10-18-2016 02:29 PM
Hello, everyone! I'm working with a dataset in SAS and have hit a road block.
I'm working with a sample of vaccination data to determine up-to-date status of children.The data have weights as a separate variable and were collected stratified by two factors (county and school type). The database contains individual-level data on roughly 8,000 children. Based on how I have worked with the database, the individuals are coded as 1 (up-to-date) or 0 (not up-to-date) for each vaccine of interest. Up-to-date status was determined for whether or not the children are up-to-date at certain points in time (i.e. school start, 30 days after school start, 60 days after school start, etc.) so that I may be able to determine the change in vaccination coverage over time. Again, all up-to-date variables are shown as 1's and 0's.
What is the best method for determining whether or not the increases in vaccination coverage for the whole population between two points in time are statistically significant?
10-18-2016 03:54 PM
One way is a TTest on the "mean" or proportion.
Here's a brief example
data have; input time status count; datalines; 1 1 2300 1 0 5700 2 1 4000 2 0 4000 ; run; proc ttest data=have; freq count; var status; class time; run;
This is a test for difference between the groups. A signficant result would mean the the mean proportion of vaccinations differed between time 1 and time 2.