I have stress measures (continuous measures between 0 and 10 scales) and would like to see if the stress measures show any relationships with time, air temperature, activity, dehydration, and possibly interactions between these variables. Fifteen subjects, who involved in consistent working, were simultaneously measured (all variables mentioned above) every 5 minute for 2 hours in the morning, then 2 hours intermission, then 5 minute for 2 hours in the afternoon. Missing values existed due to long-hour fieldwork. The dehydration variable was ordinal (0 – 4) and the rest of explanatory variables (weather, activity) are continuous. The original data look like this:
Time subject temp activity stress dehydration
0 27 17.513 6022 3.60 0
0 13 17.513 4224 3.31 0
0 23 17.513 5941 3.86 0
0 7 17.513 6501 3.44 0
0 11 17.513 6031 2.40 0
5 1 17.513 7249 4.26 0
5 21 17.513 8871 0.16 0
5 7 17.513 7141 2.79 0
5 13 17.513 4621 2.25 0
5 12 17.513 9642 3.48 0
10 10 17.513 6389 4.19 0
10 7 17.513 5363 3.70 0
(continues..)
I thought generalize estimating equation (xtgee) might be useful [like “xtgee: stress temperature activity dehydration (and interaction terms) i(subject) t(time)”], although I have never used this analysis before and am not sure if this idea is on the right track. Is it appropriate to use this analysis under the condition of time break (2 hours) in the experiment? In addition, exchangeable correlation matrix is appropriate to use in this case? Finally, should I create categorical variables in activity and temperature (e.g. high, medium, low) to simplify interactions?
.
Thank you
Miyo
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
