Paradata analyses to inform population-based survey capture of pregnancy outcomes: EN-INDEPTH study

Table 2 Question characteristics associated with answer correction probabilities and frequencies

Characteristics	Model (1)						Model (2)
Characteristics	Part 1 correction (yes/no)			Part 2 number of corrections			Part 1 correction (yes/no)			Part 2 number of corrections
Variables	Coef	Robust Std. Err.	p value	Coef	Robust Std. Err.	p value	Coef	Robust Std. Err.	p value	Coef	Robust Std. Err.	p value
Type, Ref (single)
Numerical computational	0.056	0.012	< 0.001	0.050	0.007	< 0.001	-0.138	0.014	< 0.001	0.026	0.008	0.001
Date-related	− 0.531	0.025	< 0.001	0.055	0.013	< 0.001	-0.411	0.028	< 0.001	0.072	0.015	< 0.001
Free-text	0.850	0.146	< 0.001	0.031	0.061	0.611	0.864	0.164	< 0.001	0.041	0.075	0.581
Multi-select	2.248	0.145	< 0.001	0.590	0.118	< 0.001	2.337	0.158	< 0.001	0.598	0.123	< 0.001
Nature, Ref (regular)
Death-related	− 0.166	0.009	< 0.001	0.079	0.004	< 0.001	-0.187	0.011	< 0.001	0.078	0.005	< 0.001
TOP-related	0.205	0.015	< 0.001	0.040	0.007	< 0.001	0.316	0.017	< 0.001	0.055	0.008	< 0.001
Structure, Ref (no notification)
Yes, with notification	0.400	0.012	< 0.001	− 0.041	0.007	< 0.001	0.486	0.014	< 0.001	− 0.026	0.008	0.001
Module, Ref (FPH)
FBH+	0.015	0.009	0.101	− 0.003	0.003	0.248	0.012	0.009	0.169	− 0.003	0.003	0.333
Constant	− 2.661	0.007	< 0.001	0.125	0.002	< 0.001	-2.818	0.056	< 0.001	0.1465	0.016	< 0.001
Observations	3,340,189			250,608			2,247,142			152582
R-squared	0.004			0.004			0.014			0.014
Root MSE	0.263			0.511			0.250			0.494

Model (1) unadjusted, model (2) adjusted for respondent’s characteristics (age, education, parity, wealth quintile) and location, with interviewer clustering accounted as a fixed effect. Both models accounted for clustering of individual responses within individual women (interview)
FPH full pregnancy history module, FBH+ full birth history module with additional questions on pregnancy losses

ISSN: 1478-7954