Analysis
An online survey was conducted to test the change in perceptions of respondents after viewing various print ads about different brands of chewing gum. The following analysis describes selected results of this survey, utilizing various statistical comparison tests (i.e. t-Tests, chi-square tests, Pearson correlation). The brands of interest include Extra, Eclipse and Orbit. Many of the comparisons are based upon Brand Index Scores and Advertising Index Scores which are described below.
1. Correlated t-test Comparison
Table 1.1 Brand Index Score |
||
Brand |
Mean |
Std.Dev. |
Eclipse |
27.0 |
4.7 |
Extra |
25.7 |
6.2 |
Orbit |
29.9 |
6.1 |
N = 67 |
||
Table 1.2 Paired Correlated t-Tests using Brand Index Score |
|
Brands |
t |
Eclipse & Extra |
1.73* |
Extra & Orbit |
-4.35* |
Orbit & Eclipse |
3.46* |
*p ≤ 0.15 |
|
As seen from Table 1.1, 1 Brand Index Score was calculated for each brand based on the answers of 67 respondents to a series of 10 Lickert questions about the brand. According to the collected data, in 85 or more samples out of every 100 samples drawn from the same population of this sample, the paired sample tests would be about what they are in this sample. In other words, because all t-test results were significant, an inference can be made that the positive perception toward the Orbit brand is higher than the Eclipse brand, which in turn, is higher than Extra brand.
2. Independent t-Tests
Table 2.1 Brand Index Score Change for Extra |
||||
Change in Perception after Ad Exposure |
N |
Mean |
Standard Deviation |
t-Ratio |
More favorable |
22 |
25.5 |
6.4 |
-1.79* |
Less favorable |
20 |
28.5 |
3.9 |
|
*p ≤ 0.15 |
||||
Table 2.2 Advertising Index Score Change for Extra |
||||
Change in Perception after Ad Exposure |
N |
Mean |
Standard Deviation |
t-Ratio |
More favorable |
22 |
6.8 |
3.4 |
2.57* |
Less favorable |
20 |
4.4 |
2.8 |
|
*p ≤ 0.15 |
||||
For these tests, an Advertising Index Score and the Brand Index Score were used to determine the frequency of the change in perceptions toward the Extra brand after viewing the Extra ad. The Advertising Index Score was calculated by taking the sum of a maximum of 14 positive descriptors selected by the respondent about the Extra print advertisement. Because both Brand Index Score and Advertising Index Score comparisons were found to be significant, all the results shown can be projected to the population from which the sample was drawn.
These tests show that about a third of the respondents reported a more favorable view toward the brand after viewing the ad and another third of the respondents reported a less favorable view toward the brand after viewing the ad. Interestingly, the overall brand index score mean (i.e. consumer perception) was lower for the “more favorable” group than the “less favorable” group; however, the “more favorable” group appeared to like the Extra print advertisement slightly more than the “less favorable” group.
3. Chi-Square Significance Tests
Table 3.1 Chi-Square Significance Tests |
||||
BIS: Extra |
||||
Above Median |
Below Median |
|||
Change in Perception Immediately After Ad Exposure: Extra |
More favorable |
Count |
9 |
12 |
% within Change in Perception group |
42.9% |
57.1% |
||
% within Brand Index Score Median group |
28.1% |
44.4% |
||
% of Total |
15.3% |
20.3% |
||
Remained the same |
Count |
9 |
11 |
|
% within Change in Perception group |
45% |
55.0% |
||
% within Brand Index Score Median group |
28.1% |
40.7% |
||
% of Total |
15.3% |
18.6% |
||
Less favorable |
Count |
14 |
4 |
|
% within Change in Perception group |
77.8% |
22.2% |
||
% within Brand Index Score Median group |
43.8% |
14.8% |
||
% of Total |
23.7% |
6.8% |
||
Pearson Chi-Square = 5.80* |
All the results shown are statistically significant, and therefore, the expected immediate changes in perceptions compared with the Brand Index Score of the Extra brand can be projected to the population from which this sample was drawn.
Among the people who had an above-median brand index score of the Extra brand, these respondents were pretty evenly dispersed among the three different categories of change in perception. However, those respondents who were part of the below-median brand index score were more likely to report a same or more favorable perception toward the Extra brand.
