
Markku M. answered 02/04/20
PhD in Biostatistics with Years of Practical Analysis EXP
When we talk about generalizability, we mean out ability to apply the results of an experiment to the broader population.
The point of an experiment is to collect data on a sample from a population, analyze the data from that sample and apply the results to the population (with some a small degree of variability). So we want our results from a study to be as close to the values we would get if we could survey every single person in our population of interest. So results are more generalizable when the sample you are analyzing is every representative of the population of interest.