Data transformation from survey ratings to variables

[Laurence Meeuwis]

Hi!
I am researching the effect of cosmetic combinations on the perception of leadership characteristics. I collected data on 24 images: 6 treatment versions of 4 images. The data that I collected consists of ratings of each photo on 6 leadership characteristics, hence 6 dependent variables. I, therefore, collected 24x6=144 rating values, each rating being given by approximately 25 respondents. The 6 treatment versions consist of 2 independent variables: brow enhancement (yes or no) and lipstick colour (none, pink, red).

Since I basically collected 144 ratings, Stata has imported the data as 144 variables.

How can I transform the data such that I obtain 6 dependent variables and 2 independent variables from the 144 var's?

I think the problem with my data currently is that the variables pertain to the images rather than to the subject, which would be the case for e.g. control variables like gender, age etc.

Rather than having the current variables in Stata listed as separate variables, I need to ensure that I group the data such that I obtain:

Dependent variables:
Intelligence
Dominance
Honesty&integrity
Self-confidence
Leadership effectiveness
Attractiveness

Independent variables:
Brows
Lipstick colour

NB: the data from the four images can be aggregated.


Hope someone can help!

[Mike Lacy]

There's a lot about your data that we can't know without seeing it, and so can't effectively help you. I'd encourage you to take a look at the FAQ about how to use the -dataex- command to show an example of your data here, and then repost. I know that's going to produce an unwieldy wide file, but unless one has a lot of experience with data analysis, it's almost impossible to describe a data structure in words, so I think that's the best alternative we have in this case. Also, if you can create an example (simulated is fine) that shows how you want your data to end up, that would help as well. Finally, if we knew the precise data analysis you wanted to do, that would help us in leading you to an optimal data structure.

[Laurence Meeuwis]

Thank you! I have posted my updated question here: https://www.statalist.org/forums/for...survey-answers