Figure 5From: Investigating causality in human behavior from smartphone sensor data: a quasi-experimental approach Balance check. Standardized difference between treated and control samples for each confounding variable when the applied treatment is (a) the variable \(U^{u, d}_{t_{i}}\) and (b) the variable \(O^{u, d}_{t_{i}}\). The standardized difference for all the confounding variables is less than 0.1, thus the groups are balanced.Back to article page