I would like to investigate if the incidence of an outcome that occurred during childhood has changed across the last decade. However, the studies reviewed investigated such outcomes retrospectively, using a retrospective questionnaire. How could I account for both variables (time + age), since a study that was published recently might had a samples of elderly individuals, while a study that was published twenty years ago might have used a sample of adolescents.
Looking at two variables simultaneously, I'd think meta-regression. I know R and Stata have routines for that. I'd assume SAS and SPSS do as well.
However, it isn't that simple because age is an average. Even if both groups had the same mean age of 12, one group might have range 11-13 and the other, 8-16.
You say this is a childhood outcome but elderly individuals may have been surveyed? Is it practical to classify the studies by time elapsed since the event?
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis