Multiple Regression of Circa 2002 Urban HIV Prevalence on circa 1995 STI Prevalence and Other Socioeconomic Variables| Category | (1) Basic sample with syphilis | (2) Basic sample with gonorrhea | (3) Augmented sample with syphilis | (4) Augmented sample with gonorrhea |
| Age of epidemic (urban low) in 2002a | 0.177 (4.07)*** | 0.131 (2.00)* | 0.12 (2.90)*** | 0.08 (1.70)* |
| Per capita national income, median 1996-2001 | -1.344 (2.82)*** | -1.177 (2.30)** | -0.929 (3.63)*** | -0.949 (4.45)*** |
| Income equality, Gini index, 1990s | 0.125 (0.04) | 3.186 (0.83) | 5.349 (2.77)*** | 5.258 (2.79)*** |
| Percentage of the population that is Muslim, 1999 | -0.026 (4.05)*** | -0.031 (3.50)*** | -0.015 (3.45)*** | -0.02 (4.10)*** |
| Urban male-to-female ratio for those age 20-39, 1999 | 6.178 (2.03)** | 6.199 (1.40) | 3.27 (1.29) | 3.038 (1.19) |
| Urban high-risk population dummy (1 if yes, 0 is no) | 1.17 (2.83)*** | 0.564 (0.85) | 0.894 (2.64)*** | 0.37 (0.71) |
| Logit syphilis in low-risk group, 1995b | 0.31 (2.46)** | | 0.233 (1.78)* | |
| Logit gonorrhea in low-risk group, 1995b | | 0.502 (1.92)* | | 0.479 (1.86)* |
| Constant | 0.297 (0.05) | -0.791 (0.11) | -2.117 (0.63) | 0.054 (0.02) |
| Number of observations | 56 | 38 | 181 | 180 |
| Number of countries | 40 | 29 | 101 | 100 |
| R2 | 0.57 | 0.56 | 0.5 | 0.51 |
Source: Authors' calculations. Syphilis and gonorrhea prevalence in 1995 (George Schmid, personal communication, April 5, 2004). HIV prevalence circa 2002 is from the urban low-risk tables in the U.S. Bureau of Census database on HIV prevalence. Other variables are from World Bank data.
* = significant at 10 percent; ** = significant at 5 percent; *** = significant at 1 percent.
Note: The figures in parentheses are robust t-statistics.
a. Age of epidemic is defined as the number of years since the first case of HIV/AIDS was reported.
b. The logarithm of the ratio of the prevalence to 1 minus the prevalence of the given STI for the low risk population in 1995.