Analysis
Basic statistical analyses were conducted on a data set from a consumer brand preference survey. This survey asked respondents their opinions of three brands of body lotion; Vaseline, Aveeno and Olay. The purpose of this survey was to assess whether respondents’ brand preferences changed after viewing magazine advertisements for each brand.
1. Correlated (paired) t-tests were completed using Brand Index Scores from a survey sample of 69 and 72 respondents. The Brand Index Scores were calculated from 10 Likert questions which assessed respondents’ opinions of the brands. Fifty is the highest possible score for each respondent and 10 is the lowest possible score. A higher score indicates the respondent is more favorable towards the brand.
Table 1A: Brand Index Scores
|
Brand |
Mean Brand Index Score |
Sample Size (N) |
Standard Deviation |
Pair 1 |
Vaseline |
33.3 |
69 |
6.3 |
Aveeno |
34.6 |
69 |
6.6 |
|
Pair 2 |
Vaseline |
33.2 |
69 |
6.3 |
Olay |
34.0 |
69 |
6.8 |
|
Pair 3 |
Aveeno |
34.8 |
72 |
6.5 |
Olay |
33.9 |
72 |
6.7 |
Table 1A shows the Mean and Standard Deviation of Brand Index Scores for each brand. Aveeno had the highest mean Brand Index Score of 34.6 on the sample size of 69 and a mean Brand Index Score of 34.8 on a sample size of 72. A higher Brand Index Score indicates the more respondents favored Aveeno. However the mean Brand Index Scores were very close, indicating that respondents liked the brands similarly. The Standard Deviation indicates how close responses were to the mean Brand Index Score. Responses for Olay showed a greater variance from the mean, whereas Vaseline had the least variances from the mean out of the 3 brands. The results however for standard deviation for all 3 brands were again very similar overall.
Table 1B: Paired Samples t-tests for Vaseline, Aveeno and Olay
Brand Index Score Paired Sample |
t-ratio |
Vaseline and Aveeno |
1.21* |
Vaseline and Olay |
0.61 |
Aveeno and Olay |
1.02 |
*P<=.15
Paired sample t-tests for all 3 brands were performed to determine the statistical significance of the mean Brand Index Scores. The t-ratio for Vaseline and Aveeno indicates statistical significance. In 85 or more samples out of every 100 samples drawn from the same population as this sample, the mean scores for Vaseline and Aveeno would be about what they are in this sample. The results of this sample can be projected to the population for this test. For the Brand Index Score paired samples: Vaseline and Olay; and Aveeno and Olay, the t-ratio indicates there is no statistical significance and results from these samples cannot be project to the larger population.
2. Between-Groups t-tests were calculated based on the Brand Index Scores and the Advertising Index Scores for Aveeno. The Brand Index Scores indicate how the respondent feels towards the brand itself after seeing the advertisement. The Ad Index Scores measures how the respondent feels towards the ad itself. For both the Brand Index Score and the Ad Index Score, respondents were grouped according to those whose score moved up after seeing the ad and those whose score moved down post ad exposure. For the Brand Index Score, the highest possible score would be 50 and therefore, a higher score indicates the respondent liked the brand more after seeing the ad. For the Ad Index Score the highest possible score is 10 and therefore, a higher score indicates the respondents liked the ad more.
Table 2A: Brand Index Score Change: Aveeno
Change Score |
Sample Size |
Mean |
Standard Deviation |
t-ratio |
Moved Up |
19 |
36.6 |
5.1 |
1.35* |
Moved Down |
29 |
34.6 |
5.3 |
*P<=.15
Table 2A illustrates that more respondents (a total of 29) became less favorable towards the brand after seeing the ad. The asterisk beside the t-ratio number indicates that the results are statistically significant. In 85 or more samples out of every 100 samples drawn from the same population as this sample, the mean scores for Aveeno would be about what they are in this sample. The results of this sample can be projected to the population for this test.
Table 2B: Ad Index Score Change: Aveeno
Change Score |
Sample Size |
Mean |
Standard Deviation |
t-ratio |
Moved Up |
19 |
5.3 |
2.2 |
.88 |
Moved Down |
31 |
4.6 |
2.5 |
*P<=.15
Table 2B shows that more respondents’ Ad Index Score moved down. This indicates that respondents were felt less favorably towards the ads after seeing them. The t-ratio for the Ad Index Score indicates that the results of this test are not statistically significant and therefore can not be applied to the general population.
