The impact of self-interviews on response patterns for sensitive topics: a randomized trial of electronic delivery methods for a sexual behaviour questionnaire in rural South Africa

Table 1 Respondent characteristics by response status and intention-to-treat arm

	PAPI	CAPI	CASI	ACASI	Z	p ^†	Total	%	Z	p ^‡
A. Full sample (n = 504)
Gender
Male	61	62	65	64			252
Female	65	63	62	62	0.24	0.97	252
Age
18–29	42	42	39	39			162
30–49	41	41	42	44			168
≥ 50	43	42	46	43	0.56	1.00	174
Location
Urban	33	31	30	32			126
Peri-urban	31	30	34	31			126
Rural	62	64	63	63	0.46	1.00	252
B. Contacted sample (n = 355)
Gender
Male	49	46	50	45			190	0.75
Female	44	43	46	42	0.02	1.00	175	0.69	2.24	0.14
Age
18–29	24	25	29	25			103	0.64
30–49	33	29	30	29			121	0.72
≥ 50	36	35	37	33	0.62	1.00	141	0.81	12.8	0.002
Location
Urban	22	20	14	18			74	0.59
Peri-urban	21	26	28	23			98	0.78
Rural	50	43	54	46	3.86	0.70	193	0.77	15.8	<0.001
C. Interviewed sample (n = 340)
Gender
Male	48	44	45	44			181	0.95
Female	42	40	41	36	0.15	0.99	159	0.91	2.77	0.10
Age
18–29	24	24	25	23			96	0.93
30–49	31	28	26	26			111	0.92
≥ 50	35	32	35	31	0.45	1.00	133	0.94	0.69	0.71
Location
Urban	21	19	13	16			69	0.93
Peri-urban	20	24	24	22			90	0.92
Rural	49	41	49	42	3.20	0.78	181	0.94	0.39	0.82

PAPI Paper and pen interview, CAPI computer-assisted self-interview, CASI computer-assisted personal interview, ACASI audio computer-assisted personal interview. All non-test values are counts
^† \( {\chi}_3^2 \) tests for gender, \( {\chi}_6^2 \) tests otherwise, for difference across arms and stratification cells by each stratifying variable
^‡ \( {\chi}_1^2 \) tests for gender, \( {\chi}_2^2 \) tests otherwise, for difference in proportion of allocated individuals being contacted (panel B) and the proportion of contacted individuals interviewing (panel C) across levels of each stratifying variable

ISSN: 1471-2288