ANOVA/MANOVA
A two-way factorial analysis of variance (ANOVA) and a multivariate analysis were conducted on this data set in order to observe any statistically significant relationships between pre to post ad exposure change score, gender, and the interaction between the change score and gender. When using a factorial design, we can draw conclusions regarding effects of the independent variables separately, as well as the combined effects of the independent variables. For the ANOVA test, only one Likert item was used (“Packaging is appealing”) for the dependent variable, while gender and whether respondents moved up, stayed the same, or moved down on the constant-sum scale, were the two factors in the equation. For MANOVA, all ten brand characteristic questions were used, while the independent variables were the same employed in the analysis of variance. For purposes of this analysis, only the Winterfresh brand responses will be used.
ANOVA : Ad Exposure Change Scores / Gender & Package Appeal |
Gender |
Up-Movers |
Same (no change) |
Down-Movers |
Male |
mean = 3.6
s.d. = 0.5
n = 11 |
mean = 3.5
s.d. = 1.0
n = 16 |
mean = 3.1
s.d. = 0.9
n = 7 |
Female |
mean = 3.5
s.d. = 1.1
n = 17 |
mean = 3.4
s.d. = 0.6
n = 19 |
mean = 3.5
s.d. = 0.5
n = 13 |
ANOVA : Between-Subjects Effects |
Source |
Sum of Squares |
Degrees of Freedom |
Mean Square |
F-Ratio |
Between Groups: Gender |
0.001 |
1 |
0.001 |
0.001 |
Between Groups: Change Score (Up-Same-Down) |
0.70 |
2 |
0.35 |
0.52 |
Between Groups: Gender by Change Score Movement |
0.76 |
2 |
0.38 |
0.57 |
Within Groups: Error |
51.29 |
77 |
0.67 |
-- |
Total |
1038.00 |
83 |
-- |
-- |
* indicates p <= 0.15
There is little difference between the mean scores on the “package is appealing” brand characteristic and those that are male and those that are female. There was a diminutive difference between the male and female down-movers with the males having a mean score of 3.1, whereas the females appeared to have found the package slightly more attractive, with an average score of 3.5. The group that found Winterfresh’s packaging most appealing were male up-movers (positive pre to post ad exposure change scores), with an overall average mean of 3.6. The group that was least impressed by the chewing gum’s packaging were male down-movers, with a mean score of 3.1. It is interesting to note that all the female means, whether they were up-movers, down-movers, or stayed the same, to a great extent, did not diverge from each other. It can be said that all individuals, whether male or female found the packaging to be relatively good-looking. Overall, these results signify that gender does not essentially make a difference in how one sex views an aesthetically pleasing package. It does not make a difference if you are male or female and if you liked or disliked the brand’s packaging.
Based on the sum of squares inferential statistic used to test the significance of gender, pre to post ad exposure change score on the constant-sum scale, and the interaction effect between gender and change score, we can not project any of our findings onto the population as a whole. According to our findings, there is no significant main effect of gender, no significant effect of movement (change score on the constant-sum scale), and no significant interaction effect between gender and the ad exposure change score, because every F-ratio was found to not be significant. Therefore, in 85 or more of 100 samples drawn from the same population as this sample, the mean scores of the independent variable, gender, can not be projected as having the same magnitude as in this sample. Furthermore, in 85 or more of 100 samples drawn from the same population as this sample of 83 respondents, we can not project the mean scores of the pre to post ad exposure change score groups (up, same, or down-movers) onto the population as a whole. Lastly, in 85 or more of a 100 samples drawn form the same population as this sample, we can not project the mean scores of the interaction between the pre to post ad exposure change scores and whether the respondent was male or female onto the populace. In summation, we can not project any of the six cells above onto the entire population.
