Consumer Preference Survey:
The Story of Three Pasta Sauces.
Table of Contents
Frequency Table: Change Scores
Frequency Table: Brand Preference
Paired t-test of Female Subjects
Brand Index Scores for each brand were achieved by adding the 10 corresponding coded Likert items and then obtaining the mean from 72 online respondents. The lower the score, the more favorably the brand was viewed. Correlated t-tests were then conducted comparing all three Brand Index Scores in all possible pairings. The purpose of the paired t-test was to find if the there was a significant difference between the Brand Index Scores of the pairings.
Brand Index Scores
Pasta Sauce Brand |
Sample Size |
Mean |
Standard Deviation |
Ragu |
72 |
28.0 |
7.9 |
Bertolli |
72 |
27.6 |
4.4 |
Newman’s Own |
72 |
18.9 |
5.6 |
Paired Sample t-test
Paired Samples |
t |
Ragu vs. Bertolli |
0.26 |
Ragu vs. Newman’s Own |
6.38* |
Bertolli vs. Newman’s Own |
10.58* |
* p≤ .15
The ‘moved up’ and ‘moved down’ scores were achieved by subtracting the pre-test evaluation from the post-test evaluation of each brand. Respondents with a positive score viewed the brand more favorably after viewing the ad. The Ad Index Score was achieved by adding all the corresponding positively worded questions (excluding Likert questions) asked of the brand after seeing the ad. The purpose of the between groups t-test was to find if there was a significant difference between the Brand Index Scores of up and down movers as well as the Ad Index Scores of these same subjects.
Brand Index Score change for Newman’s Own pasta sauce
Change Score |
Sample Size |
Mean |
Standard Deviation |
Moved Up |
39 |
17.7 |
5.3 |
Moved Down |
33 |
20.3 |
5.6 |
Ad Index Score Change for Newman’s Own pasta sauce
Change Score |
Sample Size |
Mean |
Standard Deviation |
Moved Up |
39 |
11.4 |
2.0 |
Moved Down |
33 |
7.3 |
2.7 |
Independent Samples t-test
Variable |
t |
Brand Index Score |
2.04* |
Ad Index Score |
7.13* |
* p≤ .15
A chi-squared significance test was run to determine if there was a significant relationship between up movers, down movers, and those that had the same score pre- and post-test and being above or below the median Brand Index Score for Newman’s Own pasta sauce. In this case, the median was 21 and the frequencies for each category are shown above.
Crosstabulation for Newman’s Own pasta sauce
|
Above Median |
Below Median |
|
Up |
Count |
22 |
11 |
% of Row |
66.7% |
33.3% |
|
% of Column |
62.9% |
42.3% |
|
% of Total |
36.1% |
18.0% |
|
Same |
Count |
13 |
11 |
% of Row |
54.2% |
45.8% |
|
% of Column |
37.1% |
42.3% |
|
% of Total |
21.3% |
18.0% |
|
Down |
Count |
0 |
4 |
% of Row |
0% |
42.6% |
|
% of Column |
0% |
100.0% |
|
% of Total |
0% |
42.6% |
|
Chi-squared for Newman’s Own pasta sauce
|
Chi-squared |
df |
Pearson Chi-Square |
6.65* |
2 |
* p≤ .15
Frequency Table: Change Scores
This is a frequency table showing simply how many of the 72 respondents moved up or down or stayed the same in their pre- to post-test scores. These are purely descriptive statistics of the sample.
Change Score |
Ragu |
Bertolli |
Newman’s Own |
Moved Up |
1 |
17 |
39 |
Same |
28 |
32 |
29 |
Moved Down |
43 |
23 |
4 |
Frequency Table: Brand Preference
This is a frequency table showing how many of the 72 online respondents preferred Newman’s Own over Ragu or vice-versa according to each brand’s particular Brand Index Score for each respondent. In this particular comparison, there were no respondents who liked both equally, so the results are split into 53 liking Newman’s Own more and 19 liking Ragu more. These are purely descriptive statistics of the sample.
|
Frequency |
Likes Newman’s Own more than Ragu |
53 |
Likes Newman’s Own less than Ragu |
19 |
This is a correlation test to determine the strength of the relationship between the Brand Index Scores of Ragu and Bertolli.