4. Pre-Post Stimulation Results
Table 4.1 Changes in Perception after Exposure to Ad |
||
Brand |
Change in Perception |
N |
Eclipse |
More favorable |
26 |
Remained the same |
29 |
|
Less favorable |
12 |
|
Extra |
More favorable |
22 |
Remained the same |
25 |
|
Less favorable |
20 |
|
Orbit |
More favorable |
12 |
Remained the same |
23 |
|
Less favorable |
32 |
|
Out of this sample of 67 respondents, the perceptions of respondents became more favorable, remained the same, or became less favorable toward each brand after he or she was exposed to the ads.
Within each brand category, the perceptions toward the brand of the majority of respondents remained the same, except for that of the Orbit brand. For this particular sample, the Orbit ad seemed to negatively influence the most respondents while the Eclipse ad seemed to positively influence the most respondents.
5. Comparison of Brands using Brand Index Score
Table 5.1 Comparison of Brand Index Scores |
|
|
N |
Eclipse > Orbit |
21 |
This sample revealed that the Eclipse Brand Index Score was higher than the Orbit Brand Index Score for 21 respondents. This means that more people responded more favorably toward the Eclipse brand than the Orbit brand.
6. Correlation Coefficient between Two Brands
Table 6.1 Comparison of Brand Index Scores |
|
|
Pearson Correlation |
Extra, Orbit |
0.17 |
A comparison of the correlation coefficients for Extra and Orbit shows that a low to moderate positive correlation exists between the brands. This means that growing positive attitudes toward the Extra brand are minimally to moderately related to the growing positive perceptions toward the Orbit brand.
7. Subset of Respondents: Computing Correlation Coefficient Between Eclipse and Orbit gum among people who self-reported that they chew gum daily
Table 7.1 Comparison of Brand Index Scores within Daily Gum Chewers group: Eclipse and Orbit |
|
Pearson Correlation (r) |
0.60 |
N+ |
19 |
|
|
+N is based on the number of people who chew gum daily |
|
From this sample, 19 out of the 67 respondents chew gum on a daily basis. Based on this group, a moderate to high positive correlation exists (r = 0.60) between the brand index scores of Eclipse and Orbit chewing gum. This means that among daily gum chewers, while their perception of the Eclipse brand becomes more positive, their perception of the Orbit brand also becomes more positive.
A linear regression test was performed on the data of three chewing gum brands (Eclipse, Extra, and Orbit) to determine the association between the dependent variable (“change score” of each respondent) and the independent variable (each respondent’s overall attitude toward the brand, using the responses from 10 Likert questions). The change score describes the change in attitude toward the brand immediately after viewing the print ad, and therefore, was calculated by subtracting the pre-score from the post-score of each of the 67 respondents.
Table 1.1 Linear Regression Model Summary for EXTRA |
|||
Multiple correlation coefficient (R) |
Coefficient of multiple determination (R2) |
Standard Error of Estimate (Se) |
F-ratio to test multiple correlation coefficient |
0.6 |
40.0% |
2.1 |
3.66* |
*p ≤ 0.15 |
|||
In 85 or more samples drawn from 100 of the same population as this same sample it would be expected that the multiple correlation coefficient would be about what it is in this sample. The multiple correlation coefficient, or R, is .6 and that the coefficient of multiple determination, or R2, is at 40%. This means that only 40% of the variance of the change score for Extra is explained by the 10 Likert items for Extra.
In other words, a moderate to high connection exists between how respondents view the Extra brand and how they view the Extra brand immediately after exposure to the print ad. The coefficient of multiple determination at 40% means that the quality of this assessment is low to moderate.
Table 1.2 Linear Regression Coefficients for EXTRA |
||||
Variable |
Unstandardized Coefficient (b) |
Standardized Coefficient (Beta) |
t-ratio |
Importance |
(Constant) |
2.9 |
|
2.39* |
|
Is a good chewing gum |
0.4 |
0.2 |
1.20 |
X |
Feels refreshed after chewing |
-0.7 |
-0.3 |
1.8* |
X |
Would not buy |
-0.8 |
-0.4 |
2.87* |
X |
Feels desirable after chewing |
0.3 |
0.1 |
0.79 |
|
Is better than brushing teeth |
-0.6 |
-0.2 |
1.49* |
X |
Would recommend to family |
0.3 |
0.2 |
1.24 |
X |
Must always carry it wherever |
0.5 |
0.2 |
1.25 |
X |
Gains confidence after chewing |
-0.4 |
-0.1 |
0.91 |
|
Feels sophisticated after chewing |
1.2 |
0.4 |
2.53* |
X |
Is better than other two brands |
-1.1 |
-0.4 |
2.95* |
X |
*p ≤ 0.15 |
||||
From Table 1.2, the constant and five out of the ten Likert questions were significant. That is, in 85 or more samples out of 100 samples drawn from the population of this sample, it would be expected that these six items would be about what it is in this sample because the t-ratio is significant. Also to note, eight out of the ten total items were deemed important because their absolute value of beta was higher than the highest absolute value divided by two.
