# Comparing two samples of ordinal data over time

#### scottkirk11

##### New Member
I am having trouble determining what test to run and how to input the data for a simple comparison of a before and after situation. I would appreciate any help you can provide. Specifically what test to run and how to input the data into SPSS or Excel.

The hypothesis I am testing to see if there is a statistical difference between the two results or not. The hope is that the publicity recieved over the time period affected the population's familiarity with the company.

I took two different samples at different times. One was taken in March and one in September. The first sample was 407 and was randomly pulled from a population of 1.3 million using a random digit dialer program. The second sample was 303 and was also randomly pulled from the same population using a random digit dialer program.

The question was: How familiar are you with Go! Airlines?
Answer options were: Very familiar, Somewhat familiar, Not very familiar, Not a all familiar

The data for March is:
Very Familiar: 114
Somewhat Familiar: 216
Not Very Familiar: 49
Not at all Familiar: 28
TOTAL (N) = 407

The data for September is:
Very Familiar: 155
Somewhat Familiar: 124
Not Very Familiar: 15
Not at all Familiar: 9
TOTAL (N) = 303

I know this is probably a really simple question, but I appreciate your help! Let me know if you have any questions: scottleekirk@gmail.com

#### IronMan

##### New Member
Use a Chi Squared test of homogeneity. Wikipedia explains the formulas, you just need to use Excel to calculate the the right tail probability of hte final value of the Chi Squared test statistic at 3 degrees of freedom.

If alpha>p, reject Ho.