which one is more accurate to represent an variable of populations?
How to calculate an unweighted variables?
There is no unique answer to this question, it depends on situation.
For which analysis do you use the population variables? What do you consider under weighting?
Population indicators could be divided into an absolute (eg. population size, number of deaths, number of immigrants etc.) and relative (eg. mortality rate, immigration rate, population change rate...).
If you have big differences in population sizes, the results will be influenced by it. For example, in the USA dies 3 million people per year and in Lesotho 30 thousands. But it doesn't mean that people more often die in the USA than in Lesotho. If we normalize absolute numbers of death by total population (USA 300 mil, Lesotho 2 mil), we'll see that people in Lesotho die twice as likely than in the USA. For this situation, it is better to use relative (unweighted) indicator.
Sometimes the population size is an important factor, as for the calculation of demographic resources. If we have a very small settlement with 10 people, it could happen that there no one dies in the year, thus the relative mortality rate is 0. A bigger town with 200 000 people surely will have mortality rate of 10 per thousand people. If we think about demographic prosperity of these settlements, it is obvious that the bigger town has a better future. The small settlement is likely to fade away in near future, despite it had zero mortality in the year. For this situation, it is recommended to weight indicators by the population size.
If you work some survey on a disproportional sample (80% males and 20% females), you'll have to divide the females' answers by the number of females in the survey or multiply it with the factor of 4 to achieve more reliable results (50 : 50).
- Badges
- Science topic
- Similar topics
- Applied Mathematics
- Calculations