This means that for this particular sample, eight out of the ten brand attributes explain why immediate attitude toward the brand changed better than the other two attributes. However, only five of these attributes can be applied to estimating the population. These five attributes include “Feeling refreshed after chewing Extra”, “Would not buy Extra gum”, “Chewing Extra is better than brushing my teeth”, “Chewing Extra makes me feel sophisticated”, and “Extra is better than the other two brands”.
Table 2.1 Linear Regression Model Summary for ECLIPSE |
|||
Multiple correlation coefficient (R) |
Coefficient of multiple determination (R2) |
Standard Error of Estimate (Se) |
F-ratio to test multiple correlation coefficient |
0.3 |
8.4% |
2.0 |
0.51 |
*p ≤ 0.15 |
|||
Table 2.1 shows that the linear regression model for this sample reports a low multiple correlation coefficient at 0.3 and a low coefficient of multiple determination at 8.4%. Because the F-ratio was not significant, the results of this sample cannot be projected to the population. In other words, even though this sample reveals a very low association between the overall attitude toward the Eclipse brand with the immediate change in opinion of the Eclipse brand, the results cannot be applied to the population.
Table 2.2 Linear Regression Coefficients for ECLIPSE |
||||
Variable |
Unstandardized Coefficient (b) |
Standardized Coefficient (Beta) |
t-ratio |
Importance |
(Constant) |
0.3 |
|
0.15 |
|
Is a good chewing gum |
-0.5 |
-0.2 |
1.0 |
X |
Feels refreshed after chewing |
0.4 |
0.2 |
1.0 |
X |
Would not buy |
0.1 |
0.1 |
0.54 |
X |
Feels desirable after chewing |
0.0 |
0.0 |
0.08 |
|
Is better than brushing teeth |
0.3 |
0.1 |
0.75 |
X |
Would recommend to family |
0.1 |
0.0 |
0.10 |
|
Must always carry it wherever |
0.3 |
0.1 |
0.80 |
X |
Gains confidence after chewing |
0.1 |
0.1 |
0.40 |
X |
Feels sophisticated after chewing |
-0.3 |
-0.1 |
0.74 |
X |
Is better than other two brands |
-0.1 |
-0.1 |
0.36 |
X |
*p ≤ 0.15 |
||||
From Table 2.2, none of the results were significant for the Eclipse brand. However, eight out of the ten total items were deemed important because their absolute value of beta was higher than the highest absolute value divided by two. This means that, for this particular sample, eight out of the ten brand attributes explain better than the other brand attributes how the immediate attitude of respondents changed. However, none of the attributes can be applied to the population.
Table 3.1 Linear Regression Model Summary for ORBIT |
|||
Multiple correlation coefficient (R) |
Coefficient of multiple determination (R2) |
Standard Error of Estimate (Se) |
F-ratio to test multiple correlation coefficient |
0.6 |
34.4% |
1.8 |
2.94* |
*p ≤ 0.15 |
|||
The results in Table 3.1 are significant. Therefore, in 85 or more samples drawn from 100 of the same population as this same sample it would be expected that the multiple correlation coefficient would be about what it is in this sample. The multiple correlation coefficient, or R, is .6 and that the coefficient of multiple determination, or R2, is at 34.4%. This means that only 34.4% of the variance of the change score for Orbit is explained by the 10 Likert items for Orbit.
In other words, a moderate to high connection exists between how respondents view the Orbit brand and how they view the Orbit brand immediately after exposure to the print ad. The coefficient of multiple determination at 34.4% means that the quality of this assessment is low to moderate.
