Skip to main content

Table 1 Socio-demographic responses by ZEST study population

From: The influence of interviewers on survey responses among female sex workers in Zambia

 

Univariate

Bivariate regression b

N

% / Median [IQR]

ICCc

Male

IVR

Female

IVR

p-value

Study site

965

     

 Livingstone

 

49.7%

    

 Kapiri

 

25.4%

    

 Chirundu

 

24.9%

    

Age

965

25 [21–30]

3.9%

26.2

26.3

0.99

Ever married

965

69.9%

12.5%

68.7%

75.5%

0.69

Currently divorced/separated

965

24.5%

10.4%

26.8%

18.4%

0.55

Education

965

    

0.65

 No formal education

 

11.2%

 

7.3%

11.8%

 

 Primary (up to 9 years)

 

46.6%

 

44.6%

53.0%

 

  > 9 years

 

42.2%

 

48.1%

35.2%

 

Able to read and write

959

75.3%

9.1%

82.8%

67.7%

0.06

Monthly income (ZMW)

949

    

0.38

 No income

 

21.3%

 

2.6%

24.8%

 

  < 250 kwacha

 

13.0%

 

9.9%

38.4%

 

 251–500 kwacha

 

24.8%

 

39.1%

29.6%

 

 501–1000 kwacha

 

25.9%

 

37.4%

6.2%

 

 1001–1500 kwacha

 

7.7%

 

6.9%

0.6%

 

  > 1500 kwacha

 

7.4%

 

4.2%

0.4%

 

Financial situation

962

    

0.020

 Very poor

 

14.7%

 

4.8%

24.1%

 

 Poor

 

37.8%

 

32.8%

54.1%

 

 Just getting by

 

36.3%

 

46.5%

18.9%

 

 Comfortable

 

10.0%

 

14.2%

2.6%

 

 Very comfortable

 

1.2%

 

1.8%

0.3%

 

Mobile phone ownership

965

85.0%

6.3%

86.4%

85.2%

0.94

Self-perceived relative SES a

965

3 [2–5]

 

3.1

3.0

0.94

Any income from non-sex-work

965

30.2%

7.5%

30.6%

27.3%

0.87

  1. ICC Intraclass Correlation Coefficient, IQR inter-quartile range, IVR interviewer, SES socio-economic status, ZMW Zambian Kwacha: 1 Kwacha ~US$ 10
  2. a10-point scale
  3. bAll bivariate regressions included study site fixed effects and interviewer random intercepts. Values for male and female IVR are marginal predicted values based on regression coefficients. c ICC is the proportion of all variance in a model without interviewer gender attributable to variation in interviewer identity; not available for Poisson or ordered logistic models. P-value is for a χ2 test, adjusted for multiple testing across all results shown in Tables 1, 2, 3 and 4 using the Benjamini-Hochberg method. P-values <0.05 shown in bold