“…For sleep‐onset difficulties and night‐waking trajectories, we used multiple logistic regressions and, for nighttime in bed trajectories, we used multiple multinomial logistic regressions. Potential confounding factors included in the models were identified from literature (Bat‐Pitault et al., ; Blair et al., ; Johansson, Ludvigsson, & Hermansson, ; Karraker & Young, ; Machado et al., ; Nevarez et al., ; Reynaud et al., ; Sadeh, Tikotzky, & Scher, ) and selected using the Directed Acyclic Graphs method (http://www.dagitty.net; Textor, Hardt, & Knüppel, ). All the models were then adjusted for family income, marital status, maternal educational level, maternal age at delivery, maternal depressive symptoms during pregnancy, maternal smoking during pregnancy, preterm birth, child's gender, child's regurgitation problems from birth to 8 months old, and maternal return to work within the first year.…”