Hi, can any one guide how could we calculate national figure for a health statistic where subnational figures are available but no national figure is reported?
Hi. The guiding principle here tends to be weighted averaging, supplemented as needed with inference for missing subareas. So for example, suppose you have % of population aged 6-18 in each of a country's 6 subareas. You could just multiply each of those 6 percentages times the subarea population, sum the results, then divide by the sum of the population in the 6 subareas. What if the country has 7 subareas and there are no data on the age distribution in the 7th? Then you need to use the data from the subareas you have to estimate for the missing subarea. A first approximation is the average for the other subareas. But often one can do better. If you have the data for 96 out of 100 communities, for example, a better approximation will come from communities with similar population size and rurality, preferably in the same part of the country (eg, the mountains, along the coast). Perhaps you can even run a regression using 9 things you already know about all 100 cities to predict the unknown characteristic for the cities that lack data. Hope this helps. Feel free to correspond back with specifics. :-)
- Badges
- Science topic
- Similar topics
- Psychology
- Emotion