Skip to main content

Table 2 Regression results for questions with known truth

From: List randomization for eliciting HIV status and sexual behaviors in rural KwaZulu-Natal, South Africa: a randomized experiment using known true values for validation

Question / dependent variable

“I am HIV negative”

“I refused the AHRI DSS HIV test this year”

Subpopulation

Full population

HIV status (+/−) known to researchers

HIV (+) status known to individual

Full population

 

(1)

(2)

(3)

(4)

Coefficient:

Constant

0.67 (−0.57, 1.90)

0.82 (− 0.62, 2.27)

−1.36 (−4.73, 2.01)

−0.55 (−1.92, 0.83)

HIV negative

0.06 (−0.40, 0.52)

0.01 (− 0.45, 0.48)

 

0.43 (− 0.08, 0.93)

HIV test refused

0.08 (− 0.40, 0.55)

  

0.54** (0.03, 1.06)

Male

0.02 (−0.31, 0.34)

−0.05 (− 0.40, 0.30)

−0.16 (− 0.67, 0.36)

−0.08 (− 0.46, 0.31)

Age

− 0.01 (− 0.06, 0.03)

−0.02 (− 0.08, 0.03)

0.04 (− 0.13, 0.20)

0.05* (− 0.01, 0.10)

Age2†

0.02 (− 0.04, 0.07)

0.03 (− 0.03, 0.09)

−0.02 (− 0.21, 0.17)

−0.04 (− 0.10, 0.02)

Years of education

0.02 (− 0.04, 0.07)

0.02 (− 0.04, 0.08)

0.09** (0.01, 0.17)

−0.02 (− 0.08, 0.04)

Residual SE

0.84

0.83

0.75

0.98

Observations

Arm A

242

209

77

245

Arm B

210

176

76

210

Total

452

385

153

455

  1. Coefficients are shown as point estimate (95% confidence interval). * p < .10, ** p < .05. Correlation with non-sensitive item component not shown. Residual SE (standard error) is shown for model fit and ranges from 0 to 1, where a residual SE of 0 is a perfectly fit model. coefficient multiplied by 100
  2. Results are shown for the linear estimation model