Table 3.2 Linear Regression Coefficients for ORBIT |
||||
Variable |
Unstandardized Coefficient (b) |
Standardized Coefficient (Beta) |
t-ratio |
Importance |
(Constant) |
2.8 |
|
2.13* |
|
Is a good chewing gum |
-0.2 |
-0.1 |
0.50 |
|
Feels refreshed after chewing |
-0.3 |
-0.1 |
0.76 |
|
Would not buy |
0.3 |
0.2 |
1.27 |
X |
Feels desirable after chewing |
0.7 |
0.3 |
2.06* |
X |
Is better than brushing teeth |
0.2 |
0.1 |
0.69 |
|
Would recommend to family |
-0.5 |
-0.3 |
1.80* |
X |
Must always carry it wherever |
-0.4 |
-0.2 |
1.37 |
X |
Gains confidence after chewing |
-0.1 |
-0.0 |
0.19 |
|
Feels sophisticated after chewing |
-0.9 |
-0.4 |
2.14* |
X |
Is better than other two brands |
-0.1 |
-0.0 |
0.21 |
|
*p ≤ 0.15 |
||||
Table 3.2 shows that the constant and three out of the ten Likert questions were significant. That is, in 85 or more samples out of 100 samples drawn from the population of this sample, it would be expected that these four items would be about what it is in this sample. Also to note, five out of the ten total items were deemed important because their absolute value of beta was higher than the highest absolute value divided by two.
This means that for this particular sample, five out of the ten brand attributes explain better the differences in immediate attitude change than the other five attributes. However, only three of the important attributes can be applied to estimating the population. These three attributes include “I feel desirable after chewing Orbit gum”, “I would recommend Orbit gum to my family”, and “I feel sophisticated after chewing Orbit gum”. The other two attributes that were seen to explain the change in attitude include “I would not buy Orbit gum” and “I must carry Orbit gum wherever I go”.
Multiple Regression Equation for Orbit Chewing Gum
Change score of Orbit chewing gum = 2.13 - 0.2(“Is a good chewing gum”) - 0.3(“Feels refreshed after chewing”) + 0.3(“Would not buy”) + 0.7(“Feels desirable after chewing”) + 0.2(“Is better than brushing teeth”) - 0.5(“Would recommend to family”) - 0.4(“Must always carry it wherever”) - 0.1(“Gains confidence after chewing”) - 0.9(“Feels sophisticated after chewing”) - 0.1(“Is better than other two brands”)
A discriminant analysis was performed on the chewing gum consumer preference survey data. The purpose of this analysis was to discover if the 10 brand attributes, or Likert items, could accurately discriminate between those who reported a more favorable attitude toward the Extra chewing gum brand after viewing the ad and those who reported a less favorable attitude toward the Extra brand after viewing the ad. The independent variables were the 10 Likert items and the dependent variables were the movers who became more favorable (up-movers) or less favorable (down-movers).
Respondents who neither became more nor less favorable were excluded from the analysis. As a result, the data from 42 out of the 67 total respondents were included in this analysis.
Table 1.1 Group Statistics |
||||
|
More Favorable |
Less Favorable |
||
Brand Attributes for Extra |
Mean |
Standard Deviation |
Mean |
Standard Deviation |
Is a good chewing gum |
3.4 |
1.0 |
3.6 |
1.1 |
Feels refreshed after chewing |
3.1 |
0.9 |
3.7 |
0.7 |
Would not buy |
3.1 |
1.3 |
3.9 |
0.8 |
Feels desirable after chewing |
2.6 |
1.0 |
2.4 |
0.8 |
Is better than brushing teeth |
1.6 |
1.0 |
1.9 |
0.6 |
Would recommend to family |
3.1 |
1.1 |
3.3 |
1.2 |
Must always carry it wherever |
2.1 |
0.9 |
2.4 |
1.1 |
Gains confidence after chewing |
2.2 |
1.1 |
2.4 |
1.1 |
Feels sophisticated after chewing |
2.2 |
0.9 |
2.3 |
0.8 |
Is better than other two brands |
2.2 |
1.1 |
2.9 |
0.7 |
Table 1.1 shows that the means for the brand attributes of the More Favorable group are slightly lower than the means of the Less Favorable group. This means that the people who became less favorable toward the brand after ad exposure rated the Extra brand a bit higher overall regarding individual attributes than the people who became more favorable toward the brand. Only for the item, “feels desirable after chewing,” did the More Favorable group report a higher score than the Less Favorable group.
