Statistics Question - Any Math Geeks?

GoingNowhere

Active Member
Joined
Jul 13, 2008
Messages
1,793
Likes
0
Points
36
Location
USA
#1
This has been stumping me for a bit now and I'm struggling to find the answer by googling. I am trying to compare two percentages to determine if they are statisticall different. The issue is that one of the groups is a subgroup of the second group.

I'm going to change the information around a little bit to keep confidential things confidential, but here's an equivalent setup:

Let's say that I have 100 student athletes out of a population of 1000 students. 70% of the athletes like to eat at Chipotle. 55% of the entire student body likes to eat at Chipotle. I have no information specific to "non-student athletes" (unless it could be computed mathematically?)

How to I determine if the percent of student athletes who like Chipotle is statistically different from the percent of the entire student body that likes Chiptole?

Chi squared tests seem to be the answer for those situations in which you have two exclusive groups (e.g. men vs. women). But can I use the same test for the group-subgroup analysis? And if so, how do I do it?

Thank you kindly!
 

Izzy's Valkyrie

Very Food Agressive
Joined
Dec 10, 2008
Messages
5,946
Likes
0
Points
36
Location
Washington DC
#2
I love stats when I was in school but don't remember specifics. That said, are you trying to compare the whole population to the athlete population or the non-athlete population to the athlete population?
 

GoingNowhere

Active Member
Joined
Jul 13, 2008
Messages
1,793
Likes
0
Points
36
Location
USA
#3
Well if I had the data for the non-athlete population, I could do athletes vs. non-athletes and then all would be good and well because each group would be exclusive of the other. Actually, that would be my ideal. But the only information that I have is athletes vs the entire student body. I don't mind doing this comparison, but I don't know how.

Unfortunately, I am not the original source of the data and it might be a pain to get the exact numbers that I would like. I already had to ask for 'n' values for each population and that took several days to come through. If I can work with what I have, that would be best.
 

eddieq

Silence! I ban you!
Staff member
Joined
Jul 31, 2007
Messages
8,833
Likes
3
Points
38
Location
PA
#4
Well, Student Athletes at 70% of 100 is 70 Chipotle Enjoyers (CE)
Total Students at 55% of 1000 is 550 CE's

Since we know that Student Athletes are Students, you know that of those 550, 70 are Athletes. So Student non-athletes that are CE's would be 550 - 70 or 480. 1000 Students total and 100 are athletes, so the other 900 students are non-athletes.

480/900 is 53.3% of Student non-athletes are CE's.
 

GoingNowhere

Active Member
Joined
Jul 13, 2008
Messages
1,793
Likes
0
Points
36
Location
USA
#5
Firstly, thank you! I appreciate that - that's all I needed :)

Secondly, I am embarrassed at the extent to which my math skills have gone down the tube. Four years ago I would have laughed at my current state! So anyhow, thank you!
 

Members online

No members online now.
Top