3. Chi-Square Significance Test: Vaseline
Table 3: Pre-to-Post ad Exposure in Relation to Median Brand Index Scores
|
Above Median |
Below Median |
|
Moved Up |
Count |
12 |
15 |
Stayed the Same |
Count |
10 |
13 |
Moved Down |
Count |
12 |
4 |
Chi-Squared=4.67*
*P<=.15
Table 3 indicates that after seeing the advertisement for Vaseline, most respondents (a total of 27) felt more favorable towards the brand and their scores moved up. A total of 23 survey respondents had no change of opinion and 16 of the respondents were less favorable towards Vaseline after seeing the ad. This table also indicates which respondents were above or below the median Brand Index Score for Vaseline. Of the respondents whose scores moved up after seeing the ad, 44.4% of respondents were above the median. Most notably however, 75% of respondents whose score moved down after seeing the ad were still above the median Brand Index Score. The chi-square significance test indicates that these results are statistically significant. In 85 or more samples out of every 100 samples drawn from the same population as this sample, the median scores for Vaseline would be about what they are in this sample. The results of this sample can be projected to the population for this test.
4. Frequency Test:
Table 4: Respondents who moved up, down or stayed the same in pre-to-post advertisement exposure
Brand |
Moved Up |
Stayed the Same |
Moved Down |
Vaseline |
33 |
28 |
19 |
Aveeno |
19 |
30 |
31 |
Olay |
29 |
32 |
19 |
Table 4 calculated the frequencies for Vaseline, Aveeno and Olay. This table calculated frequencies based on pre-to-post exposure to each brands magazine advertisement. Frequencies for each brand were categorized to show whether respondents felt more or less favorable towards the brand after seeing the advertisement. Scores that moved up indicates that after seeing the advertisement the respondent felt more favorable towards the brand. Scores that went down indicates that respondents felt less favorable towards the brand after seeing the advertisement. Notably more respondents’ scores went down after seeing the Aveeno advertisements indicating that Aveeno’s ad had the least influence in promoting favorable brand opinions. This table also indicates that the advertisement for Vaseline had the best influence out of the three brands on promoting a favorable brand opinion among respondents. For Vaseline, more respondents’ scores moved up after seeing the ad. Olay had similar results to Vaseline however most of the respondents scores remained the same pre-to-post ad exposure.
5. Brand Index Score Comparison: Vaseline and Aveeno
Table 5: Brand Index Score Comparison for Vaseline and Aveeno
Brands |
Frequency |
Vaseline & Aveeno |
26 |
Table 5 displays the result of a Frequency Test. Out of 80 respondents there was a frequency of 26 people who have a higher Brand Index Score for Vaseline than Aveeno. This indicates that 26 out of a total of 80 respondents liked Vaseline more than Aveeno.
6. Simple Correlation Coefficient
Table 6: Correlation of Brand Index Scores
Brands |
Correlation |
Vaseline & Aveeno |
.05 |
*P<=.15
In Table 6 the Correlation Coefficient was calculated based on Brand Index Scores for Vaseline and Aveeno. A Correlation of .05 indicates there is a positive relationship between the two brands but it is a weak relationship. This indicates that people who like Vaseline also like Aveeno. These results however are not statistically significant and can not therefore be applied to the general population.
7. Simple Correlation Coefficient
Table 7: Correlation of Brand Index Scores: Women
Brands |
Correlation |
Vaseline & Aveeno |
.03 |
*P<=.15
In Table 7 the Correlation Coefficient was calculated based on Brand Index Scores for Vaseline and Aveeno and only Female respondents were selected. In this case there was a correlation of .03 indicating a positive correlation between the brands but indicates this is a weak correlation. The correlation of .03 has gone down from a correlation of .05 as seen in Table 6. This means there is a slightly weaker correlation between the two brands among female respondents. In this case the results are not statistically significant and can not be used to apply to the general population.
Multiple Linear regression Analyses were conducted for three brands of body lotion, Vaseline, Aveeno and Olay. An online survey was conducted by e-mail and 80 responses were collected. The independent variables are the 10 Likert questions about each brand. The dependent variable is the change in the respondent’s feeling towards the brand after exposure to the 3 advertisements on a constant-sum scale.
The Multiple Correlation Coefficient (R) measures the relationship between the respondent’s feelings towards the brand and their feelings towards the brand’s ad. Each of the 3 brands in this survey showed a moderate correlation. When the multiple correlation coefficient equals 1.0 this is a strong positive relationship and when the multiple correlation coefficient is -1.0 this would indicate a strong negative relationship.
