We conducted a study looking at the "applicability of expert judgment in occupational hygiene decision making and effect of training". 13 hygienists (8 experienced and 5 non-experienced) predicted the exposure intensity in 7 different exposure scenarios before and after a course about exposure assessment. In other words we have 14 measurements from each hygienist (7 before educational course and 7 after completion of educational course)
Now I am confused in using a mixed effect model for data analysis
1- Should I consider measurements in different scenarios (e.g. 1-7) as a correlated and repeated measurements (because each hygienist have made 7 predictions and all of these predictions are suffering from error belongs to this subject).
Or
2- Is it necessary to use before and after measurements of each hygienist as two repeated measurements and?
Warmly welcome your comments!
I am not quite sure what exactly the question is that you ask? What are your explanatory and response variables? Need some more details...
In the meantime check out the following paper, which might help,...
Baayen, R. Harald, Douglas J. Davidson, and Douglas M. Bates. "Mixed-effects modeling with crossed random effects for subjects and items." Journal of memory and language 59.4 (2008): 390-412.
Explanatory variables: sex, experience (categorized as experienced and non experienced )
Response variable: prediction error ( calculated as relative error for each subject in each scenario
But I am not sure about the role of these variables: prediction time(before and after education) , and exposure scenarios (categorized as 1-7)
Hm... prediction time would be another response variable if you are interested to see if it varies depending on experience. Not sure about the exposure scenario.
With 13 subjects already looking at sex and experience as explanatories takes 4 degrees of freedom, if you add further fixed and random effects your power decreases further.. So in general with only 13 people measured I am a bit "worried" that you have so many variables that you want to integrate.
Clearly you have to ask yourself what are the variables of interest (fixed effects) and which variables you are not interested to estimate, but want to control for as part of the experimental design (random effects).
That means... keep it as simple as possible...
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS