Good question. The most common reason for oversampling a particular group is to be able to make claims particular to that group. The reality is that, in this case, the oversampling ended up with about 50 extra black respondents, which is too few to skew the overall data set, but enough to be able to make claims about African-Americans.
You usually have to oversample minorities because if you did just a simple random sample you wouldn't end up with a representative amount of them. For instance, African Americans are roughly 12 or 13% of the population. In a random sample, they wouldn't come out as 12 or 13% of the sample. So, you over sample them to make the sample more representative of the population. I'm just adding on to what J. Morgan said.
2 Comments:
Good question. The most common reason for oversampling a particular group is to be able to make claims particular to that group. The reality is that, in this case, the oversampling ended up with about 50 extra black respondents, which is too few to skew the overall data set, but enough to be able to make claims about African-Americans.
You usually have to oversample minorities because if you did just a simple random sample you wouldn't end up with a representative amount of them. For instance, African Americans are roughly 12 or 13% of the population. In a random sample, they wouldn't come out as 12 or 13% of the sample. So, you over sample them to make the sample more representative of the population. I'm just adding on to what J. Morgan said.
Post a Comment
<< Home