The Coefficient of Multiple Determination (R2) refers the variance in the respondents’ feelings towards the ad considering their feelings towards the brand which is explained as a percentage. The closer the value of R2 is to 100%, the stronger the correlation is between them. The Coefficient of Multiple Determination was low for each of the 3 brands indicating a weak correlation overall.
The Standard Error of the Estimate (SE) measures the variation in the predicted values that is used to develop confidence intervals around the predicted value. Each Standard of the Estimate in Table 1 can be interpreted as high because the Coefficient of Multiple Determination is low.
The F-ratio included in this table indicates whether the data is statistically significant and if it can therefore be applied to a larger population.
Table 1: Correlation Between Brand Change Score and Brand Index Items
Brand |
Multiple Correlation Coefficient |
Coefficient of Multiple Determination |
Standard Error of the Estimate |
F-ratio |
Vaseline |
0.5 |
26.2% |
1.9 |
2.17* |
Aveeno |
0.4 |
14.0% |
2.1 |
1.06 |
Olay |
0.4 |
16.9% |
2.0 |
1.31 |
*P<=.15
Vaseline
Table 1 indicates a Coefficient of Multiple Determination (R2) of 26.2%. This would indicate a weak relationship with only 26.2% of variant in change score for Vaseline is explained by the 10 Likert items for Vaseline. There doesn’t seem to be a connection between how people view the brand and the way they view the ad. Vaseline has the highest Coefficient of Multiple Determination out of the 3 brands. The Multiple Correlation Coefficient for Vaseline is 0.5 which indicates the data has a moderate correlation. The Multiple Correlation Coefficient is slightly higher than that of the other 2 brands. The F-ratio for Vaseline also indicates the data is statistically significant. In 85 or more samples out of every 100 samples drawn from the same population as this sample it would be expected that the multiple correlation coefficient would be about 0.5 as it is in this sample.
Aveeno
In table 1 the multiple regression analysis for Aveeno indicates a Coefficient of Multiple Determination of (R2) of 14.0%. This would indicate a weak relationship with only 14.0% of variant in change score for Aveeno is explained by the 10 Likert items for Aveeno. There doesn’t seem to be a connection between how people view the brand and the way they view the ad. Aveeno had the lowest value of the 3 brands for the Coefficient of Multiple Determination. This indicates a weaker relationship than that found for Vaseline. The Coefficient of Multiple Correlations is 0.4 which indicates the data has a moderate correlation. The F-ratio for Aveeno indicates the data is not statistically significant and can not therefore be applied to a larger population.
Olay
In Table 1 the multiple regression analysis for Aveeno indicates a Coefficient of Multiple Determination of (R2) of 16.9%. This would indicate a weak relationship with only 16.9% of variant in change score for Olay is explained by the 10 Likert items for Olay. This relationship is lower than that of Vaseline and slightly higher than that for Aveeno. There doesn’t seem to be a connection between how people view the brand and the way they view the ad. The coefficient of multiple correlations is 0.4 which indicates the data has a moderate correlation. The F-ratio for Olay indicates the data is not statistically significant and can not therefore be applied to the population.
Table 2: Standardized and Unstandardized Coefficients
|
Vaseline |
Aveeno |
Olay |
||||||
Likert Item |
b |
Beta |
t-ratio |
b |
Beta |
t-ratio |
b |
Beta |
t-ratio |
Constant |
1.9 |
|
1.16 |
-2.3 |
|
1.14 |
-0.2 |
|
0.11 |
Good brand |
0.6 |
0.3 |
1.36 |
1.1 |
0.4 |
1.90* |
0.7 |
0.2 |
1.10 |
Prefer over other brands |
-0.7 |
-0.3 |
2.25* |
-0.2 |
-0.1 |
0.63 |
-0.3 |
-0.1 |
0.78 |
Would not use |
0.1 |
0.1 |
0.49 |
0.0 |
0.0 |
0.00 |
-0.5 |
-0.3 |
1.42 |
Would purchase |
0.1 |
0.0 |
0.14 |
-0.1 |
-0.1 |
0.23 |
0.2 |
0.1 |
0.55 |
Prevents dryness |
0.9 |
0.3 |
1.63* |
0.5 |
0.2 |
0.65 |
0.4 |
0.1 |
0.71 |
Feels oily on skin |
-0.1 |
0.0 |
0.18 |
0.3 |
0.1 |
0.74 |
-0.6 |
-0.2 |
1.53* |
Makes skin softer |
-1.1 |
-0.4 |
2.37* |
-0.9 |
-0.3 |
1.42 |
-0.2 |
-0.1 |
0.27 |
Feel more attractive using it |
-0.2 |
-0.1 |
0.50 |
0.6 |
0.2 |
1.66* |
0.6 |
0.2 |
1.32 |
Trust the brand |
0.0 |
0.0 |
0.07 |
-0.7 |
-0.3 |
1.01 |
0.2 |
0.1 |
0.36 |
Would recommend to friends |
-0.3 |
-0.1 |
0.72 |
0.1 |
0.0 |
0.11 |
-0.4 |
-0.2 |
0.82 |
*P<=.15
b=Unstandardized Coefficient
Beta=Standardized Coefficient
In Table 2, Beta is a Standardized Coefficient that allows direct comparison between coefficients. Beta tells the relative importance of each independent variable (Likert item) in explaining the variant of the dependent variable (change score). The Likert items are important that have at least half the size of the largest Beta. In Table 2, b is the Unstandardized Coefficient which is the regression coefficient. The t-ratio is used in Table 2 to test the constant term and the regression coefficient. The t-ratio determines statistical significance of the data, whether or not the data may be applied to a larger population.
