What test should I use to compare these two groups?

#1
I'm not really sure where to start with this, I haven't done stats in a long time. I need to check whether the differences between 2 groups (A and B) doing the same task is statistically significant. The two groups are not paired. There are 3 outcomes for each group: 1) completed all of the task, 2) completed part of the task, 3) did not attempt the task.

I think one null hypothesis would be that group A is more likely to complete the task than group B and then second null hypothesis to test would be that group A is more likely to complete at least some of the task. So for the first null hypothesis I think I would need to compare the proportion of each group completing the task, and then for the second I would need to compare the proportion of each group not attempting the task.

But that's as far as I can get, I'm not too sure which statistical test would be best to test these hypotheses and am not really sure what to search for to find out!

Any suggestions would be greatly appreciate.
 
#3
I would recommend a Chi-square test for Homogeneity.
Thanks!

As I'm comparing the groups rather than against an expected result would one of the groups then become the expected values? I've only ever used Chi squared for inheritance patterns before - would it be able to test a one-tailed hypothesis like one group is more likely to complete the task than the other, or does it just show that the groups are significantly different without giving a direction for the difference?
 

Miner

TS Contributor
#4
Just create a contingency table with the two groups in rows and the three outcomes in columns. Enter the counts for each cell and run the Chi-square analysis.