Explore BrainMass

Chi Square or ANOVA or T-Test(s)

I have collected data for my pschology experiment and I am uncertain which analysis test to run. Although they will all be a challenge since this is my first year of University and we haven't done much stats work yet.

Participants were shown an image and asked to choose what they thought the people in the picture were likely to do next. There were two conditions:

1. an image with a neutral title 'what do you think the group of friends will do next'

2. the same image with a negative title 'what do you think the gang of youths will do next'

Each particpant was then asked to select what they thought the group/gang would do next from a list of 6 activities i.e. play football, smoke drugs (3 negative and 3 neutral)

I am trying to show a connection between using negative language and invoking negative sterotypes. Also though I'd like to check whether gender and age have any impact.

I thought I could run a number of T-tests to find, but that doesn't seem to be the most effective way. I looked in my stats book at Chi-Square but the example doesn't seem to apply to me and it looks very confusing....I am really not sure how to figure this out other than manually.

I've attached the participant data and an example of the neutral image and question that the participants saw.


Solution Preview

Let's see if I can set you on the right path with respect to using the Chi-Square or the "t" test analysis. First of all, as an applied statistician (PhD) and psychologist (EdD) I want you to do something for me. I want you to put all your fears and anxieties about statistics aside as statistics are nothing more than the addition, subtraction, division, and multiplication processes we all learned in high school. Statistics are only expressed differently is all. :-) Statistical processes are not magical and should not be haunting. Statistical processes have been developed to help the research investigator determine whether or not relationships, differences or effects exist in a measurement situation. In order to determine these we must have independent (treatment) and dependent (measurement or predictive) variables to examine. Therefore, there are certain statistical process that permit you to establish whether or not a relationship exists between two variables; whether there is a difference between two variables; and whether of not one variable produces an affect on another variable. Of course studies are not always limited to two variables as many studies involve multiple variables. ...