Profile Analysis

[Podcast~15 minutes]

Before you start your profile analysis, you must make sure that all the measures are on the same scale. Here are the questions were will answer using profile analysis:

1. Are the profiles parallel? (look for the interaction)
2. If the profiles are parallel, then are they coincident? (Did the groups score the same on each variable? look for the between group difference)
3. If the profiles are coincident, then are the profiles level? (Are the means on all variables equal to the same constant? look for the different on the within subjects factor)


Example from Stevens

In a study of love and marriage, a sample of husbands and wives were asked to respond to the following questions:

1. What is the level of passionate love you feel for your partner?
2. What is the level of passionate love that your partner feels for you?
3. What is the level of companionate love that you feel for your partner?
4. What is the level of companionate love that your partner feels for you?

The responses to all four question were on a Likert-type scale for 1 (none at all) to 5 (a tremendous amount). We wish to determine whether the profiles for the husbands and wives are parallel. There were 30 husbands and 30 wives that responded.

[Download Data]

SPSS Run

 

SPSS Output

 

 


Results

    A repeated measures ANOVA, with one within subjects factor (i.e., response to four items) and one between subjects factor (i.e., husbands and wives) was conducted to examine the responses of husbands and wives on the four survey items. All items were rated on the same scale, 1 to 5. The means and standard deviations across husbands and wives for the four survey items is reported in Table 1.

       The test of parallelism indicates that parallelism is tenable (F(3, 174)=.916 , p=.057). The profiles can be considered coincident (i.e., the same), F(1, 58)=1.711 , p=.196. This suggests that any differences in the differences for husbands and wives on the four items can be consider due to sampling error. The test of equal means across the four survey items indicate that there was a differences between the means, F(3, 174)=15.098 , p<.001. Post hoc analyses using Bonferroni inequality indicated that items 1 and 2 were significantly higher than items 3 and 4.