Table 2.1 Discriminant Function Coefficients |
|||
Brand Attributes for Extra |
Standardized Discriminant |
Unstandardized Discriminant |
Importance |
(Constant) |
|
-3.4 |
|
Is a good chewing gum |
-0.2 |
-0.2 |
|
Feels refreshed after chewing |
0.4 |
0.4 |
X |
Would not buy |
0.7 |
0.6 |
X |
Feels desirable after chewing |
-0.2 |
-0.2 |
|
Is better than brushing teeth |
0.2 |
0.2 |
|
Would recommend to family |
0.0 |
0.0 |
|
Must always carry it |
-0.4 |
-0.4 |
X |
Gains confidence after chewing |
0.7 |
0.6 |
X |
Feels sophisticated after chewing |
-0.5 |
-0.6 |
X |
Is better than other two brands |
0.6 |
0.6 |
X |
Table 2.1 shows how well each brand attribute contributes to the overall More Favorable or Less Favorable attitude of the respondent toward the Extra brand after being exposed to the ad. Out of the 10 items, six of the indicated items were deemed as important based upon standardized discriminant function coefficients that were above the absolute value of the highest score divided by two. These attributes were “feels refreshed after chewing”, “would not buy”, “must always carry [Extra]”, “gains confidence after chewing”, “feels sophisticated after chewing”, and “is better than other two brands”. These discriminant function coefficients illustrate the relationship between each brand attribute and how much the respondents liked the brand after viewing the ad. For example, the more respondents felt that they gained confidence after chewing Extra, the more they liked the Extra brand after viewing the ad.
Table 3.1 Group Centroids and Statistical Significance |
|||
Group Centroids |
Wilks’ Lambda |
Chi-squared |
|
More Favorable |
Less Favorable |
||
-0.5 |
0.6 |
0.75 |
10.32 |
*p ≤ 0.15 |
|||
Table 3.1 shows that the Group Centroids, or average of the discriminate function scores, are -0.5 for the More Favorable group and 0.6 for the Less Favorable group, which are far apart and distinguishable. In general, this means that the More Favorable group is quite distinct from the Less Favorable group. However, because this test was not found to be significant, the difference between these Centroid scores cannot be projected to the population.
Table 4.1: Classification Matrix for Extra Chewing Gum |
||
|
Predicted More Favorable Respondents |
Predicted Less Favorable Respondents |
Actual More Favorable Respondents |
16 |
6 |
Actual Less Favorable Respondents |
6 |
14 |
Note: 71.4% of original grouped cases correctly classified. |
||
Table 4.1 shows that 71.4% of the More Favorable and Less Favorable groups are correctly predicted by this discriminant analysis. The observed t ratio of 3.19 indicates that these results are significant. In other words, the accuracy of these group predictions according to the disicriminant analysis can be projected to the population. That is, in 85 or more samples out of every 100 samples drawn from the same population of the sample, it would be expected that the accuracy rate would be about what it is in this sample. Overall, this discriminant analysis is a pretty good predictor of the number of More Favorable and Less Favorable respondents.
An online brand preference survey was given to 67 respondents about Eclipse, Extra, and Orbit chewing gum. A univariate analysis of variance (ANOVA) was conducted to determine if a relationship between the dependent variable (the Likert item “Chewing Eclipse chewing gum is better than brushing my teeth”) and two independent variables exist. The independent variables include the (1) after-ad-exposure change score and the (2) gender of the respondents. Additionally, a multivariate analysis of variance (MANOVA) was also conducted to determine the relationship between these same two independent variables but with all 10 Likert items posed to respondents about the Eclipse brand.
Table 1.1 Group Mean Scores and Standard Deviation |
|||
Eclipse Chewing Gum |
Gender |
||
Male |
Female |
||
Change in Perception after Ad Exposure |
More favorable |
Mean = 1.5 |
Mean = 1.9 |
Stayed the same |
Mean = 1.3 |
Mean = 1.8 |
|
Less Favorable |
Mean = 1.3 |
Mean = 2 |
|
Table 2.1 Analysis of Variance (ANOVA) |
|||||
Eclipse Chewing Gum |
Sum of Squares |
Degrees of Freedom |
Mean Square |
F-ratio |
|
Source of Variance |
Between Groups Change Score |
0.15 |
2 |
0.1 |
0.12 |
Between Groups Gender |
2.47 |
1 |
2.5 |
3.96* |
|
Within Groups Change Score by Gender |
0.15 |
2 |
0.1 |
0.12 |
|
*p ≤ 0.15 |
|||||
Table 1.1 depicts the descriptive statistics for the Likert item, “Chewing Eclipse chewing gum is better than brushing my teeth” and the change in perception of the brand after ad exposure for males and females. The women had the highest mean scores overall which indicates that women are more likely than men to consider Eclipse chewing gum to be better than brushing one’s teeth.
From Table 2.1, the data shows that only the F ratio for Between Groups Gender is significant, meaning that in 85 or more samples out of every 100 samples drawn from the same population as this sample of 67 people, it would be expected that the mean scores among the gender groups would be the same magnitude as in this sample.