Vaseline
Four of the 10 Likert brand attributes were important in this analysis as indicated by the Beta value: “Vaseline is a good brand”; “Prefer Vaseline over other brands”; “Vaseline prevents dryness”; and “Vaseline makes skin softer”. Of these four brand attributes 3 of them: “Prefer vaseline over other brands”; “Vaseline prevents dryness”; and “Vaseline makes skin softer” are statistically significant as indicated by the t-ratio. In 85 or more samples out of every 100 samples drawn from the same population as this sample it would be expected that the coefficients for the brand attributes “Prefer vaseline over other brands”; “Vaseline prevents dryness”; and “Vaseline makes skin softer” would be about what they are in this sample.
Aveeno
Four of the Likert items are important in this analysis as indicated by the Beta value: “Aveeno is a good brand”; “Aveeno makes skin softer”; “Feel more attractive using Aveeno”; and “Aveeno is a brand they trust”. Of these four brands the t-ratio indicates that 2 of these Likert items are statistically significant: “Aveeno is a good brand” and “Feel more attractive using Aveeno” and can be applied to the general population. In 85 or more samples out of every 100 samples drawn from the same population as this sample it would be expected that the coefficients for the brand attributes “Aveeno is a good brand” and “Feel more attractive using Aveeno” would be about what they are in this sample.
Olay
Five of the Likert items are important in this analysis as indicated by the Beta value: “Would recommend Olay to friends”; “Olay is a good brand”; “Feel more attractive using Olay”; “Olay feels oily on skin”; and “would not use Olay”. Of these only the brand attribute “Olay feels on oily skin” is statistically significant as indicated by the t-ratio and therefore in 85 or more samples out of every 100 samples drawn from the same population as this sample it would be expected that the coefficient for the brand attribute “Olay feels oily on skin” would be about what it is in this sample.
Multiple Regression Equation for Vaseline:
Change Score for Vaseline= 1.9 (Constant) + 0.6 (“Vaseline is a good brand”) - 0.7 (“Prefer Vaseline”) + 0.1 (“Would not use Vaseline”) + 0.1 (“Would purchase Vaseline”) + 0.9 (“Vaseline prevents dryness”) - 0.1 (“Vaseline feels oily”) - 1.1 (“Vaseline makes skin softer”) -0.2 (“Feel more attractive using Vaseline”) + 0.0 (“Trust Vaseline”) – 0.3 (“Would recommend Vaseline”)
A multiple discriminant analysis was conducted to determine if respondents’ attitudes to the Olay brand affected whether the respondents’ were “up-movers” or “down-movers”. “Up-movers” are those respondents who had a positive change score for Olay on the constant sum scale after viewing the advertisements, while the “down-movers” had a negative score. The categorical dependent variable is group membership. The independent variables are the scores on the 10 brand Likert items.
