Skip to main content
. 2020 Dec 21;15(12):e0243188. doi: 10.1371/journal.pone.0243188

Table 2. Linear regression model assessing associations between reported exercise changes during the COVID-19 pandemic and Edinburgh Postnatal Depression Survey (EPDS) score.

Variable B coefficient (SE, 95% CI) p-value
Intercept 12.4 (1.07, 10.4–14.5) <0.001
Age (years) -0.078 (0.036, -0.148-(-0.009)) 0.028
Weeks pregnant at time of survey 0.024 (0.014, -0.003–0.051) 0.085
Race/ethnicity:
    White reference
    Hispanic, Latino, or Spanish origin 1.32 (0.503, 0.332–2.30) 0.009
    Black or African American 0.062 (1.07, -2.04–2.16) 0.954
    Asian -0.991 (0.672, -2.31–0.327) 0.140
    American Indian or Alaskan Native -0.340 (1.48, -3.24–2.56) 0.818
    Other -0.449 (0.858, -2.13–1.23) 0.601
Household income:
    < $49,999 reference
    $50,000 - $99,999 -0.861 (0.424, -1.69-(-0.029)) 0.042
    $100,000+ -1.26 (0.446, -2.14-(-0.390)) 0.005
Education level:
    Less than a bachelor’s degree reference
    Bachelor’s degree -0.841 (0.349, -1.52-(-0.157)) 0.016
    Degree beyond a bachelor’s degree -1.08 (0.359, -1.79-(-0.379)) 0.003
Financial stress (no vs. yes) 2.29 (0.246, 1.81–2.78) <0.001
High-risk pregnancy (no vs. yes) 0.573 (0.290, 0.003–1.14) 0.049
Exercise routine changed (no vs. yes) 0.906 (0.246, 0.423–1.39) <0.001

Beta coefficients are provided with standard errors, 95% confidence intervals, and p-values for each variable included in the model.