Table 1.
Uni- and multivariate (controlling for prevalence of circumcision) regression analyses of the relationship between the migration variables and country-level peak HIV prevalence.
| Univariate | Multivariate | |||||||
|---|---|---|---|---|---|---|---|---|
| Variable | No.a | No. (total) GHEs > medianb |
Coefficient | R2 | P-value | Coefficient | P-value | |
| Internal migration | ||||||||
| Internal migration rate | 51 | 1 (7) | −0.166 | 0.093 | 0.028 | −0.167 | 0.027 | |
| Percent internal migrants | 29 | 2 (4) | −0.145 | 0.149 | 0.038 | −0.150 | 0.033 | |
| Percent recent internal migrants | 28 | 2 (4) | −0.133 | 0.038 | 0.316 | −0.143 | 0.289 | |
| Recent urban migration (Women) | 28 | 9 (16) | 0.470 | 0.084 | 0.132 | 0.396 | 0.198 | |
| Recent urban migration (Men) | 22 | 7 (12) | 0.314 | 0.050 | 0.316 | 0.331 | 0.157 | |
| Crude migration intensity (5 year)d | 11 | 0 (0) | 0.010 | 0.002 | 0.878 | 0.001 | 0.980 | |
| Crude migration intensity (Lifetime) | 20 | 1 (4) | −0.124 | 0.025 | 0.503 | −.089 | 0.594 | |
| Courgeau’s k index (5 year) | 16 | 1 (1) | 4.13 | 0.149 | 0.140 | 4.27 | 0.115 | |
| Courgeau’s k index (Lifetime) | 15 | 0 (2) | −0.91 | 0.065 | 0.359 | −0.533 | 0.582 | |
| International migration | ||||||||
| Average annual net migration rate | 93 | 2 (13)c | 0.12 | 0.000 | 0.808 | 0.052 | 0.901 | |
| Immigration rate (1960) | 126 | 13 (19) | −0.004 | 0.000 | 0.950 | −0.003 | 0.966 | |
| Immigration rate (2005) | 141 | 5 (19) | −0.072 | 0.017 | 0.113 | −0.071 | 0.125 | |
| Emigration rate | 141 | 3 (19) | −0.162 | 0.042 | 0.014 | −0.165 | 0.015 | |
| International movement rate | 141 | 2 (19) | −0.109 | 0.046 | 0.010 | −0.108 | 0.012 | |
The column heading “No.” refers to the number of countries included in linear regression analyses. The observations were limited for most variables. See methodology section for definitions of the variables.
The column “No. (total) GHEs > median” refers to the number of countries with generalized HIV epidemics (GHEs) where the dependent variable is higher than the median for the dependent variable of the countries analyzed (total No. of countries with GHEs represented in the analysis).
Because this variable reflects the net migration rate, it ranges from −38 to +37. 2. Two out of the 13 countries with GHEs had migration rates below the .25 quantile or above the .75 quantile and this result is reported here.
Crude migration intensity is measured at the provincial or major regional level. The results for the municipal level were similar but are not shown.