Assuming that you have your data organized as two variables--one for breast cancer screening and one for colon cancer screening and these variables are 0/1 values, a simple correlation will tell you whether there is a relationship between the two. You can also test whether they are related with a McNemar test but I think there is a general expectation of ordering there.
Chi-square will tell very little about the direction of any observed relationship by itself. Look a the residuals for a better understanding.
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]] On Behalf Of Lou
Sent: Wednesday, February 21, 2007 10:02 AM
To:
[hidden email]
Subject: Quick help needed: Chi-Square
Dear all,
I have a group of women who have all undertaken both breast and bowel
cancer screening during a given time period.
I would like to establish:
(a) whether there is a significant relationship between the screening
programmes (i.e. if a woman undertakes breast screening, is it more likely
that she will have undertaken bowel screening etc);
(b) whether there are significant differences between uptake in the
screening programmes. For instance, 79.30% of the women completed breast
screening whereas only 35.36% of the same women completed bowel
screening.
Can I use a simple Chi-Square test to answer these questions or do I have
to take into account the fact that I am working with the same women (i.e.
does this mean dependencies)? My instinct tells me that I can use Chi-
Square since I am looking at independent screening programmes, but I don’t
have high confidence in this!!
Any advice please?
Thanks,
Lou
Locked
