Dear Allstat members, I am involved in the design stages of a study which aims to relate measurements of lung function to exposure to a particular pollutant for a small group of subjects (20 or 30). Both lung function and personal exposure will be recorded daily on continuous scales of measurement for 60 days. I will need a regression analysis taking account of the autocorrelation between the lung function measurements. A similar study looked at deviation from the mean lung function by subtracting a subject's mean lung function from the rest of his/her measurements. I am aware that I will have to investigate a number of possible lags for the pollutant measure. I know the mean and range of the likely pollutant exposures. I also have an estimate and confidence intervals for the effect of a unit increase in pollutant on lung function found in a similar study. Does anyone have any additional suggestions for the analysis required? Does anyone know how to perform a power calculation for this type of study? I would appreciate any references relating to either an appropriate analysis or power calculations. Thank you, Gordon Prescott Department of Public Health University of Aberdeen [log in to unmask] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%