We have processed accelerometer data that includes segments for the school day (e.g., 9am-3pm) which successfully produces output for the time windows 0-9hr, 9-15hr, 15-24hr. The devices started recording on day 1 when they were distributed and stopped recording when they were returned on day 9, so both of these days were partial and as expected in part2daysummary, no PA data are generated because the minimum wear time criterion wasn't met. We used data_masking_strategy =1 to account for the protocol (although other strategies have also been tried and produce the same results).
My question relates to the part2summary file output for the AD_N_hours_x-xhr columns, which when summed for the 3 time windows do not add to 24 hours because partial 'wear' on the incomplete days 1 and 9 are included in the calculation for these columns (the same applies to the adjacent AD_N_valid_hours_x-xhr columns). Is there a way for the values in these columns to exclude the partial days wear so the averaged recorded hours sums to 24? We can of course extract this information from the part2daysummary output if not.
thanks, Stuart