# Comparing groups. Should I use Frequency or T test ?

#### fordy_4

##### New Member
Morning everyone,

New member to the stats forum and glad to be here.

I'm learning stats as part of an internship at a company, but I am having trouble using it as I have (a) little stats experience (b) no experience with this type of software (SPSS).

I have been tasked with using SPSS for these two questions, but I am a little confused about how to get meaningful results that answer the questions. FYi; this is for a gym and they enlisted me to help with market research to understand consumers.

I have outlined the two questions I have to answer to date, with how I think I should answer them. But I need some assistance as I feel I maybe going in the wrong direction.

(1) How does the device used differ across the three psychographic groups?
Device used: 1=mobile, 2=tablet, 3=computer
Psychographic group: 1=new to healthy living, 2= social poser, 3=gym junkie

-I will run a frequency test to reveal the mean, media and mode. However, when I do this it gives it as one table.
-How would I be able to get the mean, median, mode and std. deviation for the three individual psychographic groups based on the three devices?
-Should I use a T-Test instead?

(2) How does the time of day in accessing our instagram page differ across four age groups?
Time of day: 1= Morning, 2= Midday, 3= Afternoon, 4=Night
Age groups: I don't have any age groups, so I sill need to split the groups ?

This is the most confusing part for me. Should I do another T test?

#### Berley

##### Member
T-tests are only for continuous data. Since you don't have any, you can't use a t-test for #1. I would use a chi-square test. That is on the Analyze > Descriptive Statistics > Crosstabs screen. Select "statistics" and make sure you click "chi-square."

For #2, what do you have for ages? You say you don't have "age groups" -- do you have exact ages? If you have the age groups (or can assign the proper group to each respondent), I would look at the Spearman's correlation coefficient. That one works when comparing categorical data (time of day) and ordinal data (age groups in order). That's on the Analyze > Correlate > Bivariate Correlations screen.

Perhaps others will have other suggestions as well.

#### fordy_4

##### New Member
T-tests are only for continuous data. Since you don't have any, you can't use a t-test for #1. I would use a chi-square test. That is on the Analyze > Descriptive Statistics > Crosstabs screen. Select "statistics" and make sure you click "chi-square."

For #2, what do you have for ages? You say you don't have "age groups" -- do you have exact ages? If you have the age groups (or can assign the proper group to each respondent), I would look at the Spearman's correlation coefficient. That one works when comparing categorical data (time of day) and ordinal data (age groups in order). That's on the Analyze > Correlate > Bivariate Correlations screen.

Perhaps others will have other suggestions as well.
Thank you very much for that. For #2, I broke made groups with the split groups tool, so I had four even numbered groups for the ages.