| Date: | Fri, 9 Jun 2000 09:24:16 -0400 |
| Reply-To: | Chunki Fong <chunki.fong@NDRI.ORG> |
| Sender: | "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU> |
| From: | Chunki Fong <chunki.fong@NDRI.ORG> |
| Subject: | signif test for weekly 0/1 data |
| Content-Type: | text/plain; charset=US-ASCII |
Hi,
I have data similar to the format below
id group wk1 wk2 wk3 wk4 wk5
1 1 1 1 1 0 .
2 1 0 . 1 0 0
4 2 0 1 1 1 1
5 2 0 0 . 0 0
which include the subject id and treatment group. For each subject, a test is done each week (idicate by wk1-wk5) which 1 indicate positive ,0 indicate negative, missing indicate subject did not take the test.
I like to compare the percentage positive between treatment group 1 and 2.
I can plot the percent positive for each week in a graph by plotting the percent over the subjects within a group for each week such as
group wk1 wk2 wk3 wk4 wk5
1 0.50 1.00 1.00 0.00 0.00
2 0.00 0.50 1.00 0.50 0.50
and visually compare them.
I wonder if anyone have worked with data in this format before. Can anyone advice me of a proper significant test I can use in SPSS? Does Repeated Measure applies in this case? Thanks for any advice.
Chunki Fong
|