was created from the summation of the positively worded phrases or descriptors that were identified by respondents as applying to the advertisement.

Demographics (Section 9)

The final section of the survey asked respondents a variety of demographic questions including gender, age, education, household income, and some lifestyle related questiones regarding bottled water consumption.

CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

This study was designed to measure brand attitudes and responsiveness to advertisements using a copy testing format with pre and post ad exposure data collection. Respondents were asked to score brands prior to and immediately after viewing an advertisement. In addition, other attitudinal responses were measured after viewing the ad. Due to time constraints, there was no control group against which to compare responses.

CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

A convenience sample was used for this survey due to time and resource constraints. The respondents came from friends and colleagues. A minimum of sixty completed surveys were required in order to have a decent sample size for analysis. This study uses the results from sixty-two completed surveys for analysis.

CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

The data in this study were collected via the questionnaire contained in Appendix B. Respondents accessed the survey through a link in an e-mail sent to friends and colleagues. The results submitted by each respondent were linked through a Cold Fusion program and integrated into a master database file. This information was retrieved through Microsoft Access and all analyses were conducted using SPSS and simple mathematical formulas.

CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

• Basic Statistics Analysis

• Discriminant Analysis

• Regression Analysis

• Factor Analysis

CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

How respondents view the advertisement of three brands is assumed to be influenced by their attitudes toward the brands and their demographics. Therefore, the brand index score, which represented respondents' attitude toward each brand, was the first to be examined. Next, the relations between each brand's brand index score and pre-to-post ad exposure constant sum score, which represented how the respondents view each advertisement, was examined as well.

1. Paired t-tests for Brand Index Score

Table 1. Brand Index Score

Table 2. Paired t-tests

*p < .15

Overall, Evian had the highest mean for brand index score, with the smallest standard deviation. Fiji had the second highest mean and Dasani had the lowest mean for brand index score.
The result of the correlated t-test shows that the difference between means of brand index scores for the test on Fiji and Evian is statistically significant at an alpha level of .15. In 85 or more samples out of every 100 samples drawn from the same population as this sample of 62 people, it is expected that the means of brand index scores for Fiji and Evian would be about what they are in this sample. Therefore, we can project the results for this sample to the population; the same as for the tests on Evian and Dasani. However, the difference between means of brand index scores for Dasani and Fiji would not have a significantly different magnitude as found in this sample, so that the results for this sample cannot be projected to the population.

2. Between Groups t-tests: Fiji

Table 3. Brand Index Score Change for Fiji

Table 4. Ad Index Score Change for Fiji

*p < .15, Total Sample: 43

There were 22 individuals who changed scores that moved up after pre-to-post exposure to the ad on Fiji. The mean of brand index score for this group was 34.4 and the mean of ad index score was 7.2. On the other hand, there were 21 individuals who changed scores that moved down pre-to-post exposure to the ad on Fiji. The mean of brand index score for this group was 34.6 and the mean of ad index score was 5.1.
Respondents who moved up in pre-to-post exposure to the ad liked the Fiji ad more than respondents who moved down, but they liked the Fiji brand less than respondents who moved down. At an alpha level of .15, the ad index score for Fiji was statistically significant, but the brand index score for Fiji was not. In 85 or more samples out of every 100 samples drawn from the same population as this sample, it is expected that the mean scores for Fiji on its ad index score would be about what they are in this sample. Thus, the results for this sample can be projected to the population. However, this does not hold true for the mean scores for Fiji on its brand index score. The results for this sample can not be projected to the population.
Overall, respondents who improved their perception of Fiji after viewing the ad had more favorable perception of the Fiji ad, not the brand, than those who felt negative about Fiji after viewing the ad.

3. Chi-Squared Significance Test: Fiji

Table 5. Brand Index Score Relative to Median for Fiji

Chi-Squared=1.36, Total Sample: 57

It shows how many people who increased, stayed the same, and decreased in how likely they liked the Fiji brand as compared before they viewed the ad and were also either above or below the median brand index score for Fiji which was 33. 36.8% (7 respondents) of those who moved up, 50.0% (9 respondents) of those who remained the same and 55.0% (11 respondents) of those who moved down after pre-post testing for Fiji were above the median. 63.2% (12 respondents) of those who moved up, 50.0% (9 respondents) of those who remained the same and 45.0% (9 respondents) of those who moved down after pre-post testing for Fiji were below the median. It shows that the proportion of each box is nearly identical, so that each association is weak. Moreover, with a chi-squared value of 1.36, it is not statistically significant at an alpha level of .15, so that it is not expected to see the same distribution when 85 or more samples out of every 100 samples are drawn from the same population as this sample. Therefore, we can not project the results for this sample to the population.

