Background-In a community-based prevalence study, sampling was done from different geographical regions using probability proportional to sample size method and random sampling of individuals within that frame. After the collection of data, I find that the characteristics of the sample differ between the different regions-in all demographic and socio-economic characteristics.
What impact would this have on the results-the primary objective being point prevalence of a disease? I need help with this, please.
Hello Gomathi,
Congratulations on completing what sounds like a pretty well-planned sampling study. Your results would indicate that:
1. The crude (overall) incidence rate differs from adjusted rates; hence
2. Geographic region and the selected demographic / socio-economic characteristics are important explanatory/moderator variables for accounting for differences in observed incidence rates.
Good luck with your work.
Depending on your research objectives you can create independent samples according to the different regions (in the same context as you extracted the elements in the sample in a proportionate way). Then you can define zero hypotheses for the differences of the average values by region (by samples) and apply parametric statistics, ie a parametric test (If the purpose of your research is statistical inference.
I would like to expand on what David Morse means when he says "Geographic region and the selected demographic / socio-economic characteristics are important explanatory/moderator variables for accounting for differences in observed incidence rates."
In terms of a regression analysis, these are tests of whether the effects of your predictors differ between the two regions. These should be tested by interaction effects using the region as a dummy variable. If you are not familiar with dummy variables or this kind of interaction effect, a good introduction is Hardy, Regression with Dummy Variables.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size