Welcome to this website which is a report on a brand preference study for 3 brands of facecare cleansing systems.

To learn more please use the navigation to the left, which also appears on every page, or move progressively through the report with the links at the bottom of each page.

Thank you for visiting!

 

Go to the Executive Summary »

This report is comprised of consumer’s facecare brand preferences after taking a survey that was based in copy-testing methods.

In general, copy-testing seeks to find out what sorts of advertisements affect people’s purchase decisions before companies run the ads. The more effective advertisers can make an ad ahead of time, the more secure companies can feel about efficiently spending their ad dollars.

In this case, the survey employed copy-testing to determine the following:

For this survey, respondents viewed print advertisements in 2007 magazines that can be found on most grocery store check-out aisles. The facecare brands for which they were surveyed include Proactive, Clinique and pHisoderm. The ads used can be viewed to the right or within the survey respondents took in Appendix B.

62 respondents completed a 15 minute online survey about brand preferences and their responses were run through a statistical computer program called SPSS. This program enabled the data to show many different analyses, seen to the left and in the analysis section of this report.

For more detailed information, please view the methodology, analyses, summary, conclusions and report with statistical frequencies.

Thanks again for visiting!

Go to the Methodology »

To gather data on Proactive, Clinique and pHisoderm, I created an online survey using Dreamweaver. The online format had a layout specific to industry standards for collecting data.

Due to time restraints, it was designed as a single page but appears to be many different pages due to hyperlinks between many anchor tags. Respondents were sent a link via email to answer questions.

After responding, the data was sent to a database where I pulled from it for analyses.

Please read more about the specific methodology in the following sections or skip to Appendix B to see the actual survey respondents took.

Go to the Structure of the Questionnaire »

This survey consisted of nine sections which can be described in 4 larger groups. They are described briefly below:

Sections 1 and 2 – These sections were the “pre-exposure” and looked at respondents likelihood to purchase brands before viewing ads.

Section 3 – Respondents viewed ads with copy enlarged for easy reading. Though respondents were encouraged to look at the ads only 15 seconds each and to not look back at ads during subsequent questions, a limitation of the study is that we could not control this.

Section 4-8  – This section included Lickert questions about respondent’s perception for each brand, yes/ no questions about the perception of each brand and yes/no questions about specific characteristics of each ad. This was the “post-exposure” section which was later compared to the “pre-exposure” to obtain change scores.

Section 9 – The final section consisted of several demographic questions and inquiries related to personal care of respondent’s own faces, facecare product buying habits and frequency and reasons for visiting a dermatologist.

Go to the Design »

This survey was primarily interested in the pre and post exposure responses and how the advertisements played a part in the differences of these groups.

There was no control group, only this pre vs. post differential, and the differences along the 21 point constant sum scale (10 to -10) were what helped classify respondents into an “up” group (moving up the scale), “down” group (moving down the scale) and a “same” group (no movement).

Go to the Sampling Description »

Time and monetary constraints limited this survey to a non-random convenience sample. 60 respondents were needed to create sufficient statistical significance and this survey actually has 62 respondents.

An email was sent to family and friends soliciting help and asking them to forward on the email to any others who might help me, creating a sort of snowball effect. Below is the letter that was sent.

Hi all!

I hope everyone is doing well!

I created the survey below for a huge research requirement for my graduate
program.

Will you please fill it out in the next few days? It should only take 10
minutes. I need a ton of responses to get a good sample size.

http://www.ciadvertising.org/sa/summer_07/adv380j/lauralouise/survey/test1.html

I really appreciate it! Next time I see you I'll give you a big hug!


Laura

P.S. - If I've missed anyone who you think could help, please forward this on!!

After clicking the link in the email I sent, respondents were directed to the survey. They filled in text fields and checked boxes, but mostly clicked radio buttons.

Once respondents reached the end of the survey, they clicked a submit button which sent the data to an Access database via ColdFusion. I was then able to download the Access database file to use the data in SPSS.

Because each field was coded with a value I could easily and numerically evaluate each respondent’s answers.

 

 

Go to the Analysis »

After collecting data, I ran several analyses via SPSS to better understand the implications it might hold for advertisers.

Check out the links below for each of the analyses or visit the Conclusions page to understand the implications for the brand's advertising.

Go to the Basic Statistics Analysis »

After collecting data on respondent’s face care brand preferences between brands Proactive, Clinique and pHisoderm, I ran several tests on the data to understand it more thoroughly.

Below you will find correlated (paired) t-tests on each brand index score, a between groups t-test on the pre and post exposure to the advertisements that were shown and a chi-square significance test to determine if a relationship exists between respondents those that moved up and down the brand index score. Additionally, I did several analyses of the frequency in the changes of the brand index scores in various types of comparisons. Finally, I calculated the simple correlation coefficient between two brands and then ran the same test while only looking at female respondents.

Paired samples t-test

Table 1. Paired Samples Statistics

Table 2. Paired Samples Test

*p<=.15

Between groups t-test

Table 3.Between groups t-test

*p<=.15

Chi-square significance test

Table 4.* Chi square significance test

*p<=.15

Pre and post brand index score comparison

Table 5. Proactive pre and post brand index score comparison

Proactive and Clinique brand index score comparison

Table 6. Proactive and Clinique brand index score comparison

Correlation Coefficient between 2 brands

Table 7. Simple correlation coefficient between Proactive and Clinique

*p<=.15

Correlation Coefficient between 2 brands for females

Table 8. Simple correlation coefficient between Proactive and Clinique for females

*p<=.15

Go to the Regression Analysis »

Table 1. Proactive linear regression model summary

*p<=.15

Table 2. Proactive linear regression coefficients

*p<=.15

Table 3. Clinique linear regression model summary

*p<=.15

Go to the Discriminant Analysis »

Below is a discriminant analysis for the facecare system Proactive between the likelihood of buying a brand before and after seeing an ad and the brand perception based on ten different questions asked of respondents.

Table 1. Group Statistics

*p<=.15

Table 2. Classification Results

	Predicted group Membership
	Positive responses for the brand on 10 Lickert items	Less positive responses for the brand on 10 Lickert items
Those whose perception for the brand increased after seeing ads	12	2
Those whose perception for the brand decreased after seeing ads	3	10

*p<=.15

 

 

 

Go to the ANOVA/MANOVA Analysis »

Below are is an Analysis of Variance (ANOVA) and a Multivariate Analysis of Variance (MANOVA). The ANOVA looks at only one Lickert item and the change score for the brand proactive and how that differs by gender, whereas MANOVA is an extension of the ANOVA and looks at the ten Lickert items for Proactive, the change score for the brand and how they all differ by gender.

Table 1. Analysis of Variance Table

Table 2. Between Subjects Effects

*p<=.15

Table 3. Multivariate Analysis of Variance