Teun Remmers

154 | Chapter 9 during school hours, while the child and one of their parents received a verbal and written invitation for an electronic questionnaire. Measurement Socio-demographic measures Directly after measurement, children and parents filled in an electronic questionnaire focusing on aspects such as birthdate of the child, residential address, perceived physical and social environment, child's (travel) behaviours and homework. We assessed whether children from divorced parents were residing at two locations, and whether children potentially moved home between baseline and follow-up measurements. Schools provided detailed class timetables for the data-collection period. Accelerometer and GPS loggers In this study, accelerometers (GT3X, ActiGraph, Pensacola, Florida) were set to record data at 10 second epochs. Actilife version 6.11.9 was used for initialization and downloading. The GPS logger used in this study (BT-Q1000XT, Qstarz International Co, Taipei, Taiwan) showed relatively good static spatial accuracy compared to other units (36), and acceptable dynamic accuracy (37). The manufacturer's software QTravel version 1.46 was used for initialization and downloading. GPS devices were also set to record data at 10 second epochs while recording parameters such as date, time, longitude, latitude, and speed. GPS and accelerometer devices stopped logging data when storage capacity was full (33). Data analysis Data management Accelerometer and GPS data were processed using the Personal Activity and Location Measurement System (PALMS), which allows users control over most parameter settings in a web-based application (32, 38). In order to handle the data-load, data were processed in PALMS separately for each school and time-point (i.e. baseline and follow-up). PALMS categorized intensity of accelerometer activity into sedentary time (ST), light PA (LPA), moderate PA (MPA), and vigorous PA (VPA) according to Evenson's cut-points (39), which performed best in free-living activities of 5-15 year-old children (40). We defined non- wear time as  20 consecutive minutes of zero counts (41). Datasets were cleaned based on extreme speed (i.e. threshold  130 kmph) and changes in elevation (i.e. threshold  1000 meters) in the GPS data. Algorithms as described in Carlson et al. (2015) were used for trip and trip mode classification (e.g. pedestrian, bicycle) from the GPS data (31). As the sample in the study of Carlson et al. (2015) consisted of commuting cyclists that were expected to accumulate higher cycling speeds, the present study deviated from these thresholds by using the 10-25 kmph bicycling speed-threshold. Invalid GPS points were imputed from the last known valid point, for up to 10 minutes. Finally, PALMS matched accelerometry and GPS data based on start- and end-times of the GPS logger.