Table 1: Up-Mover and Down-Mover Group Statistics for Olay
Brand Attribute (Likert Item) |
Up Movers |
Down Movers |
Unstandardized Discriminant Function Coefficient |
Standardized Discriminant Function Coefficient |
||
Mean |
Standard Deviation |
Mean |
Standard Deviation |
|||
Olay is a good brand |
4.0 |
0.6 |
3.7 |
0.8 |
1.0 |
0.7 |
Prefer Olay |
3.0 |
1.2 |
3.0 |
1.0 |
0.0 |
0.0 |
Would not use Olay |
3.5 |
1.1 |
3.4 |
1.1 |
0.0 |
-0.1 |
Would purchase Olay |
3.4 |
0.9 |
3.2 |
1.1 |
0.3 |
0.3 |
Olay prevents dryness |
3.7 |
0.5 |
3.6 |
0.6 |
0.7 |
0.4 |
Olay feels oily on skin |
3.2 |
0.7 |
3.6 |
0.5 |
-1.3 |
-0.9 |
Olay makes skin softer |
3.6 |
0.6 |
3.5 |
0.6 |
-0.4 |
-0.3 |
Feel more attractive using Olay |
3.3 |
0.8 |
2.9 |
1.0 |
0.7 |
0.6 |
Trust brand Olay |
3.9 |
0.7 |
3.6 |
1.0 |
0.0 |
0.0 |
Would recommend Olay |
3.3 |
1.0 |
3.2 |
1.2 |
-0.6 |
-0.6 |
n=Sample Size
Table 1 indicates that the mean scores for both the up-movers and down-movers are similar which would indicate very few differences between the two groups’ perceptions of the Olay brand. Discriminant function coefficient analyses were conducted to determine the relative importance of variables (brand attributes) in explaining differences between groups. “Olay is a good brand”, “Olay feels oily on skin”, “feel more attractive using Olay”, and “would recommend Olay” appear to be the most important attributes in determining whether respondents were up-movers or down-movers.
Table 2: Group Centroids and Inferential Statistics
Group Centroids |
Wilks’ Lambda |
Chi-squared |
|
Up-Movers |
Down-Movers |
||
0.4 |
-0.6 |
0.78 |
9.70 |
*p<=.15
Table 2 indicates that chi-square is not statistically significant which signifies that the group centroids cannot be applied to a larger population because Wilks’ Lambda converted to chi-squared is not statistically significant. The Wilks’ Lambda value tests the significance of the group centroid values which are far apart and distinguishable from one another.
Table 3: Classification Matrix
Group Membership |
Predicted Group Membership |
Total |
|
Up-Movers |
Down-Movers |
||
Up-Movers |
22 |
5 |
27 |
Down-Movers |
11 |
7 |
18 |
64.4% of original grouped cases are correctly classified
Classification results indicate that only 7 out of 18 down-movers were accurately classified and 22 out of 27 up-movers were accurately classified. The classification matrix indicates that 64.4% of the original grouped cases were classified correctly which would indicate a moderate to slightly higher rate of accuracy.
t-ratio:
to= (0.64 – 0.5) / √ [(0.64 x 0.36) / 45] + [(0.5 x 0.5) / 45] = 1.36*
*p<=.15
The t-ratio indicates that the result, 64.4% of original grouped cases classified correctly, is statistically significant and can be applied to the population. The observed t-ratio of 1.36 is greater than the critical value of 1.04. In 85 or more samples out of 100 samples it can be expected classification accuracy would be about what it is in this case.
A two-way ANOVA analysis of variance was conducted to determine the relationship between one continuously measured dependent variable and two categorical independent variables. The dependent variable was the Likert item “would purchase Olay”. The independent categorical variables for these tests were gender (male/female) and change score (up-movers/non-movers/down-movers).
A MANOVA multivariate analysis of variance was also conducted to determine the relationships between the independent variables, gender (male/female) and change score (up-movers, non-movers, down-movers) and the dependent variables are all 10 Likert items.
ANOVA
Table 1: Mean Scores for Different Groups for Likert item “Would purchase Olay”.
Change Score |
Gender |
|
Male |
Female |
|
Up-Movers |
Mean: 3.2 |
Mean: 3.5 |
Non-Movers |
Mean: 2.9 |
Mean: 3.4 |
Down-Movers |
Mean: 2.7 |
Mean: 3.6 |
Table 1 shows descriptive data for Likert item, “Would purchase Olay” and pre and post exposure change scores for males and females. Table 1 indicates that female, down-movers had the highest mean.
Table 2: Tests of Between Group Effects
|
F-ratio |
Change Score |
0.36 |
Gender |
5.30* |
Change Score by Gender |
0.34 |
*p<=.15
Table 2 indicates that only the F-ratio for gender is statistically significant. In 85 or more of 100 samples drawn from the same population as in this sample it would be expected that the difference in the mean Likert scores for “ would purchase Olay body lotion” between the male and female groups would be about what it is here. The results of this test may be projected to the population. The F-ratio results for change score and change by gender were not found to be statistically significant and may not therefore be projected to the population.