4. Pre-Post Ad Exposure Change Score Frequencies

Table 6. Frequency Count of Pre-Post Exposure Change Score by Brand

Fiji had the most respondents (22 respondents) with improved brand perception and almost same respondents (21 respondents) with lowered brand perception after exposure to the ad. Dasani had the least respondents (13 respondents) with improved brand perception and the most respondents (30 respondents) with lowered brand perception after exposure to the ad. Evian was the second highest moving up (21 respondents) and the lowest mowing down (12 respondents) in favorability in pre-to-post exposure. Overall, these results indicate that respondents favor Fiji's ad the most, Evian's ad the second, and Dasani's ad the least.

5. Brand Index Score Comparison: Fiji and Evian

Table 7. Comparison by Brand Index Score for Fiji and Evian

Out of the 62 respondents, 27 individuals (43.5% of the sample) rated Fiji higher than Evian, and 35 individuals (56.5% of the sample) rated Fiji lower than or equal to Evian after viewing the ad. The result shows that more respondents felt more favorably toward Evian than toward Fiji after viewing the ad.

6. Simple Correlation Coefficient: Fiji and Evian

Table 8. Correlation Coefficient for Fiji and Evian

*p < .15

The simple correlation coefficient between brand index scores for Fiji and Evian is .3. Although there is a moderately weak positive linear association between brand index scores for Fiji and Evian, this is statistically significant at an alpha level of .15. In 85 or more samples out of every 100 samples drawn from the same population as this sample of 62 people, it is expected that the correlation between brand preferences for Fiji and Evian would be about the same as found in this sample. Therefore, we can project the results of the survey to the population. 

7. Simple Correlation Coefficient: Fiji and Evian for Women

Table 9. Correlation Coefficient for Fiji and Evian (Women)

*p < .15

A simple correlation coefficient test was conducted between brand index scores for Fiji and Evian to determine if any relationship exists among women. 35 people out of the total sample of 62 people were used in this analysis. There is a moderately weak positive association between brand index scores for Fiji and Evian for female sample with a correlation coefficient of .4. The correlation coefficient value for the female sample is larger than one for the total sample, so that it can be inferred that brand index scores for Fiji and Evian is more strongly associated among women than men. Also, this result is statistically significant at an alpha level of .15. In 85 or more samples out of every 100 samples drawn from the same population as this sample, it is expected that the correlation between brand index scores for Fiji and Evian would be about what they are in this sample. Therefore, we can project the results for this sample to the population.

CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

To find out the relationship between each brand attribute and the pre-to-post ad exposure change score, multiple regression tests were conducted for each brand with the brand attributes as the independent variables and the pre-to-post ad exposure change score as the dependent variable. 62 people participated in this on-line survey.

1. Multiple Regression Analysis for Fiji

Table 1. Correlation between Change Score and Brand Index Items

*p < .15

Table 2: Unstandardized and Standardized Coefficients for Fiji

*p < .15

Because the R Squared is low, the Brand Index Score is not very highly associated with the Change Score. This low and positive correlation means that only 24.9% of how much the respondents liked the Fiji ad (as defined by a positive Change Score) is explained by how much they liked the Fiji brand. The F value (1.69) is statistically significant so that in 85 or more samples out of every 100 drawn from the same population as this sample, it would be expected that the coefficient of multiple determination would be about the same as it is in the sample.
The items of "Thirst Quencher," and "Pure" are significant so that in 85 or more samples drawn from the same population as this sample that the correlation would be about what it is in this sample for these two items. The items that are important variables in explaining variance in the Change Score include those described as "Thirst Quencher," and "Pure."
The positive items, such as "Good,""Prefer," "Healthy," "Thirst Quencher," and "High Social Status" mean that as these item values increase, so does Change Score. The negative items, such as, "Do Not Like Taste," "Too Expensive," "Pure," "Trust," and "Not Drink" mean that these are negatively correlated with the Change Score, that is, as the values for these items increase, the Change Score decreases. The standard error of estimate is at 1.9, indicating that the respondents were an average of 1.9 points away from the regression line.

