Skip to main content
. Author manuscript; available in PMC: 2016 Jan 31.
Published in final edited form as: Demography. 2015 Feb;52(1):259–280. doi: 10.1007/s13524-014-0364-z

Table 4.

Nonlinear two-stage estimator: Second-stage estimates for HIV-negative men

Divorced After 2004 Number of Sexual Partners
in Year of Follow-up (continuous)
Increase in Condom Use
With Spouse/Partner
After 2004



(1)
Probit
(2)
IV Probit
(3)
OLS
(4)
IV Regression
(5)
Probit
(6)
IV Probit
Learned HIV Status 0.015 (0.013) 0.025 (0.034) −0.063 (0.052) −0.030 (0.105) 0.066* (0.033) 0.033 (0.074)
Age in 2004 −0.001* (0.001) −0.002* (0.001) −0.002 (0.002) −0.002 (0.002) −0.006** (0.001) −0.006** (0.001)
Region (relative to Mchinji, central)
  Rumphi, north −0.003 (0.018) −0.002 (0.019) 0.188* (0.072) 0.192** (0.072) 0.103* (0.044) 0.093* (0.039)
  Balaka, south 0.004 (0.016) 0.004 (0.017) 0.130 (0.067) 0.130 (0.067) 0.038
(0.036)
0.037
(0.034)
Education (relative to no education)
  Primary 0.021 (0.016) 0.025 (0.021) −0.059 (0.092) −0.057 (0.091) −0.051 (0.050) −0.052 (0.046)
  Secondary 0.020 (0.029) 0.020 (0.024) −0.116 (0.124) −0.111 (0.125) −0.027 (0.062) −0.036 (0.070)
Observations 775 775 750 750 595 595

Notes: Robust standard errors clustered by village are shown in parentheses. Marginal effects are reported for all binary outcomes; OLS coefficients are reported for the continuous outcome (Model 4). For Model 4, the value of Wooldridge’s robust score test of endogeneity is 0.159, with a p value of .6910.

p < .10;

*

p < .05;

**

p < .01