# Thread: Statistical significance: T-test or something else? (Percentages from two samples)

1. ## Statistical significance: T-test or something else? (Percentages from two samples)

Hi guys,

I have the following data and I want to show that the negative change for group 1 is statistically significantly different from the negative change for group 2. And along the same lines, I also want to compare positive change between group 1 and group 2.

The number of observations is 312 for group 1 and 972 for group 2 and the means are in percentages. The percentages are obviously quite different - but how do I prove that they are statistically significantly different (ideally with p-values)?

Key Indicator-----------First Group------------Second Group

------------------Observations---Mean-----Observations----Mean

1 Grade change
negative----------35-------------11,22%--------2--------------0,20%

no change------149------------47,76%-------316-----------32,51%

positive---------128-------------41,03%-------654-----------67,28%

Is it possible to do a t-test here or what would you recommend?

2. ## Re: Statistical significance: T-test or something else? (Percentages from two samples

I don't understand why you are calling these percentages "means". Aren't they just percentages? 35/312 = 0.112 (11.2%). That isn't the mean of anything.

I would do a 2 (group) x 3 (change type) chi-square. I could also see dropping the non-changes and doing a 2x2 chi-square to see if, among those who changed, the pattern differed. I think it will be highly significant either way.

 Tweet

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts