Mate Preference

[Star7]

Hi guys, I'm a psychology student and my recent report is on mate preference.

Basically the lecturer gathered data by asking everyone to write on a piece of paper our preference for a potential mate. Our own gender is indicated.

So what I have in the end (provided by the lecturer) is a huge list of words describing characteristics of a potential mate starting from a to z.

E.g. Cook, Charming, Dance, Elegant, Funny etc.

Unlike previous essays or reports, the raw data are not numbers. That's where and why I'm lost.

As far as I'm concern there's really no IV nor DV, just a collection of preferences and I was thinking descriptives instead of inferential statistics.

However, even with descriptives, I've no idea how to go about doing this when the raw data are all words. I've thought about attaching a value to each label (word) but I'm still pretty confused.

All help appreciated.

Oh and I'm using SPSS.

[mirkamali]

I don't know much about psychology, but I think that you can assign some codes to their responses and use the number of each codes to make some inferences.

You can compare frequencies between subgroups (divided by Gender). If you had more variables you could classify responses, and find some correlations.

Also, You can join some resposes and derive some indices.

[bugman]

As mirkamali suggested you coul apply codes for the particualar preference. To simplify the analysis you could code several preferences (eg.funny, outgoing, chraming etc...) into one (i.e. personnality) or cook, clean etc.. into "domestic" for example. you can then count these responses and classify them by gender and then construct a contingency table test for an association between gender and attributes, say.

Phil

[Star7]

I've thought about what both of you had said and I've an article that might provide some help.

I've found an article whereby factor analysis was conducted to reveal the underlying structure of 76 preferences.

The analysis yield 9 factors: Kind-Considerate, Socially Exciting, Artistic-Intelligent, Religious, Domestic, Professional Status, Likes Children, Politically Conservative and Easygoing-adaptable.

What I have in mind is to code each of those labels into values, 1-9. Gender would be 1-2.

I'll proceed to enter the gender first, followed by categorizing the respective preference into one of those nine categories.

Something like
Gender: 1 (female)
Preferences: 5 (domestic - can cook)

With this I can enter everything into SPSS as numbers and possible explore gender differences in mate preference.

However, here's the problem, every person (male or female) wrote down an unlimited number of preferences, by entering data in this manner, I would definitely inflate the number of males and females. Would this be a problem?

What method of analysis would you suggest then, for a comparison of gender differences? 2 categorical variables - Chi sq? I'm not sure.

[bugman]

Factor analysis is good for finding the most important variables (the ones that explain the most amount of the total variance in the data set). This is ususally followed by further analysis. One option is path analysis (this might suit your data set - see link below) or as previously mentioned, the factors (if they are counts) can be analysed in a cross-tab format.

http://faculty.chass.ncsu.edu/garson/PA765/path.htm


Im not quite sure what you mean by inflating the number of males and females

[CowboyBear]

Star - this really sounds to me like a qualitative research project, I think the folks have made some good suggestions but perhaps you need to look at the problem from a different angle - a lot of great qualitative research goes on in psychology, it's important not to force everything into being a "number" problem. One technique that *might* work is content analysis, which sorta toes the line between quantitative and qualitative research. (Qualitative research definitely isn't my area, but just a suggestion!)

[Star7]

To Bugman: Well I was thinking of using the results from the FA conducted by other researchers to base my data on.

Regarding the inflation of males and females, it goes something like this:
For example, if a female wrote down 5 preferences, when I enter the data, it would be something like this:

Gender, Preferences
1, 3
1, 2
1, 5
1, 4
1, 6

Instead of 1 female, there would be 5 >.<

To CowboyBear: Yea I've thought about it being a qualitative project, but because I've had no experience with qualitative projects, I thought I'll try to do it the "number" way. >.<

[bugman]

You treat it as multivariate data (as in Factor analysis) so that instead of:

1,3
1,4
1,5 etc...


it reads,

male 1: 1,2,5,6
male 2: 1,5
female 1: 1,2,3,4,5,6,7,8,9
female 2: 1,3,4,5

etc...

This shouldn't be a problem in factor analysis