Correlation between Ragu and Bertolli Brand Index Scores
|
Correlation Coefficient |
Ragu and Bertolli Brand Index Scores |
-0.6 |
Paired t-test of Female Subjects
The original sample of 72 respondents altered to comprise of only female respondents. The Brand Index Scores for Ragu and Newman’s own of this new pool of 35 females was analyzed to determine if there was a difference in the ways the brands were viewed.
Pasta Sauce Brand |
Sample Size |
Mean |
Standard Deviation |
Ragu |
35 |
27.5 |
7.2 |
Newman’s Own |
35 |
17.5 |
4.9 |
Paired Sample t-test
Paired Samples |
t |
Ragu vs. Newman’s Own |
5.01* |
* p≤ .15
A regression analysis was run for each of the pasta sauce brands to determine the relationship between certain independent variables and a dependent variable. In this case, the independent variables were the respective Likert questions and the dependent variable was the move score coded as ‘up’, ‘same’, or ‘down’. It is useful to see which of the independent variables can best explain the difference in the dependent variable.
RAGU
R: Multiple correlation coefficient |
R-squared: coefficient of multiple determination |
Standard error of the estimate |
F ratio to test multiple R |
0.8 |
58.5% |
1.5 |
8.61* |
* p≤ .15
Brand attributes |
Unstandardized coefficients (b) |
Standardized coefficients (beta) |
t |
Constant |
-3.8 |
|
2.63* |
Ragu is a good pasta sauce |
-1.6 |
-.7 |
1.81* |
Ragu tastes good |
-0.2 |
-.1 |
0.31 |
Ragu is a good value |
1.5 |
0.6 |
1.77* |
Ragu is not a healthy pasta sauce |
0.5 |
0.2 |
0.95 |
Ragu is a brand that I trust |
1.2 |
0.4 |
1.67* |
I would not recommend Ragu to others |
-0.6 |
-0.3 |
1.19 |
Ragu is a high quality pasta sauce |
1.7 |
0.8 |
4.39* |
Ragu is an authentic tasting pasta sauce |
-1.0 |
-0.5 |
3.04* |
Ragu is a pasta sauce I would be loyal to without switching |
-1.0 |
-0.4 |
3.13* |
I would not buy Ragu |
0.3 |
0.2 |
0.82 |
* p≤ .15
The independent variables in this test were the Likert items in the brand survey testing perception of various attributes of the brand; the dependent variable was the move score, or the increase or decrease in likelihood of buying the brand after seeing its ad. The coefficient of multiple determinants for Ragu is 58.5%. This is a relatively middling score indicating some connection between the perception of the brand and the effectiveness of the ad. The standard error of the estimate is 1.5, which is also middling, meaning that the each respondent is an average of 1.5 points away from the regression line, that line being the line which most accurately describes the scatterplot of respondent answers to the Likert items.
As to the Likert items and constant, those indicated with an asterisk are significant predictors of the change score. Those highlighted in red are the most important in determining whether the change score was ‘up’, down’, or ‘same’.
BERTOLLI
R: Multiple correlation coefficient |
R-squared: coefficient of multiple determination |
Standard error of the estimate |
F ratio to test multiple R |
0.7 |
50.7% |
0.8 |
7.09* |
* p≤ .15
Brand attributes |
Unstandardized coefficients (b) |
Standardized coefficients (beta) |
t |
Constant |
0.6 |
|
0.85 |
Bertolli is a good pasta sauce |
-0.3 |
-.2 |
0.48 |
Bertolli tastes good |
0.4 |
0.2 |
0.96 |
Bertolli is a good value |
0.8 |
0.4 |
0.76 |
Bertolli is not a healthy pasta sauce |
-0.2 |
-0.1 |
0.41 |
Bertolli is a brand that I trust |
0.2 |
0.1 |
0.35 |
Bertolli is a high quality pasta sauce |
-1.5 |
-0.8 |
2.36* |
Bertolli is an authentic tasting pasta sauce |
1.2 |
0.6 |
1.61* |
Bertolli is a pasta sauce I would be loyal to without switching |
-0.4 |
-0.2 |
0.89 |
I would not buy Bertolli |
-0.8 |
-0.5 |
1.73* |
* p≤ .15
The independent variables in this test were the Likert items in the brand survey testing perception of various attributes of the brand; the dependent variable was the move score, or the increase or decrease in likelihood of buying the brand after seeing its ad. The coefficient of multiple determinants for Bertolli is 50.7%. This is a neither strong nor weak score indicating there exists some connection between the perception of the brand and the effectiveness of the ad. The standard error of the estimate is 0.8 meaning that the each respondent is an average of 0.8 points away from the regression line, that line being the line which most accurately describes the scatterplot of respondent answers to the Likert items.