MANOVA
Table 3: Multivariate Analysis of Variance
Brand Attribute (Likert Item) |
Male |
Female |
||||||||||
n=18 |
n=58 |
|||||||||||
Up-Movers |
Non-Movers |
Down-Movers |
Up-Movers |
Non-Movers |
Down-Movers |
|||||||
n=5 |
n=6 |
n=7 |
n=22 |
n=25 |
n=11 |
|||||||
Mean |
Std. |
Mean |
Std. |
Mean |
Std. |
Mean |
Std. |
Mean |
Std. |
Mean |
Std. |
|
Olay is a good brand |
4.0 |
0.7 |
3.3 |
0.5 |
3.4 |
0.5 |
4.0 |
0.7 |
3.8 |
0.9 |
3.8 |
0.9 |
Prefer Olay |
3.8 |
0.8 |
2.8 |
0.8 |
2.7 |
0.5 |
2.9 |
1.2 |
3.0 |
1.0 |
3.2 |
1.2 |
Would not use Olay |
3.0 |
0.7 |
2.8 |
1.2 |
2.9 |
0.9 |
3.6 |
1.1 |
3.5 |
1.0 |
3.7 |
1.2 |
Would purchase Olay |
3.2 |
0.4 |
3.0 |
0.9 |
2.7 |
0.8 |
3.5 |
1.0 |
3.4 |
0.9 |
3.6 |
1.2 |
Olay prevents dryness |
3.2 |
0.4 |
3.3 |
0.5 |
3.4 |
0.5 |
3.8 |
0.4 |
3.4 |
0.9 |
3.7 |
0.6 |
Olay feels oily on skin |
3.2 |
0.4 |
3.2 |
0.4 |
3.1 |
0.4 |
3.2 |
0.8 |
3.4 |
0.8 |
3.8 |
0.4 |
Olay makes skin softer |
3.2 |
0.4 |
3.3 |
0.5 |
3.0 |
0.0 |
3.7 |
0.6 |
3.4 |
0.8 |
3.8 |
0.6 |
Feel more attractive using Olay |
3.2 |
0.4 |
3.0 |
0.0 |
2.3 |
0.5 |
3.2 |
0.8 |
2.8 |
0.8 |
3.3 |
1.0 |
Trust Olay brand |
4.0 |
0.0 |
3.3 |
0.5 |
3.3 |
0.8 |
3.8 |
0.7 |
3.6 |
0.9 |
3.8 |
1.0 |
Would recommend Olay |
3.4 |
0.5 |
3.0 |
0.6 |
2.7 |
1.0 |
3.3 |
1.0 |
3.2 |
0.9 |
3.6 |
1.2 |
Std. Dev. = Standard Deviation
Table 3 indicates that the brand attribute “Olay is a good brand” had the highest mean score for male up-movers, and female up-movers. The brand attribute “Trust Olay brand” also had the highest mean score for male up-movers.
Table 4: Multivariate Tests
|
Wilks’ Lambda |
F-ratio |
Change Score |
0.82 |
0.64 |
Gender |
0.77 |
1.80* |
Change by Gender |
0.66 |
1.43* |
*P<=.15
The F-ratios in Table 4 indicate that there is one significant main effect for gender and the 10 Likert items and a significant interaction effect which can therefore be projected to the population. In 85 or more of 100 samples drawn from the same population as in this sample it would be expected that the results for gender and change by gender would be about what they are in this sample.
A factor analysis was conducted for the three brands of body lotion: Vaseline, Aveeno and Olay to determine how the brand attitude Likert items for each brand were related to each other. The Eigenvalues are the sum of squared factor loadings for a given factor and when an Eigenvalue is divided by the number of variables in the factor analysis the result is the percent of variance accounted for by a factor. A Factor Matrix was created to determine factor loadings and a Varimax Rotated Matrix was created in an attempt to overcome ambiguous results from the Factor Matrix. In the factor analysis however the Varimax Rotated Matrix did not provide less ambiguous results for any of the 3 brands. Factor loadings indicate correlations between the original variables and the factors. Communality is the amount of variance explained by all factors for a given variables’ original variance. Based on results from the Factor Matrix, mean attitude scores for each brand were calculated. The means of the attitude scores were compared using a t-test for statistical significance.