MANOVA: Ad Exposure Change Score / Gender & Brand Characteristics |
Brand Characteristic |
Male |
Female |
Up-Movers
n = 11 |
Same
n = 16 |
Down-Movers
n = 7 |
Up-Movers
n = 17 |
Same
n = 19 |
Down-Movers
n = 13 |
mean |
s.d. |
mean |
s.d. |
mean |
s.d. |
mean |
s.d. |
mean |
s.d. |
mean |
s.d. |
1. Is a flavorful gum. |
4.0 |
0.5 |
3.9 |
0.6 |
3.7 |
0.5 |
3.9 |
0.4 |
3.5 |
1.2 |
4.3 |
0.5 |
2. Is an excellent brand of gum. |
3.8 |
0.6 |
3.6 |
0.7 |
3.3 |
0.8 |
3.8 |
0.6 |
3.4 |
0.9 |
3.6 |
1.3 |
3. Has a long lasting flavor. |
3.9 |
0.7 |
3.6 |
0.8 |
3.1 |
0.7 |
3.5 |
0.9 |
3.3 |
1.1 |
3.7 |
0.6 |
4. Is too overwhelming. |
3.6 |
0.8 |
3.8 |
0.8 |
3.0 |
0.8 |
3.4 |
1.1 |
3.3 |
0.9 |
3.5 |
1.0 |
5. Keeps your breath fresh. |
3.8 |
0.4 |
3.6 |
1.2 |
3.6 |
0.5 |
3.8 |
0.7 |
3.5 |
1.0 |
3.6 |
1.3 |
6. Has an odd texture. |
3.6 |
0.5 |
3.8 |
0.8 |
3.0 |
0.8 |
3.0 |
1.1 |
3.5 |
0.8 |
3.5 |
0.7 |
7. Packaging is appealing. |
3.6 |
0.5 |
3.5 |
1.0 |
3.1 |
0.9 |
3.5 |
1.1 |
3.4 |
0.6 |
3.5 |
0.5 |
8. Is too sugary. |
3.5 |
0.8 |
3.1 |
0.9 |
3.0 |
0.6 |
3.1 |
0.9 |
3.3 |
1.0 |
3.2 |
0.7 |
9. Gets hard after chewing. |
3.2 |
0.8 |
3.0 |
1.0 |
2.1 |
0.9 |
3.2 |
0.7 |
3.1 |
0.9 |
3.2 |
0.8 |
10. Prefer over other gums. |
2.8 |
1.3 |
2.8 |
1.2 |
2.7 |
1.1 |
2.7 |
0.9 |
2.3 |
1.2 |
2.7 |
1.0 |
MANOVA: Multivariate Tests |
Source |
Wilks’ Lambda |
F - Ratio |
Hypothesis Degrees of Freedom |
Error Degrees of Freedom |
Between Groups: Gender |
0.91 |
0.71 |
10 |
68 |
Between Groups: Change Score (Up-Same-Down) |
0.76 |
0.99 |
20 |
136 |
Between Groups: Gender by Change Score Movement |
0.75 |
1.07 |
20 |
136 |
* indicates p <= 0.15
According to the tables above, there is little to moderate difference between the mean scores of all the brand characteristics and those that are male and female and whether they moved up, down, or stayed the same after ad exposure. The most variance between mean scores arose with Likert item # 9 (“Gets hard after chewing”), who had the lowest mean score of 2.1. Brand attribute # 5 (“Keeps your breath fresh”) had steady and similar mean scores across the board, as well as Likert item # 7 (“Packaging is appealing”) , # 8 (“Is too sugary”) and # 10 (“Prefer over other gums”).
Based on the Wilks’ Lambda inferential statistic to test for significance of pre to post ad exposure change score, gender, and the interaction between the change score and specified sex, we find again that we can not project our results onto the entire population. The “Multivariate Tests” table above shows that the Error Degrees of Freedom (within group) is much larger than the Hypothesis Degrees of Freedom (between group), which greatly decreases the likelihood that the F-ratio is significant. The Wilks’ Lambda was not found significant in any of the three effects. When looking at each brand characteristic with each independent variable (gender and ad exposure change score) and the interaction between the two, it appears that a significant relationship does not exist between the two. Therefore, due to lack of statistical significance, in 85 or more samples out of every 100 drawn from the same population, the above 60 mean scores cannot be projected to the population. It can be concluded that there is no major effect or relationship between how the respondents rated the ten Likert items about the brand, their gender, and their like or dislike of the brand’s ad.