Skip to main content

Table 4 Item response rates for partner-specific sexual behaviours with most-recent sexual partner

From: 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

 

ITT analysis

AT analysis

 

PI arms

SI arms

Z

p

φ

PI arms

SI arms

Z

p

φ

Total number of respondents

154

 

131

    

181

 

104

    

 How first met a

  Known since childhood

28

18%

20

15%

   

32

17%

16

15%

   

  Through a mutual friend

4

3%

7

5%

   

5

2%

6

6%

   

  At work, school, university

47

31%

35

27%

   

53

29%

29

28%

   

  Online

1

1%

2

2%

   

1

1%

2

2%

   

  At a sporting event

0

0%

2

2%

   

1

1%

1

1%

   

  At a religious event

10

6%

13

10%

   

14

8%

9

9%

   

  At a friend/relatives’

4

3%

4

3%

   

5

3%

3

3%

   

  At a shebeen or club

0

0%

1

1%

   

1

1%

0

0%

   

  At the river

9

6%

6

5%

   

12

6%

3

3%

   

  On the street

17

11%

4

3%

   

17

9%

4

4%

   

  In town

12

8%

3

2%

   

13

7%

2

2%

   

  Other

16

10%

19

15%

   

20

11%

15

14%

   

  Declined to answer

6

4%

15

11%

5.92

0.01

0.14

7

4%

14

13%

8.91

0.003

0.18

 Relationship at last sex b

  Conjugal relationship

34

22%

20

15%

   

44

18%

10

6%

   

  Steady relationship

64

42%

34

26%

   

72

32%

26

17%

   

  Ex-steady relationship

47

31%

48

37%

   

54

30%

41

39%

   

  Known to one-another

2

1%

9

7%

   

4

2%

7

7%

   

  Not known to one-another

1

1%

2

2%

   

1

1%

2

2%

   

  Declined to answer

5

3%

18

14%

10.5

0.001

0.19

5

3%

18

17%

18.8

<0.001

0.26

 Still in a sexual relationship

  Yes

102

66%

83

63%

0.07

0.79

0.02

117

35%

68

65%

1.37

0.24

0.07

  Declined to answer

2

1%

10

8%

7.04

0.01

0.16

2

1%

10

10%

11.9

0.001

0.20

 Age difference of partner

  Median (IQR), women

4

[0–7]

4

[0–8]

0.03

0.85

0.01

4

[0–7]

3

[0–8]

0.09

0.76

0.02

  Median (IQR), men

−3

[−5.5–0]

−3

[−6–0]

0.29

0.59

0.03

−3

[−6–0]

-3

[−5.5–0]

0.03

0.87

0.01

  Declined to answer

13

8%

4

3%

3.25

0.07

0.11

14

8%

3

3%

2.45

0.12

0.09

 Partner a household member

  Yes

66

43%

57

44%

0.37

0.54

0.04

82

25%

41

39%

0.12

0.73

0.02

  Declined to answer

2

1%

10

8%

7.04

0.01

0.16

2

1%

10

10%

11.9

0.001

0.20

 Ever used a condom

  Yes

72

47%

58

44%

0.00

1.00

0.00

78

23%

52

50%

2.8

0.09

0.10

  Declined to answer

0

0%

7

5%

8.44

0.004

0.17

0

0%

7

7%

12.5

<0.001

0.21

 Frequency of condom use

  Never

21

14%

25

19%

   

24

12%

22

17%

   

  Sometimes

25

16%

16

12%

   

28

14%

13

10%

   

  Most of the time

26

17%

15

11%

   

26

10%

15

9%

   

  Always

72

47%

56

43%

3.33

0.34

0.11

78

43%

50

48%

2.52

0.47

0.09

  Declined to answer

0

0%

2

2%

2.52

0.11

0.09

0

0%

2

2%

3.05

0.08

0.10

 Condom use at first sex

  Yes

44

31%

44

34%

0.05

0.82

0.01

47

15%

41

39%

3.99

0.05

0.12

  Declined to answer

15

10%

8

6%

1.71

0.19

0.08

15

8%

8

8%

0.11

0.74

0.02

 Condom use at last sex

  Yes

42

29%

43

33%

0.44

0.51

0.04

48

15%

37

36%

2.25

0.13

0.09

  Declined to answer

13

8%

9

7%

0.37

0.54

0.04

13

7%

9

9%

0.12

0.72

0.02

  1. ITT Intention-to-treat, AT As-treated, PI personal interview arms (PAPI, CAPI), SI self-interview arms (CASI, ACASI), IQR inter-quartile range. Z test statistics compare PI to SI arms. In each pair the upper value is a \( {\chi}_1^2 \) comparison of the proportion of affirmative responses amongst valid responses for binary outcomes and a non-parametric Kruskall-Wallis test with k − 1 degrees of freedom for continuous and ordinal variables. The lower value is a comparison of item non-response rates using a \( {\chi}_1^2 \) test. φ is the effect size associated with the relationship between interview arm and the outcome of interest. a This question had not been asked in recent annual surveillance questionnaires. b This question had been asked in recent annual surveillance questionnaires, but the categories of responses were more precise in this trial