LPTS, You bring up good points. Self-reports are notoriously unreliable (like eye-witness testimony) however, it is still use-able data. In terms of missing da

LPTS, You bring up good points. Self-reports are notoriously unreliable (like eye-witness testimony) however, it is still use-able data. In terms of missing days, the questions can be ordered and repeated to those who miss. In terms of false correlations, you can reduce this by being careful about how you use and analyze the data-set. Multiple regression analysis can be dangerous--it's all in how it is used. If you limit the items you correlate to those for which you have a hypothesis, minimally this data should be valuable in that manner.