# Thread: What tests to unleash on the data?

1. ## What tests to unleash on the data?

Hey everyone!

So here's some background: I'm doing a course on data-analysis [SPSS] and my teacher likes the idea of us finding our own solutions, so we don't get much help from him (We're allowed, and cheered on to find help on the internet). So that brings me here! I'm not too good at this and I have an exam in a couple of days, in which I need to do five analysis on one data-set. I've tried my best, but am hoping that I can get some tips here. I'm using version 23.

The question that the data should answer (I have a bit of room in this, but it should HELP get in the direction of an answer) is: What is the relationship between gender and smiles on characteristics? People[m/f] get to see a picture (m/f with a smile or no smile) and need to judge three characteristics.

Three characteristics are being measured: dominance, attractiveness and trustability. For each of these characteristics you have two conditions (smile/no smile). However, the data is made so that every column is made up of one of these combinations (ex. Variables are: "dominance/s/f, dominance s/m, dominance ns/f...etc), under which the score was presented. So i have a lot of columns and I am NOT allowed to alter this.

Now I started of using compute to make total scores of all the characteristics, and total scores of [S] and [NS]. So I made five new variables.

I proceeded to first make graphs of the total scores of characteristics to check for normality (this counts as analysis one)

I then proceeded to do a MANOVA with gender as fixed factor, and the total scores of all characteristics as dependent variable (this counts as analysis two). This tested the effect of gender on judgement of characteristics.

But then I get stuck.. I've considered using difference score of the totals of [S]/[NS] and using that as a dependent variable in an ANOVA, with maybe gender as a fixed factor, but I'm not sure what I would be testing.

I also want to compare the two totals of [S/NS] to see if they significantly differ(to test the effect of S). Though I'm not sure how to do this since the variables are "total S" & "total NS" with scores underneath. I don't have a grouping variable and im not allowed to restructure =/

I'm sorry for the long story. Any help at all will be greatly appreciated!

 Tweet