Skip to main content
. 2022 Nov 21;22:463. doi: 10.1186/s12905-022-02054-9

Table 5.

Difference-in-differences models for SABAS scores and abortion knowledge

Variable Coefficient Std. Error 95% Confidence Interval p-value
Model 1: Outcome = Total SABAS Score
 DID (Treatment*Time)a −9.33 1.49 −12.25 −6.42 < 0.001
Adjustment Covariates
  Treatment (Intervention = 1) 1.40 1.06 − 0.67 3.48 0.185
  Time (Endline = 1) 10.04 1.77 6.58 13.51 < 0.001
  Any school −5.04 0.89 −6.78 −3.29 < 0.001
  Any school*Timeb −4.59 1.26 −7.05 −2.12 < 0.001
  Orthodox Christian 3.99 1.09 1.84 6.14 < 0.001
  Orthodox Christian*Timeb −2.01 1.48 −4.92 0.89 0.175
  Knowledge of any indications of legal abortion −3.23 1.02 −5.23 −1.23 0.002
  Knowledge of any indications of legal abortion*Timeb −3.46 1.32 −6.05 −0.86 0.009
  Attended any safe abortion eventsc 1.53 0.75 0.07 2.99 0.040
Model 2: Outcome = Knowledge of one or more indications of legal abortion
 DID (Treatment*Time)a 0.26 0.06 0.15 0.37 < 0.001
Adjustment Covariates
  Treatment (Intervention = 1) −0.14 0.04 −0.22 − 0.06 < 0.001
  Time (Endline = 1) 0.32 0.06 0.19 0.44 < 0.001
  Any school 0.15 0.03 0.09 0.21 < 0.001
  Any school*Timec −0.02 0.05 −0.11 0.07 0.672
  Orthodox Christian −0.08 0.04 −0.16 0.002 0.056
  Orthodox Christian*Timeb 0.02 0.06 −0.09 0.13 0.764
  Attended any safe abortion eventsc 0.21 0.03 0.15 0.26 < 0.001

aDID is the difference-in-differences variable, an interaction term of treatment group (Boset vs. Fogera) and time (baseline vs. endline); 1 = intervention group at endline

bTime-varying covariates with time-varying effects on the outcome are adjusted for using an interaction term between the covariate and time

cTime-varying covariates with time-invariant effects on the outcome are adjusted for by including the term in the model