I am looking for a few peer-reviewed references (other than Israel, 1992) that support the argument that sample size determination is not necessary in situations where a research study population is the full population (e.g. in some record-linkage studies). Any help would be greatly appreciated.
Your statement is self-evident, and so well known in the scientific and statistical community, that you need no reference.
A related post appears at https://www.researchgate.net/post/What_conditions_justify_studying_the_entire_population_instead_of_selecting_a_sample
The error of estimation increases with time i.e. the longer the interval the higher the error
Ian Kennedy - agreed, except when a journal asks for one nonetheless!
State that the population size is N and that the sample size is n, where n=N.
Perhaps the journal referee does not fully understand the meaning of the word full!
- Badges
- Science topic
- Similar topics
- Correction