I have a question concerning two different samples, with the first being more uniform that the second.
a) Chance errors are likely to be smaller...
- using the first set of subjects
- using the second set of subjects
- no way to tell
b) Bias is likely to be smaller...
- using the first set of subjects
- using the second set of subjects
- no way to tell
What I think:
When our samples are more uniform, we compromise representativeness. For example, if we use a random sample, that sample tends to be more representative of the whole population, and thus less uniform than, say, a sample of a group of your friends. So, I think a) is using the first set of subjects (the more uniform set), and b) is using the second set of subjects (the less uniform set). Can someone confirm if my justifications are correct? Thanks!