As to the Likert items and constant, those indicated with an asterisk are significant predictors of the change score at the p≤ .15 level.Those highlighted in red are the most important in determining whether the change score was ‘up’, down’, or ‘same’.
NEWMAN’S OWN
R: Multiple correlation coefficient |
R-squared: coefficient of multiple determination |
Standard error of the estimate |
F ratio to test multiple R |
0.7 |
46.5% |
2.0 |
5.30* |
* p≤ .15
Brand attributes |
Unstandardized coefficients (b) |
Standardized coefficients (beta) |
t |
Constant |
0.4 |
|
0.32 |
Newman’s Own is a good pasta sauce |
-1.1 |
-.2 |
-0.81 |
Newman’s Own tastes good |
11.7 |
3.4 |
3.87* |
Newman’s Own is a good value |
4.3 |
1.6 |
5.61* |
Newman’s Own is not a healthy pasta sauce |
10.3 |
2.0 |
5.05 |
Newman’s Own is a brand that I trust |
-2.0 |
-0.4 |
1.07 |
I would not recommend Newman’s Own to others |
-10.5 |
-3.0 |
5.70* |
Newman’s Own is a high quality pasta sauce |
0.4 |
0.1 |
0.22 |
Newman’s Own is an authentic tasting pasta sauce |
-14.1 |
-4.3 |
4.62* |
Newman’s Own is a pasta sauce I would be loyal to without switching |
-1.0 |
0.0 |
0.15 |
I would not buy Newman’s Own |
2.1 |
0.46 |
1.03 |
* p≤ .15
The independent variables in this test were the Likert items in the brand survey testing perception of various attributes of the brand; the dependent variable was the move score, or the increase or decrease in likelihood of buying the brand after seeing its ad. The coefficient of multiple determinants for Newman’s Own is 46.5%. This is a neither strong nor weak score indicating there exists some connection between the perception of the brand and the effectiveness of the ad. The standard error of the estimate is 2.0 which is quite high, meaning that the each respondent is an average of 2.0 points away from the regression line, that line being the line which most accurately describes the scatterplot of respondent answers to the Likert items.
As to the Likert items and constant, those indicated with an asterisk are significant predictors of the change score at the p≤ .15 level. Those highlighted in red are the most important in determining whether the change score was ‘up’, down’, or ‘same’.
A discriminant analysis is a test performed to examine how well independent variables classify members of a sample into different groups. It also tells us which independent variables are the most important in achieving that distinction. In this case, the independent variables are all the Likert questions from the survey regarding Newman’s Own and the dependent variables are the move scores: ‘up’, ‘down’, or ‘same’.
