Quality indicators for performance of organizations often come in the form of counts, e.g. number of employees with OHS injuries at any given period of time or a period. But over successive periods of time the same people may be captured in this count. Is the only solution to restructure the QI to reflect the number of employees that suffered injuries over a set period of time which would automatically exclude them from the next capture (unless of course they injured themselves again)?
Linda,
The solution sounds reasonable to me. What are your misgivings about it?
Pat
No misgivings. I just know that I will probably get the data in the other form as well.
Dear Linda,
if I understand you correctly, then you have kind of longitudinal count data. Isn't then one way to analyse such data a generalised linear mixed model. There you can specify rather flexible the dependence structure between consecutive observations of one organization. Of course the type of dependence structure has to be the same for all organizations or at least for the members of one group, if you can group the data. Maybe this paper
http://de.scribd.com/doc/117716839/Bolker-Et-Al-2009-General-Mixed-Model
helps you to decide if a GLMM for count data is what you want.
Best regards, Stephan
- Badges
- Science topic