Teun Remmers

General discussion | 185 Previous studies that investigated objective environmental determinants have traditionally focused on the residential neighborhood. However, people tend to spend considerable time outside their residential neighborhood (77, 78). Alternatively, environmental exposure may be defined by multiple locations (i.e. 'anchor points') and often travelled routes between these anchor points. GPS data may also allow researchers to analyze a participant's environmental exposure based on exact GPS locations. This means that environment-PA relationships are investigated within the same specific geographical location, measured within the same time frame (18-21, 56, 58, 79-82). With this so-called contemporaneous momentary design, the specificity of environmental exposure assessment may be improved (83). This design may be useful in epidemiological studies where geographical exposure (for a specified amount of time) can be defined when the GPS location is within a certain distance. In PA-studies however, the contemporaneous momentary design may be less useful as PA is a complex interplay between spontaneous and planned behavior; involving memory of location of PA facilities, time- and capacity constraints, social interactions, and compensation mechanisms. Also, contemporaneously momentary designs are vulnerable to selective daily mobility. This might occur when environments are deliberately visited for PA (e.g. sports grounds). In this way, it is not the environment that influences PA, but a participant's pre-conceived choice that biases directionality of the relationship between the determinant and PA (34, 49). Consequently, modern GPS analyses present researchers with a trade-off between improved specificity of environmental exposure assessment versus the potential for causal inference (22, 34, 49). Future studies are therefore encouraged to incorporate frameworks from both health sciences (3) and transportation research (72), in order to make informed decisions about children's exposure environments (or activity spaces) based on distances between school and homes, and knowledge of potential other frequently visited anchor points. In addition, future studies may also exploit the potential of GPS logging for investigating social environmental determinants. As one study showed, it is possible to investigate locations of joint child-parent PA based on the nearness of the parent's and child's GPS locations (84); future studies may apply this by objectively measuring children's social interactions and locations where joint PA takes place. Altogether, although objective measurements provide researchers with reliable estimates of specific attributes of the environment, the perceived environment and objective environment are considered to be conceptually different. Therefore, both can be complementary in PA research.