Up-Movers
Likert Item |
Mean |
Standard Deviation |
Newman's Own is a good pasta sauce. |
1.6 |
0.5 |
Newman's Own tastes good. |
1.9 |
0.8 |
Newman's Own is a good value. |
2.2 |
0.8 |
Newman's Own is a not healthy pasta sauce. |
1.5 |
0.5 |
Newman's Own is a brand that I trust. |
1.5 |
0.6 |
I would not recommend Newman's Own to others. |
1.5 |
0.6 |
Newman's Own is a high quality pasta sauce. |
1.5 |
0.5 |
Newman's Own is an authentic tasting pasta sauce. |
2.0 |
0.8 |
Newman's Own is a pasta sauce that I would be loyal to without switching. |
2.6 |
0.7 |
I would not buy Newman's Own. |
1.5 |
0.6 |
Down Movers
Likert Item |
Mean |
Standard Deviation |
Newman's Own is a good pasta sauce. |
1.73 |
0.5 |
Newman's Own tastes good. |
2.1 |
0.7 |
Newman's Own is a good value. |
2.4 |
1.1 |
Newman's Own is a not healthy pasta sauce. |
1.7 |
0.5 |
Newman's Own is a brand that I trust. |
1.9 |
0.4 |
I would not recommend Newman's Own to others. |
2.0 |
0.8 |
Newman's Own is a high quality pasta sauce. |
1.6 |
0.5 |
Newman's Own is an authentic tasting pasta sauce. |
2.1 |
0.7 |
Newman's Own is a pasta sauce that I would be loyal to without switching. |
3.2 |
1.4 |
I would not buy Newman's Own. |
1.6 |
0.6 |
Standard Discriminant Function Coefficients and Discriminant Function Coefficient
Likert Item |
Standard Discriminant Function Coefficient |
Discriminant Function Coefficient |
Newman's Own is a good pasta sauce. |
-1.5 |
-3.1 |
Newman's Own tastes good. |
-4.6 |
-6.3 |
Newman's Own is a good value. |
-2.3 |
-2.4 |
Newman's Own is a not healthy pasta sauce. |
-3.6 |
-7.5 |
Newman's Own is a brand that I trust. |
2.1 |
4.1 |
I would not recommend Newman's Own to others. |
6.2 |
9.5 |
Newman's Own is a high quality pasta sauce. |
1.2 |
2.4 |
Newman's Own is an authentic tasting pasta sauce. |
6.6 |
8.7 |
Newman's Own is a pasta sauce that I would be loyal to without switching. |
-0.7 |
-0.7 |
I would not buy Newman's Own. |
-3.2 |
-5.8 |
Wilks’ Lambda
Wilks’ Lambda |
Chi-squared |
0.36 |
66.17* |
* p≤ .15
Group Centroids
Change Score |
Group Centroids |
Up-Movers |
-1.2 |
Down-Movers |
1.4 |
Confusion Matrix
|
Predicted Up-Movers |
Predicted Down-Movers |
Total |
t |
Actual Up-Movers |
37 |
2 |
39 |
5.89* |
Actual Down-Movers |
5 |
28 |
33 |
*at α=.15, tc=1.04, t achieved by: _________0.903-0.5________
√ [(0.903)(0.097) + (0.5)(0.5)]
72 72
Of the 72 respondents to the survey, 39 were classified as up-movers and 33 as down-movers based on the difference of their pre and post-test scores of the Newman’s Own pasta sauce ad. These respondents were also asked 10 Likert questions regarding the perceived attributes of the product with the same 5 point scale applied to each question (1 being ‘strongly agree’ and 5 being ‘strongly disagree’). The means and standard deviations for each question for both groups are reported above. A discriminant analysis was performed to see if the up-movers and down-movers could be separated into statistically significant groups and to see which Likert questions were most influential in the division. The Chi-squared figure shown is an indirect way of testing Wilk’s Lambda and it was found to be significant. This means that the group centroids of each group, the ‘center’ of each group as projected onto the discriminant function can be projected to the population.
Also, the most important Standard Discriminant Function Coefficients for the Likert questions in determining group affiliation from which the group centroids are drawn are marked in red. The confusion matrix has successful group prediction rate of 90.3%. That is to say of the respondents that took the test, 90.3% were correctly predicted to be either up-movers or down-movers.
An ANOVA test is used to determine the individual effects of two or more discretely measured independent variables on a continuously measured dependent variable as well as the combined effect of the independent variables. Here the independent variables are gender (male and female) and the move score (whether a respondent’s initial perception of the brand (Newman’s Own) increased, decreased, or stayed the same after viewing its ad). The dependent variable is ‘Newman’s Own is a good pasta sauce’.
