Which Statistical test Should I Use?

• Jan 24th 2011, 02:10 PM
Warren
Which Statistical test Should I Use?
I don't know what test to use to see if there is a relationship between two sets of data when they rely on each other.

Example:

# of people who slipped on wet floor: day 1 - 7. Day 2 - 8. Day 3 - 7.
# of people who slipped on dry floor: day 1 - 3. Day 2 - 2. Day 3 - 3. . .
Basically the number of people is always 10. I don't think I can do a chi square test because if one number changes, then so does the other. Is there a specific test for this situation?
• Jan 24th 2011, 02:26 PM
pickslides
Quote:

Originally Posted by Warren
I don't think I can do a chi square test because if one number changes, then so does the other.

What do you mean by this?

Have you seen the Wilcoxon signed rank test?

Look here for an example,

Wilcoxon signed-rank test - Wikipedia, the free encyclopedia

give it a test drive with your data!
• Jan 24th 2011, 02:44 PM
Warren
I don't know if I'm right about this, but I don't think I can do a chi square test because the variables have to add up to ten (if one is 7 then the other has to be 3). I thought that this was the reason that you can't use percentages in a chi square test.