What is the difference between Rim Weighting, sample balancing by iterative proportional fitting, data weighting and ranking? Actually there isn’t anything different, all these terms mean more or less the same thing.
So, what do all these different words describe?
Imagine that you’ve done a survey in a population that is composed of 50 males and 50 females. You’re trying to study it accurately, so you’re looking for the same composition in your sample. As it turns out fewer men responded to your survey. Instead of having achieved the perfect 50/50 balance: You only have 40 males.
To acquire balanced and representative results for your sample, you’ll have to correct the mismatch between sample and population. Rim Weighting is the method for you! It compensates for the mismatch. You calculate a weight factor for the males giving them a higher weight in the sample than one. By doing this the weighted results match the composition of the population.
When you’re actually doing research only controlling for gender might not be the most prudent thing to do. You’ll also want to control for other variables such as age, household size or income and region. Even though these demographic variables may be interesting for your research, the most important variables to control for are whether or not the persons in your sample have come into contact with your product or service.
Our tool at sampleweighting.com provides you with an easy-to-use solution for performing Rim Weighting across many variables. You’ll easily select the necessary variables and set the preferred targets to achieve a representative sample. If you’re looking for simple solutions and an intuitive interface our tool is the thing for you. We’re sure you’re going to have fun using it and that your samples will be balanced and fair.