Table 3.1 MANOVA Descriptive Statistics for Eclipse Chewing Gum (Eclipse) |
||||||||||||
|
Male |
Female |
||||||||||
Change in Attitude: |
More |
Same |
Less |
More |
Same |
Less |
||||||
Brand Attributes |
Mean |
SD |
Mean |
SD |
Mean |
SD |
Mean |
SD |
Mean |
SD |
Mean |
SD |
Is a good chewing gum |
3.5 |
0.5 |
3.7 |
0.8 |
3.7 |
1.2 |
3.7 |
0.6 |
3.7 |
0.7 |
3.9 |
0.9 |
Feels refreshed after chewing |
3.3 |
0.5 |
3.0 |
0.6 |
2.3 |
2.1 |
3.6 |
0.6 |
3.6 |
1.1 |
3.6 |
0.5 |
Would not buy |
3.5 |
0.5 |
3.2 |
0.8 |
4.0 |
1.0 |
3.8 |
1.3 |
3.5 |
0.9 |
3.6 |
1.4 |
Feels desirable after chewing |
3.2 |
0.8 |
2.3 |
1.2 |
2.0 |
1.7 |
2.6 |
0.8 |
2.5 |
1.1 |
2.8 |
1.3 |
Is better than brushing teeth |
1.5 |
0.8 |
1.3 |
0.8 |
1.3 |
0.6 |
1.9 |
0.7 |
1.8 |
0.8 |
2.0 |
1.0 |
Would recommend to family |
3.0 |
1.1 |
3.0 |
0.6 |
3.0 |
0.0 |
3.2 |
0.9 |
3.2 |
0.7 |
3.6 |
0.7 |
Must always carry it |
2.0 |
0.9 |
1.7 |
1.0 |
1.7 |
0.6 |
2.2 |
1.0 |
2.0 |
0.7 |
2.0 |
1.3 |
Gains confidence after chewing |
3.0 |
1.1 |
1.5 |
0.8 |
2.0 |
1.7 |
2.1 |
1.0 |
2.3 |
1.1 |
2.4 |
0.9 |
Feels sophisticated after chewing |
1.8 |
0.8 |
1.3 |
0.8 |
2.3 |
1.5 |
2.0 |
0.7 |
1.9 |
0.9 |
2.3 |
1.1 |
Is better than other two brands |
2.7 |
1.1 |
1.7 |
0.5 |
3.7 |
1.2 |
2.5 |
0.9 |
2.5 |
0.9 |
3.0 |
1.0 |
SD = standard deviation |
||||||||||||
Table 4.1 Multivariate Tests (MANOVA) |
|||
Eclipse Chewing Gum |
Wilk’s Lambda |
F-ratio |
|
Source of Variance |
Between Groups Change Score |
0.60 |
1.50* |
Between Groups Gender |
0.80 |
1.28 |
|
Within Groups Change Score by Gender |
0.69 |
1.07 |
|
*p ≤ 0.15 |
|||
Table 3.1 reflects the descriptive statistics for males and females but now in comparison with all ten Likert items. Interestingly, even though females were more likely than males to indicate that chewing Eclipse gum was better than brushing one’s teeth, this specific item overall was viewed significantly lower than other Likert items for both males and females in all change-in-attiude score groups.
From Table 4.1, the data shows that only the F ratio for Between Groups Change Score is significant meaning that in 85 or more samples out of every 100 samples drawn from the same population as this sample of 67 people, the mean scores among the change-in-attitude groups is expected to be the same magnitude as in this sample.
A factor analysis was conducted on three brands of chewing gum, Eclipse, Extra, and Orbit, to determine if certain attributes of the brand could summarize the overall attitude that was reported by 67 respondents. Therefore, using the 10 Likert items, a brand attitude score was calculated by placing related brand attributes into independent groups for each brand. A paired t-test was then executed to find the statistical significance of all the mean differences among brand attitude scores.
Eclipse Chewing Gum
Table 1.1 Factor Analysis for Eclipse chewing gum |
|||
Factors |
Eigenvalues |
% of Variance |
Cumulative Variance |
I |
3.0 |
30.3 |
30.3 |
II |
1.5 |
15.4 |
45.8 |
III |
1.3 |
12.8 |
58.6 |
IV |
1.1 |
10.8 |
69.4 |
V |
0.8 |
8.1 |
77.4 |
VI |
0.7 |
7.1 |
84.5 |
VII |
0.5 |
5.2 |
89.8 |
VIII |
0.4 |
4.2 |
94.0 |
IX |
0.3 |
3.3 |
97.3 |
X |
0.3 |
2.7 |
100.0 |
As determined by Eigenvalues ≥ 1, Table 1.1 shows that four of the ten factors are important for Eclipse chewing gum. Together, they account for 69.4% of the total variance of the Likert items.