Table 1
Factor |
Vaseline |
Aveeno |
Olay |
||||||
Eigenvalues |
% of variance |
Cumulative % of variance |
Eigenvalues |
% of variance |
Cumulative % of variance |
Eigenvalues |
% of variance |
Cumulative % of variance |
|
I |
5.2 |
52.3 |
52.3 |
5.9 |
59.1 |
59.1 |
6.2 |
61.5 |
61.5 |
II |
1.1 |
11.5 |
63.8 |
1.6 |
15.7 |
74.8 |
0.9 |
|
|
III |
1.1 |
11.1 |
75.0 |
0.7 |
|
|
0.8 |
|
|
IV |
0.6 |
|
|
0.6 |
|
|
0.5 |
|
|
V |
0.6 |
|
|
0.4 |
|
|
0.5 |
|
|
VI |
0.4 |
|
|
0.3 |
|
|
0.4 |
|
|
VII |
0.3 |
|
|
0.2 |
|
|
0.3 |
|
|
VIII |
0.3 |
|
|
0.1 |
|
|
0.2 |
|
|
IX |
0.2 |
|
|
0.1 |
|
|
0.2 |
|
|
X |
0.2 |
|
|
0.1 |
|
|
0.1 |
|
|
Table 1 indicates the Eigenvalues for each of the 3 brands. In the case of Vaseline 75.0% of the variance in the original 10 Likert items is explained by 3 factors and therefore 25.0% is lost. The first factor for Vaseline explains 52.3% of the variance of the original 10 Likert items. The second factor explains 11.5% of variance of original 10 Likert items and the third factor explains 11.1% of variance of original 10 Likert items. For Aveeno, 74.8% of variance in the original 10 Likert items is explained by 2 factors and therefore 25.2% is lost. The first factor for Aveeno explains 59.1% of the variance of the original 10 Likert items and the second factor for Aveeno explains 15.7% of the variance of the original 10 Likert items. For Olay, 61.5% of the variance in the original 10 Likert items is explained by 1 factor and therefore 38.5% is lost.
Table 2: Vaseline
Brand Attribute |
Communalities |
Factor Matrix |
Varimax Factor Matrix |
||||
I |
II |
III |
I |
II |
III |
||
Vaseline is a good brand |
0.8 |
0.9 |
-0.1 |
0.1 |
0.8 |
0.3 |
0.2 |
Prefer Vaseline |
0.6 |
0.7 |
0.2 |
-0.3 |
0.5 |
0.6 |
0.2 |
Would not use Vaseline |
0.6 |
0.7 |
0.3 |
0.3 |
0.5 |
0.2 |
0.5 |
Would purchase Vaseline |
0.7 |
0.8 |
0.1 |
0.0 |
0.7 |
0.4 |
0.3 |
Using Vaseline prevents dryness |
0.9 |
0.7 |
-0.6 |
0.1 |
0.9 |
0.0 |
-0.2 |
Vaseline feels oily on skin |
0.8 |
0.2 |
0.7 |
0.5 |
0.0 |
0.0 |
0.9 |
Vaseline makes skin softer |
0.7 |
0.8 |
-0.2 |
0.2 |
0.8 |
0.2 |
0.1 |
Feel more attractive using Vaseline |
0.8 |
0.4 |
0.3 |
-0.7 |
0.0 |
0.9 |
-0.1 |
Trust Vaseline brand |
0.8 |
0.8 |
-0.2 |
0.1 |
0.8 |
0.3 |
0.2 |
Would recommend Vaseline |
0.8 |
0.8 |
0.2 |
-0.3 |
0.5 |
0.7 |
0.3 |
In Table 2 the Factor Matrix and the Varimax Rotated Matrix for Vaseline both indicated ambiguous results for 2 Likert items and therefore the original Factor Matrix was used for this analysis. More Likert items loaded on Factor I which was used in the attitude score analysis
Table 3: Aveeno
Brand Attribute |
Communalities |
Factor Matrix |
Varimax Factor Matrix |
||
I |
II |
I |
II |
||
Aveeno is a good brand |
0.8 |
0.9 |
-0.2 |
0.9 |
0.1 |
Prefer Aveeno |
0.7 |
0.8 |
0.2 |
0.7 |
0.5 |
Would not use Aveeno |
0.8 |
0.5 |
0.7 |
0.3 |
0.8 |
Would purchase Aveeno |
0.8 |
0.9 |
0.1 |
0.8 |
0.4 |
Aveeno prevents dryness |
0.8 |
0.8 |
-0.2 |
0.9 |
0.1 |
Aveeno feels oily on skin |
0.7 |
0.3 |
0.8 |
0.0 |
0.8 |
Aveeno makes skin softer |
0.8 |
0.9 |
-0.2 |
0.9 |
0.1 |
Feel more attractive using Aveeno |
0.5 |
0.4 |
-0.5 |
0.6 |
-0.3 |
Trust Aveeno brand |
0.9 |
0.9 |
-0.2 |
0.9 |
0.1 |
Would recommend Aveeno |
0.9 |
0.9 |
0.0 |
0.9 |
0.4 |
In Table 3 the Factor Matrix and the Varimax Rotated Matrix for Aveeno both indicated ambiguous results for 1 Likert item and therefore the original Factor Matrix was used for this analysis. More Likert items loaded on Factor I which was used in the attitude score analysis.
