Table 1. Description and coding of the examined predictors.
| Predictor | Used to examine… | Coding |
|---|---|---|
| Linear age | Linear age effects | • Coded in years at T1, centereda and divided by 10b |
| Quadratic age | Quadratic age effects | • Linear age variable2 |
| Cubic age | Cubic age effects | • Linear age variable3 |
| Linear symptom changes during pregnancy | Linear changes in depressive, anxiety, and stress symptoms from early until late pregnancy | • Coded with the time span (in months) between the respective assessment wave and the time point of delivery for all measurement occasions during pregnancy (T1, T2, and T3) (i.e., coded with -7.25 (-29/ 4) for T1 (week 10 to 12 of gestation), with -4.25 (-17/ 4) for T2 (week 22 to 24 of gestation), and with -1 (-4/ 4) for T3 (week 35 to 37 of gestation))c • Coded with 0 for all measurement occasions after delivery (T5, T6, and T7) |
| Linear symptom changes after delivery | Linear changes in depressive, anxiety, and stress symptoms from two until 16 months postpartum | • Coded with the time span (in months) between the respective assessment wave and the time point of delivery for all measurement occasions after delivery (T5, T6, and T7) (i.e., coded with 2 for T5 (two months postpartum), with 4 for T6 (four months postpartum), and with 16 for T7 (16 months postpartum)) • Coded with 0 for all measurement occasions during pregnancy (T1, T2, and T3) |
| Short-term symptom changes after delivery | Abrupt and transient short-term changes in depressive, anxiety, and stress symptoms at two months postpartum (as compared to all other waves) | • Coded with 1 for T5 (two months postpartum) • Coded with 0 for all other measurement occasions (T1, T2, T3, T6, and T7) |
| Long-term symptom changes after delivery | Enduring long-term changes in depressive, anxiety, and stress symptoms more than two months after delivery (as compared to all previous waves) | • Coded with 1 for T6 (four months postpartum) and T7 (16 months postpartum) • Coded with 0 for all previous measurement occasions (T1, T2, T3, and T5) |
a The linear age variable was centered to ensure that the intercept in the overall model referred to an average-aged woman of the total sample
b The linear age variable was divided by 10 to ensure that the quadratic and cubic age effects would not become too small to be reported rounded at three decimals
c Because a few women ended their study participation during pregnancy we assumed an average length of pregnancy of 40 weeks and subtracted the average week of gestation, in which T1, T2, and T3 were conducted, from this length.