The multiple regression equation:
Pre-Post Ad Exposure Change Score for Fiji= -1.9 + .4 (Good) + .4 (Prefer) - .4 (Do Not Like Taste) - .1 (Too Expensive) + .2 (Healthy) + 1.4 (Thirst Quencher) - 1.0 (Pure) - .5 (Trust) -.2 (Not Drink) + .5 (High Social Status)

2. Multiple Regression Analysis for Evian

Table 1. Correlation between Change Score and Brand Index Items

Table 2: Unstandardized and Standardized Coefficients for Evian

*p < .15

Because the R Squared is low, the Brand Index Score is not very highly associated with the Change Score. This low and positive correlation means that only 20.9% of how much the respondents liked the Evian ad (as defined by a positive Change Score) is explained by how much they liked the Evian brand. The F value (1.34) is not statistically significant so that these correlations that apply to this sample cannot be generalized to other populations.
The items of "Prefer," and "Trust" are significant so that in 85 or more samples drawn from the same population as this sample that the correlation would be about what it is in this sample for these two items. The items that are important variables in explaining variance in the Change Score include those described as "Prefer" and "Trust."
The positive items, such as "Good," "Prefer," "Do Not Like Taste," "Thirst Quencher," "Pure," and "Not Drink" mean that as these item values increase, so does Change Score. The negative items, such as, "Too Expensive," "Healthy," "Pure," and "Trust" mean that these are negatively correlated with the Change Score, that is, as the values for these items increase, the Change Score decreases. The standard error of estimate is at 1.6, indicating that the respondents were an average of 1.6 points away from the regression line.

3. Multiple Regression Analysis for Dasani

Table 1. Correlation between Change Score and Brand Index Items

Table 2: Unstandardized and Standardized Coefficients for Dasani

*p < .15

Because the R Squared is so low, the Brand Index Score is not highly associated with the Change Score. This low and positive correlation means that only 12.8% of how much the respondents liked the Dasani ad (as defined by a positive Change Score) is explained by how much they liked the Dasani brand. The F value (.75) is not statistically significant so that these correlations that apply to this sample cannot be generalized to other populations.
The items of "Pure" is significant so that in 85 or more samples drawn from the same population as this sample that the correlation would be about what it is in this sample for this item. The items that are important variables in explaining variance in the Change Score include "Healthy," "Thirst Quencher," "Pure," and "High Social Status."
The positive items, such as "Good," "Prefer," "Healthy," "Trust," and "Not Drink" mean that as these item values increase, so does Change Score. The negative items, such as, "Do Not Like Taste," "Too Expensive," "Thirst Quencher," "Pure," and "High Social Status" mean that these are negatively correlated with the Change Score, that is, as the values for these items increase, the Change Score decreases. The standard error of estimate is at 2.0, indicating that the respondents were an average of 2.0 points away from the regression line.

CONSUMER PREFERENCE ANALYSIS

THREE BRANDS OF BOTTLED WATER

In order to know if the ten Likert items can effectively predict our respondents' attitude toward the advertisement with a better than chance accuracy, discriminant analysis for Fiji was conducted. This discriminant analysis was executed to test whether people like the advertisement can be based on the ten Likert items of the brand with two groups, 'up-movers,' those who liked the advertisement for Fiji and 'down-movers,' those who did not like the advertisement for Fiji. Those respondents who indicated no change in their pre-to-post ad exposure brand evaluations were not included in this analysis. The independent variables in this analysis are the brand attributes as indicated by ten Likert items and the categorical dependent variable is group membership (up's and down's). Forty three people out of the total sample of sixty two were used in this analysis.

Table 1. Group Mean Scores and Standard Deviation

When comparing the mean scores of each brand attribute, there are little differences between those who had positive pre-to-post ad exposure change scores and those who had negative pre-to-post ad exposure change scores. Although the largest difference between mean scores for up-movers and down-movers was 0.6, the difference was not great. The most marked appeared in the attributes "Too expensive," "Prefer," "Good," and "Pure." Since there are no items that have dramatically different mean scores between up-movers and down-movers, there is not likely to be much to discriminate.

Table 2. Standardized and Unstandardized Discriminant Function Coefficients

The item "Pure" has the biggest absolute value of standardized discriminant function coefficients, which is 1.3. The important independent variables in explaining differences in pre-to-post ad exposure change scores between up-movers and down-movers are brand attributes described as "Do Not Like Taste," "Thirst Quencher," and "Pure."  The more people think Fiji is a brand of which taste they do not like, the more they like the advertisement; the more people think Fiji is a brand that quenches their thirst, the less they like the advertisement; the more people think Fiji is a brand that tastes pure, the more people like the advertisement.

Table 3. Group Centroids and Statistical Significance

*p < .15, Degrees of Freedom=10

The group centroids, which are the average discriminant z scores of -.9 for up-movers and of .9 for down-movers, are significant, due to the statistical significance found for Wilks' Lambda at .54 and chi-squared at 22.21. Therefore, in 85 or more samples out of every 100 samples drawn from the same population as this sample of 43 people, it is expected that the group centroids would be about what it is in this sample. Therefore, we can project the results of the survey to the population.

Table 4. Classification Matrix