Table 4: Olay
Brand Attribute (Likert item) |
Communalities |
Factor Matrix |
I |
||
Olay is a good brand |
0.8 |
0.9 |
Prefer Olay |
0.6 |
0.7 |
Would not use Olay |
0.6 |
0.8 |
Would purchase Olay |
0.7 |
0.8 |
Using Olay prevents dryness |
0.5 |
0.7 |
Olay feels oily on skin |
0.3 |
0.5 |
Olay makes skin softer |
0.7 |
0.8 |
Feel more attractive using Olay |
0.5 |
0.7 |
Trust Olay brand |
0.8 |
0.9 |
Would recommend Olay |
0.7 |
0.9 |
Table 4 indicates that only a Factor Matrix was generated and in this case all 10 Likert items loaded on Factor I. There were therefore no ambiguities here.
Table 5
Brand |
Mean Attitude Score |
Standard Deviation |
Sample Size |
Vaseline |
3.4 |
0.7 |
72 |
Aveeno |
3.5 |
0.8 |
72 |
Olay |
3.4 |
0.7 |
72 |
Table 5 indicates that out of the 3 brands, Aveeno had a slightly higher mean attitude score over Vaseline and Olay which indicates that respondents liked the Aveeno brand slightly more that Vaseline and Olay.
Table 6
Brand Pair |
t-ratio |
Aveeno & Vaseline |
0.93 |
Vaseline & Olay |
0.21 |
Aveeno & Olay |
1.29* |
*P<=.15
In Table 6, the results of the paired t-test demonstrate that in 85 or more samples out of 100 samples drawn from the same population as in this sample the mean attitude scores for the pair Aveeno and Olay would be about what they are here and can therefore be projected to the population.
A Cluster analysis was conducted to determine if respondents could be grouped into clusters based on their responses to the ten brand attitude Likert items for Aveeno. The clusters were also analyzed to determine if the clusters could be differentiated by those who are frequent and non-frequent users of body lotion.
Cluster analysis groups people or objects into clusters so that people or objects in the same cluster are more similar to each other than they are to people or objects in another cluster.
Table 1: Descriptive Statistics
Brand Attribute (Likert Item) |
Cluster 1 (n=43) |
Cluster 2 (n=33) |
||
Mean |
Standard Deviation |
Mean |
Standard Deviation |
|
Aveeno is a good brand |
4.3 |
0.5 |
3.3 |
0.9 |
Prefer Aveeno |
3.6 |
0.8 |
2.3 |
0.8 |
Would not use Aveeno |
4.2 |
0.5 |
3.1 |
0.9 |
Would purchase Aveeno |
4.2 |
0.5 |
2.6 |
0.9 |
Aveeno prevents dryness |
4.2 |
0.4 |
3.2 |
0.7 |
Aveeno feels oily on skin |
3.8 |
0.7 |
3.2 |
0.5 |
Aveeno makes skin softer |
4.1 |
0.4 |
3.0 |
0.6 |
Feel more attractive using Aveeno |
3.1 |
0.8 |
2.6 |
0.9 |
Trust Aveeno brand |
4.1 |
0.4 |
3.0 |
0.8 |
Would recommend Aveeno |
4.0 |
0.5 |
2.5 |
0.8 |
Table 1 displays higher mean scores for each of the 10 Likert items for Cluster 1 over Cluster 2. This indicates there is an overall higher brand attitude among respondents in Cluster 1.
Table 2: F-ratios
Brand Attribute (Likert item) |
F-ratio |
Aveeno is a good brand |
42.15* |
Prefer Aveeno |
45.01* |
Would not use Aveeno |
44.57* |
Would purchase Aveeno |
96.12* |
Aveeno prevents dryness |
63.93* |
Aveeno feels oily on skin |
16.15* |
Aveeno makes skin softer |
94.14* |
Feel more attractive using Aveeno |
4.52* |
Trust Aveeno brand |
60.58* |
Would recommend Av |