|
|
|
|
|
|
by dredmorbius
4541 days ago
|
|
|
Size of studies isn't what influences their accuracy, it's the sampling method and methodology. You can attain statistically significant results with an n of less than 20. Jacob Nielsen is a big fan of small-number user testing, often of as 3-5 users: http://www.nngroup.com/articles/why-you-only-need-to-test-wi... And if you are interested in the sampling error, that's reduced with the square root of n. So in order to halve the error, you've got to square the sampling size. That increases costs markedly (and is why sample-based statistical methods are so powerful -- they work, if done correctly). |
|
|