Basic statistics

Paired t-tests

It seems then that Advil is the most positively-perceived brand of the three, in both the respondent sample and in the population from which the sample was drawn. Tylenol is the next most positively-perceived brand, followed by Bayer.

Independent t-tests

In this sample, up movers had a higher perception of Advil’s advertising, but a lower perception of Advil’s brand, than down movers. The higher perception of Advil’s advertising can be expected to hold true in the larger population, but the lower perception of Advil’s brand cannot.

Chi-squared test

For both up movers and down movers, there was only a slight difference between the number of respondents above and below the median brand index score for Advil. This difference cannot be projected to the population at large.

Frequency count

By a wide margin, Bayer had both the largest number of up movers (50) and the smallest number of down movers (3). Advil had the smallest number of up movers (19) and the largest number of down movers (45). Tylenol had the smallest number of respondents who stayed the same.

Comparison of brand perceptions

Sixty-eight respondents had a higher perception of the Advil brand than the Tylenol brand, as measured by the respective brands’ brand index scores. Fifty-eight respondents had a higher perception of Tylenol than Advil.

Correlation

As a respondent’s perception of the Advil brand goes up, his perception of the Tylenol brand tends to go up as well – but this tendency is very slight. This is the case both for the respondent sample and for the population at large.

Paired t-tests for heavy users

For heavy users as for all respondents, Advil is the most positively-perceived brand, Tylenol is the next most positively-perceived brand and Bayer is the least positively-perceived brand. This is true both for heavy users in the respondent sample and in the larger population.

Multiple regression analysis

These results suggest that Advil’s brand perceptions played a very small role in the variations between respondents’ changed purchase intentions after exposure to the advertisements. Only three of the brand attributes tested – Advil is a good brand, Advil doesn’t have too many side effects and Advil is made by a company that cares about my health – could be considered important.

Bayer’s brand perceptions seem to have played a vanishingly small role in explaining the variations between respondents’ changed purchase intentions after exposure to the advertisements. Only two of the brand attributes tested – Bayer is a good brand and Bayer works fast – could be considered important.

Tylenol’s brand perceptions seem to have played a very small role in explaining the variations between respondents’ changed purchase intentions after exposure to the advertisements. Only two of the brand attributes tested – Tylenol is easy to use and Tylenol doesn’t have too many side effects – could be considered important.

Discriminant analysis

In the discriminant analysis, none of the mean Likert scores for each attribute varies more than 0.2 points between up-movers and down-movers – thus there appear to be few if any great differences between the two groups’ perceptions of the Tylenol brand. "Tylenol is safer than other over-the-counter pain relief medications," "Tylenol would be prescribed by my doctor," "Tylenol is easy to use," "Tylenol is made by a company that cares about my health" and "Tylenol provides effective pain relief" appear to be the most important brand attributes in determining whether respondents were up movers or down movers.

These results suggest that, given a respondent’s brand attitude Likert scores, we can predict whether the respondent is an up-mover or a down-mover with only 63.2% accuracy. But the average discriminant scores are not statistically significant and cannot be projected to the population.

Analysis of variance

In this sample, heavy users are more likely to agree that "Tylenol provides effective pain relief" than non-heavy users. This can be projected to the population. But there doesn’t seem to be much difference between up-, down- or non-movers, nor when heavy usage and mover group are combined.

In this sample, heavy users are more likely to agree that "Tylenol would be prescribed by my doctor," "Tylenol is made by a company that cares about my health" and "Tylenol provides effective pain relief." They are also more likely to believe – wrongly – that "Tylenol prevents heart attacks." But the results of this test cannot be projected to the population.

Factor analysis

For Advil, “Advil is a good brand,” “Advil works fast,” “Advil is trustworthy,” “Advil is easy to use,” “Advil doesn’t have too many side effects” and “Advil provides effective pain relief” all load on factor I. “Advil would be prescribed by my doctor” loads ambiguously. The remainder of the Likert items load on factor II.

For Bayer, “Bayer is a good brand, “Bayer works fast,” “Bayer prevents heart attacks,” “Bayer would be prescribed by my doctor” and “Bayer provides effective pain relief” all load on Factor I. “Bayer is trustworthy,” “Bayer is easy to use” and “Bayer doesn’t have too many side effects” load on Factor II. “Bayer is safer” and “Bayer is made by a company that cares about my health” load on Factor III.

For Tylenol, “Tylenol is a good brand,” “Tylenol works fast,” “Tylenol is trustworthy,” and “Tylenol provides effective pain relief” all load on Factor I. “Tylenol prevents heart attacks” and “Tylenol is easy to use” load on Factor II. “Tylenol would be prescribed by my doctor” and “Tylenol is made by a company that cares about my health” load on Factor III. “Tylenol doesn’t have too many side effects” loads ambiguously on Factors I or II.

Advil has the highest mean attitude score with 4.0. Tylenol has a mean attitude score of 3.9. Bayer has the lowest mean attitude score with 3.4. This is true both for the sample and for the population.

Cluster analysis

Cluster 2 has a higher perception of Tylenol on all of the brand Likert items except "Tylenol is made by a company that cares about my health." The members of cluster 1 are much more likely to be male than female, while the members of cluster 2 are much more likely to be female than male. This is true both for the sample and for the population.