Move Score |
||||
|
|
Up |
Same |
Down |
Gender |
Male |
Mean= 1.6 |
Mean= 1.7 |
Mean= 2.0 |
Female |
Mean= 1.6 |
Mean= 1.7 |
Mean= 0 |
|
|
Sum of Squares |
Deg. of freedom |
Mean Square |
F-ratio |
Source |
||||
Gender |
0.01 |
1 |
0.01 |
0.03 |
Move Score |
0.65 |
2 |
0.32 |
1.38 |
Interaction effect |
0.00 |
1 |
0.00 |
0.00 |
Error |
15.63 |
67 |
0.23 |
|
* p≤ .15
The results of the ANOVA are shown above. The participants numbered 72 split into six different categories, but there are no female ‘down’ movers. The dependent variable was a Likert question measured on a 1 to 5 scale with one being ‘strongly agree’ and 5 ‘strongly disagree’. The mean scores and standard deviations of the male ‘up’ and ‘same’ movers are identical to their female counterparts.
MANOVA
A MANOVA test is identical in computation and purpose to an ANOVA except there are multiple dependent variables. In this case, the independent variables are the same as in the above ANOVA, but the dependent variables consist of all the Likert questions from the survey.
Move Score |
||||
Likert Item** |
Gender |
Up |
Same |
Down |
__good |
Male |
Mean = 1.6 |
Mean = 1.7 |
Mean = 2.0 |
Female |
Mean = 1.6 |
Mean = 1.7 |
Mean =0.0 |
|
__taste |
Male |
Mean = 1.9 |
Mean = 2.1 |
Mean = 3.0 |
Female |
Mean = 2.0 |
Mean = 1.7 |
Mean = 0.0 |
|
__value |
Male |
Mean = 2.1 |
Mean = 2.8 |
Mean = 2.0 |
Female |
Mean = 2.2 |
Mean = 1.7 |
Mean = 0.0 |
|
__healthy |
Male |
Mean = 1.4 |
Mean = 1.7 |
Mean = 2.0 |
Female |
Mean = 1.6 |
Mean = 1.7 |
Mean = 0.0 |
|
__trust |
Male |
Mean = 1.4 |
Mean = 1.9 |
Mean = 2.0 |
Female |
Mean = 1.5 |
Mean = 1.7 |
Mean = 0.0 |
|
__recom |
Male |
Mean = 1.4 |
Mean = 2.2 |
Mean = 2.0 |
Female |
Mean = 1.5 |
Mean = 1.7 |
Mean = 0.0 |
|
__qual |
Male |
Mean =1.4 |
Mean = 1.8 |
Mean = 2.0 |
Female |
Mean = 1.6 |
Mean = 1.0 |
Mean = 0.0 |
|
__auth |
Male |
Mean = 1.9 |
Mean = 2.1 |
Mean = 3.0 |
Female |
Mean = 2.1 |
Mean = 1.7 |
Mean = 0.0 |
|
__loyal |
Male |
Mean = 2.4 |
Mean = 3.9 |
Mean = 3.0 |
Female |
Mean = 2.7 |
Mean = 2.1 |
Mean = 0.0 |
|
__buy |
Male |
Mean = 1.4 |
Mean = 1.9 |
Mean = 2.0 |
Female |
Mean = 1.5 |
Mean = 1.0 |
Mean = 0.0 |
|
** __ good: (Brand of sauce)is a good pasta sauce.
__taste: (Brand of sauce) tastes good.
__value: (Brand of sauce) is a good value.
__healthy: (Brand of sauce) is anot healthy pasta sauce.
__trust: (Brand of sauce) is a brand that I trust.
__recom: I would not recommend (Brand of sauce) to others
__qual: (Brand of sauce) is a high quality pasta sauce.
__auth: (Brand of sauce) is an authentic tasting pasta sauce.
__loyal: (Brand of sauce) is a pasta sauce that I would be loyal to without switching.
__buy: I would not buy (Brand of sauce).
Inferential Statistics
Source |
Wilks’ Lambda |
F-ratio |
Hypothesis df |
Error df |
Gender |
0.06 |
89.44* |
10 |
58 |
Move Score |
0.06 |
18.98* |
20 |
116 |
Interaction effect |
0.09 |
59.28* |
10 |
58 |
* p≤ .15
The results are split into the same categories as the preceding ANOVA, but for all 10 Likert dependent variables. The descriptive statistics are shown above for all 72 respondents and are purely illustrative of the sample.