Table 1.2 Factor Matrix Eclipse chewing gum |
|||||
Brand Attributes for Eclipse |
Communalities |
Factor I |
Factor II |
Factor III |
Factor IV |
Is a good chewing gum |
0.8 |
0.6 |
-0.5 |
-0.2 |
0.2 |
Feels refreshed after chewing |
0.8 |
0.6 |
0.2 |
-0.6 |
0.0 |
Would not buy |
0.5 |
0.4 |
-0.5 |
-0.1 |
-0.2 |
Feels desirable after chewing |
0.8 |
0.2 |
0.4 |
-0.5 |
0.6 |
Is better than brushing teeth |
0.8 |
0.1 |
0.5 |
0.6 |
0.4 |
Would recommend to family |
0.8 |
0.8 |
-0.2 |
0.1 |
0.2 |
Must always carry it wherever |
0.6 |
0.6 |
0.3 |
0.4 |
-0.1 |
Gains confidence after chewing |
0.7 |
0.5 |
0.4 |
-0.3 |
-0.4 |
Feels sophisticated after chewing |
0.7 |
0.6 |
0.4 |
0.2 |
-0.4 |
Is better than other two brands |
0.6 |
0.6 |
-0.4 |
0.3 |
0.2 |
Table 1.3 Varimax-Rotated Factor Matrix Eclipse chewing gum |
|||||
Brand Attributes for Eclipse |
Communalities |
Factor I |
Factor II |
Factor III |
Factor IV |
Is a good chewing gum |
0.8 |
0.8 |
-0.0 |
0.2 |
-0.2 |
Feels refreshed after chewing |
0.8 |
0.2 |
0.4 |
0.7 |
-0.3 |
Would not buy |
0.5 |
0.5 |
0.1 |
-0.2 |
-0.4 |
Feels desirable after chewing |
0.8 |
-0.0 |
-0.0 |
0.9 |
0.2 |
Is better than brushing teeth |
0.8 |
0.0 |
0.1 |
0.0 |
0.9 |
Would recommend to family |
0.8 |
0.8 |
0.3 |
0.2 |
0.1 |
Must always carry it wherever |
0.6 |
0.3 |
0.6 |
-0.0 |
0.4 |
Gains confidence after chewing |
0.7 |
0.0 |
0.8 |
0.3 |
-0.2 |
Feels sophisticated after chewing |
0.7 |
0.1 |
0.8 |
-0.0 |
0.2 |
Is better than other two brands |
0.6 |
0.7 |
0.1 |
-0.1 |
0.2 |
The Factor Matrix in Table 1.2 shows that Factor I contains the “good brand” item and therefore is the attitudinal scale. As a result, the “feels refreshed after chewing”, “recommend to family”, “carry it wherever”, “gains confidence”, “feels sophisticated”, and “is better than the other two brands” are all loaded onto this factor because they reported a value ≥ 0.5. Incidentally, the Varimax-Rotated Factor Matrix in Table 1.3 is not utilized because the unrotated data in Table 1.2 already gives a pretty clear result; however, it has been included for the reader’s reference.
Extra Chewing Gum
Table 2.1 Factor Analysis for Extra chewing gum |
|||
Factors |
Eigenvalues |
% of Variance |
Cumulative Variance |
I |
4.1 |
40.8 |
40.8 |
II |
1.8 |
17.7 |
58.4 |
III |
0.9 |
9.1 |
67.6 |
IV |
0.7 |
7.1 |
74.6 |
V |
0.6 |
6.4 |
81.0 |
VI |
0.5 |
5.4 |
86.4 |
VII |
0.4 |
3.9 |
90.2 |
VIII |
0.4 |
3.7 |
93.9 |
IX |
0.4 |
3.6 |
97.6 |
X |
0.2 |
2.4 |
100.0 |
For the Extra chewing gum brand, 58.4% of the variance of the original Likert items is explained by two factors as shown by Table 2.1.
