“…If this was the case, it would be important to also consider how children's socio-emotional wellbeing affects maternal employment and maternal job characteristics. This would require sophisticated methodological approaches that exploit the panel data to account for reverse causation and feedback effects, e.g., simultaneous equation modelling (Marcoulides & Schumacker, 2013;Zellner & Theil, 1962), multilevel multi-process models (Haynes, Baxter, Hewitt, & Western, 2015;Lillard & Waite, 1993), or instrumental-variable regression (Angrist et al, 1996;Swanson et al, 2015). Another important factor to which more attention could be paid is the timing of exposure to poor maternal jobs.…”