Year 12s: what will be the first way you research your future options? >>

Help needed choosing statistical test

Hi,

I've collected some data but I am confused over which test to use to analyse it.

It is a between-subjects design, with participants being allocated to either a first-person condition or third person condition.

All participants answered one question on whether they thought the "bystander" scenario was moral and another question on whether they thought the "footbridge" scenario was moral (I have Yes/No, and likert rating scale data for both these questions so can use either)

What I am trying to analyse is whether participants rated the "bystander" scenario as more moral than the "footbridge" scenario and whether this is affected by the group they were in (first/third person).

Any advice on what the appropriate test is, would be greatly appreciated! (I'm using spss).

Thank you!

Reply 1

iammichealjackson

Original post by sunshinebubbles

I guess an issue here is that data from Likert scales (unless you're using the average response of many Likert questions) is unlikely to fit the assumption of having data that can be modelled with the normal distribution. Otherwise I'm thinking you could have used a random effects ANOVA model.

I guess the simplest method I would know is to convert the outcome measure into one of three codes - coded bystander and more moral, footbridge as more moral, or coded both equally moral. You can then do a chi-square test of independence to see if the condition influence the distribution of responses across the three outcomes.