A factor analysis is a test conducted to summarize variables into fewer representative factors. That is the variables used to determine any sort of conclusion can be grouped into common factors, each of which is descriptive of those constituent variables. After this is done, all the variables which load onto the same factor as the first Likert item ( ‘__ is a good pasta sauce’) can be by default classified as evaluative variables. Taking the mean score of these variables will then give us an attitude scale in which variables superfluous to the overall judgment of the brand as ‘good’ can be eliminated.
Communalities |
|||
|
RAGU |
BERTOLLI |
NEWMAN’S OWN |
Likert Item** |
h-squared |
h-squared |
h-squared |
__good |
0.8 |
0.6 |
0.6 |
__taste |
0.8 |
0.7 |
1.0 |
__value |
0.8 |
0.9 |
0.7 |
__healthy |
0.9 |
0.8 |
0.7 |
__trust |
0.8 |
0.7 |
0.8 |
__recom |
0.9 |
0.9 |
0.9 |
__qual |
0.6 |
0.7 |
0.7 |
__auth |
0.7 |
0.9 |
0.9 |
__loyal |
0.4 |
0.9 |
0.7 |
__buy |
0.6 |
0.7 |
0.8 |
Total Variance Explained |
|||||||||
|
Ragu |
Bertolli |
Newman’s Own |
||||||
Likert Item** |
Eigenvalue |
% of variance |
Cumulative% |
Eigenvalue |
% of variance |
Cumulative% |
Eigenvalue |
% of variance |
Cumulative% |
__good |
7.3 |
72.7 |
72.7 |
6.7 |
66.9 |
78.7 |
6.8 |
68.1 |
79.7 |
__taste |
0.7 |
7.0 |
|
1.2 |
11.8 |
|
1.2 |
11.6 |
|
__value |
0.7 |
6.5 |
|
0.8 |
7.6 |
|
0.7 |
6.9 |
|
__healthy |
0.6 |
5.6 |
|
0.7 |
7.0 |
|
0.6 |
5.7 |
|
__trust |
0.3 |
2.6 |
|
0.3 |
3.1 |
|
0.3 |
3.4 |
|
__recom |
0.2 |
2.4 |
|
0.2 |
1.7 |
|
0.2 |
2.2 |
|
__qual |
0.2 |
1.6 |
|
0.1 |
1.0 |
|
0.1 |
1.4 |
|
__auth |
0.1 |
1.0 |
|
0.1 |
0.7 |
|
0.0 |
0.4 |
|
__loyal |
0.0 |
0.3 |
|
0.0 |
0.3 |
|
0.0 |
0.2 |
|
__buy |
0.0 |
0.3 |
|
0.0 |
0.0 |
|
0.0 |
0.0 |
|
Ragu Factor Matrix |
|
|
Factor |
Likert Item** |
I |
__good |
0.9 |
__taste |
0.9 |
__value |
0.9 |
__healthy |
0.9 |
__trust |
0.9 |
__recom |
0.9 |
__qual |
0.8 |
__auth |
0.8 |
__loyal |
0.7 |
__buy |
0.8 |
Bertolli Factor Matrix |
||
|
Factor |
|
Likert Item** |
I |
II |
__good |
0.7 |
-0.3 |
__taste |
0.8 |
-0.1 |
__value |
0.9 |
0.4 |
__healthy |
0.9 |
-0.1 |
__trust |
0.8 |
0.2 |
__recom |
0.9 |
0.0 |
__qual |
0.8 |
-0.2 |
__auth |
0.9 |
-0.1 |
__loyal |
0.4 |
0.9 |
__buy |
0.8 |
-0.3 |
Bertolli Varimax Rotated Matrix |
||
|
Factor |
|
Likert Item** |
I |
II |
__good |
0.8 |
0.0 |
__taste |
0.8 |
0.2 |
__value |
0.6 |
0.8 |
__healthy |
0.9 |
0.3 |
__trust |
0.7 |
0.5 |
__recom |
0.8 |
0.4 |
__qual |
0.8 |
0.2 |
__auth |
0.9 |
0.3 |
__loyal |
0.0 |
1.0 |
__buy |
0.8 |
0.1 |
Newman’s Own Varimax Rotated Matrix |
||
|
Factor |
|
Likert Item** |
I |
II |
__good |
0.6 |
0.5 |
__taste |
0.2 |
1.0 |
__value |
0.8 |
0.2 |
__healthy |
0.7 |
0.5 |
__trust |
0.7 |
0.5 |
__recom |
0.9 |
0.2 |
__qual |
0.6 |
0.7 |
__auth |
0.2 |
0.9 |
__loyal |
0.8 |
0.2 |
__buy |
0.7 |
0.6 |
Newman’s Own Factor Matrix |
||
