Skip to main content
. 2013 Jun 17;5(4):217–227. doi: 10.5539/gjhs.v5n4p217

Table 3.

Comparison of performance of C values derived applying different strategies

Estimation method Correlation of e and ê*(eligible groups) Correlation of e and ê (ineligible groups) MSE PMAB NMAB Slope (SE)**
All reference groups (national) C= 239 0.60 Not applicable 216780 2.34 0.48 0.63 (0.10)
All reference groups (average of province estimates) C= 248 0.60 Not applicable 225674 2.20 0.47 0.60 (0.09)
Eligible groups via regression method C= 380 0.97 0.06 25422 1.28 0.54 0.79 (0.06)
Eligible groups via ratio method C= 308 0.95 0.63 13884 1.22 0.78 0.91 (0.06)
*

The correlation between real (e) and estimated size (ê) of reference groups is the same as correlation between prevalence of reference groups in the community with average people known by respondents in each

**

slope of regression through origin line of predicted versus real sizes