Table 2.2 Factor & Varimax-Rotated Factor Matrix Extra chewing gum |
|||||
|
|
Factor Matrix |
Varimax-Rotated Factor Matrix |
||
Brand Attributes for Extra |
Communalities |
Factor I |
Factor II |
Factor I |
Factor II |
Is a good chewing gum |
0.6 |
0.5 |
0.6 |
0.1 |
0.8 |
Feels refreshed after chewing |
0.7 |
0.7 |
0.4 |
0.4 |
0.8 |
Would not buy |
0.6 |
0.2 |
0.8 |
-0.2 |
0.8 |
Feels desirable after chewing |
0.4 |
0.5 |
-0.4 |
0.6 |
0.0 |
Is better than brushing teeth |
0.5 |
0.6 |
-0.4 |
0.7 |
0.0 |
Would recommend to family |
0.5 |
0.6 |
0.4 |
0.2 |
0.7 |
Must always carry it wherever |
0.6 |
0.8 |
-0.1 |
0.7 |
0.3 |
Gains confidence after chewing |
0.7 |
0.7 |
-0.4 |
0.8 |
0.0 |
Feels sophisticated after chewing |
0.7 |
0.8 |
-0.3 |
0.8 |
0.2 |
Is better than other two brands |
0.6 |
0.7 |
-0.0 |
0.6 |
0.4 |
Table 2.2 combines both the Factor Matrix and the Varimax-Rotated Factor Matrix for the Extra brand. Because the Factor Matrix contains more than one factor that is ≥ 0.5, the Varimax-rotated Matrix is chosen in this analysis because it is less ambiguous than the unrotated result, or Factor Matrix. This time, Factor II is the attitudinal scale since it contains the "good brand" item. Other factors related to this item include “feels refreshed”, “would not buy”, and “would recommend to family”.
Orbit Chewing Gum
Table 3.1 Factor Analysis for Orbit chewing gum |
|||
Factors |
Eigenvalues |
% of Variance |
Cumulative Variance |
I |
4.0 |
39.4 |
39.9 |
II |
2.1 |
20.7 |
60.6 |
III |
0.9 |
9.3 |
69.9 |
IV |
0.8 |
8.4 |
78.3 |
V |
0.7 |
7.1 |
85.3 |
VI |
0.5 |
4.7 |
90.0 |
VII |
0.3 |
3.1 |
93.1 |
VIII |
0.3 |
2.6 |
95.7 |
IX |
0.2 |
2.4 |
98.1 |
X |
0.2 |
1.9 |
100.0 |
Finally, Table 3.1 reveals that 60.6% of the variance of the original Likert items is explained by two factors as shown by Table 3.1.
Table 3.2 Factor & Varimax-Rotated Factor Matrix Orbit chewing gum |
|||||
|
|
Factor Matrix |
Varimax-Rotated Factor Matrix |
||
Brand Attributes for Extra |
Communalities |
Factor I |
Factor II |
Factor I |
Factor II |
Is a good chewing gum |
0.8 |
0.7 |
-0.5 |
0.9 |
0.2 |
Feels refreshed after chewing |
0.7 |
0.8 |
-0.3 |
0.8 |
0.3 |
Would not buy |
0.3 |
0.4 |
-0.4 |
0.6 |
0.0 |
Feels desirable after chewing |
0.6 |
0.7 |
0.3 |
0.3 |
0.7 |
Is better than brushing teeth |
0.5 |
0.3 |
0.6 |
-0.2 |
0.7 |
Would recommend to family |
0.6 |
0.5 |
-0.6 |
0.8 |
-0.1 |
Must always carry it wherever |
0.6 |
0.7 |
0.4 |
0.2 |
0.7 |
Gains confidence after chewing |
0.6 |
0.7 |
0.4 |
0.2 |
0.7 |
Feels sophisticated after chewing |
0.8 |
0.6 |
0.6 |
0.0 |
0.9 |
Is better than other two brands |
0.6 |
0.7 |
-0.4 |
0.8 |
0.2 |
Table 3.2 also combines both the Factor Matrix and the Varimax-Rotated Factor Matrix for the Orbit brand. Again, the Varimax-rotated Matrix is chosen in this analysis because it is less ambiguous than the Factor Matrix. Factor I is determined to be the attitudinal scale since it contains the "good brand" item. Other factors loaded onto this item include “feels refreshed”, “would not buy”, “recommend to family”, and “better than other two brands”.
Table 4.1 Descriptive Statistics for Attitude Scores |
|||
Brand |
Mean |
Standard Deviation |
Sample Size |
Eclipse |
2.7 |
0.6 |
67 |
Extra |
3.2 |
||