|
Factor |
|
Likert Item** |
I |
II |
__good |
0.8 |
0.0 |
__taste |
0.8 |
0.6 |
__value |
0.8 |
-0.3 |
__healthy |
0.9 |
-0.1 |
__trust |
0.9 |
-0.1 |
__recom |
0.9 |
-0.4 |
__qual |
0.8 |
0.1 |
__auth |
0.8 |
0.6 |
__loyal |
0.8 |
-0.4 |
__buy |
0.9 |
0.0 |
Attitude Scale (descriptive statistics) |
|||
Brand |
Mean |
Standard Deviation |
Sample Size |
Ragu |
2.8 |
0.8 |
72 |
Bertolli |
2.5 |
0.4 |
72 |
Newman’s Own |
1.9 |
0.1 |
72 |
Attitude Scale (inferential statistics) |
|
Paired Samples |
t |
Ragu vs. Bertolli |
2.68* |
Ragu vs. Newman’s Own |
6.38* |
Bertolli vs. Newman’s Own |
7.12* |
* p≤ .15
** __ good: (Brand of sauce)is a good pasta sauce.
__taste: (Brand of sauce) tastes good.
__value: (Brand of sauce) is a good value.
__healthy: (Brand of sauce) is anot healthy pasta sauce.
__trust: (Brand of sauce) is a brand that I trust.
__recom: I would not recommend (Brand of sauce) to others
__qual: (Brand of sauce) is a high quality pasta sauce.
__auth: (Brand of sauce) is an authentic tasting pasta sauce.
__loyal: (Brand of sauce) is a pasta sauce that I would be loyal to without switching.
__buy: I would not buy (Brand of sauce).
All statistics derived directly from a factor analysis are descriptive; no generalizations can be made to the population. The first table of communalities is an indication of the percentage of variance in the original variable that is explained entirely by the factors achieved from the factor analysis. The Eigenvalues are the sums of squared factor loadings for a given factor (which is the percentage of variance in the original variable explained by a factor). When divided by the number of variables, this gives the percentage of variance accounted for by a factor. Important Eigenvalues are highlighted; they are more than 1 because, if not, the factor would explain less than any one of the variables individually. The cumulative percentage explains the amount of variance explained by a factor made up of the most important variables. The next tables show how the variables (Likert items) loaded onto the factors. Those with values above 0.5 are highlighted as important loading variables. Since Ragu had only one factor, there was no varimax rotation. This is a process which objectively places variables on the loading scheme to try and achieve a more accurate matrix. In this case, the two varimax rotations were worse than the unrotated (for Bertolli and Newman’ Own) so the original component matrices were used though Newman’s Own has two variables which load on both factors. In this test, Bertolli was the only successful factor analysis to break the variables into two factors; Ragu had all ten Likerts load on one factor which is not reductive and Newman’s Own had overlapping variables. The attitude scale was constructed to achieve inferential statistics and is discussed in the ‘Conclusions’ section.