# comparing frequencies

#### mrjk

##### New Member
I have a survey question that asks which of 3 activities participants regularly participate in. They are told “select all that apply.” So they can mark any or all of the activities. As results, I get frequencies (# and %) for each activity.

Now I want to compare the frequencies of the different activities. Say out of a sample size of 80, 50 people said they participate in activity 1 and 60 people said they participate in activity 2 and 45 people participate in activity 3. I’d like to know how statistically significant the difference between those frequencies is.

I have a hard time getting my head around doing a chi square or fishers test for this, because the “activity” variable is not either/or. Or do I view it as 3 different activity variables, each having two possible answers, yes and no. Then what statistic would I use? I would even be ok with just comparing two of the activities at one time.

#### Karabiner

##### TS Contributor
Maybe a repeated-measures approach with Cochran's Q can answer you reseach question.

Regards

K.

#### mrjk

##### New Member
Maybe a repeated-measures approach with Cochran's Q can answer you reseach question.

Regards

K.