Table 3.
Multilevel Regression Models Predicting the Index Week’s Emotional Symptoms and HPA Axis Functioning from Average E2 Across Weeks and Weekly Changes in E2 (Depressed Subsample)
| Outcome at Current Week (Full Sample) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||||
| Predictor | CESD Depressed Affect Subscale |
CESD Lack of Positive Affect Subscale |
CESD Somatic Subscale |
Hopeless- ness |
Anxiety | Anger | Rejection | Guilt | Cortisol Mean (First Two Samples) |
Cortisol Waking Level |
Cortisol Awakening Response |
|
Fixed Effects
|
|||||||||||
| Intercept | −1.01 (8.58) | −1.09 (1.90) | −1.32 (3.18) | .33 (.94) | .57 (1.26) | −1.49 (1.30) | −.45 (.68) | .050 (.81) | .24 (.045) | .103 (.061) | .26** (.08) |
| VMS Severity | 1.75** (.31) | .15 (.13) | .85*** (.13) | .33 (.081) | .13 (.11) | .23 (.11) | .39*** (.080) | .17** (.042) | - | - | - |
| Avg E2 Across Weeks | 11.36** (5.05) | 3.44* (1.35) | 4.90* (2.17) | .041 (.64) | 1.52* (.67) | 1.89* (.88) | 1.31* (.65) | −.15 (.51) | .023 (.038) | .105* (.045) | −.10 (.059) |
| Abs ΔE2 Past Week | .51 (.82) | .44 (.76) | .25 (.88) | .13 (.49) | .87 (.85) | .41 (.81) | −1.90 (1.58) | .67 (.43) | −.041 (.088) | .019 (.045) | −.060 (.084) |
| Direction of ΔE2 | .10 (.51) | .41 (.58) | −.81 (.55) | .49 (.36) | .24 (.57) | −.10 (.58) | −.82 (.49) | .068 (.25) | −.056 (.040) | −.023 (.032) | −.017 (.05) |
| ΔE2 X Direction | 4.09** (1.26) | 2.17* (1.05) | 3.49** (.98) | 1.54 (.93) | −.39 (1.36) | 1.59 (1.41) | 2.55** (.95) | 2.34** (.93) | .32** (.11) | .14** (.042) | .061 (.13) |
|
Random Parameters
|
|||||||||||
| Intercept | 9.63** (3.30) | 5.33** (2.06) | 6.79* (2.98) | 3.02** (1.19) | 1.89 (1.42) | .59 (1.67) | 4.58* (1.80) | 2.38* (1.15) | .42* (.12) | .019* (.005) | .012 (.014) |
| Abs ΔE2 Past Week | 9.47** (4.26) | 4.34** (1.80) | 5.49* (2.69) | 6.32** (2.59) | 3.26** (.96) | 6.75* (2.15) | 5.68** (2.19) | 2.89** (1.18) | .045* (.021) | .025 (.024) | .044 (.05) |
| AR(1) | .51* (.23) | .19* (.08) | .88*** (.04) | .97*** (.02) | .029 (.29) | .29 (.32) | .98* (.39) | .31 (.77) | .035 (.35) | .20 (.34) | .045 (.29) |
| Residual | 12.24* (5.60) | 1.60*** (.36) | 8.28**(3.75) | .22** (.07) | 2.27*** (.69) | 2.71* (1.20) | 4.98* (1.40) | .68 (.84) | 5.43** (2.07) | .005** (.002) | .022** (.007) |
| −2 Log Likelihood | 393.0† | 341.4† | 385.4† | 165.2† | 340.3† | 330.9† | 189.7† | 179.5† | 118.3† | 115.3† | 138.49† |
| f2 for Mean E2 | .018 | .023 | .060 | .018 | .045 | .052 | .0087 | .0053 | .012 | .011 | .011 |
| f2 for E2 Changes | .056 | .065 | .098 | .025 | .035 | .028 | .084 | .083 | .16 | .20 | .055 |
Note. Standard errors are in parentheses.
*
p < .01.
**
p < .001.
Direction of E2 change is coded such that 0 indicates E2 has decreased since the previous week, and 1 indicates that E2 has increased since the previous week. Significant fixed effects are shown in bold.
†
Change in −2 Log likelihood over a null model (a model with no predictors) is significant at p < .001. ΔE2 represents the absolute value of the difference between last week’s E2 value and the index week’s E2 value.