Hi everyone,
Concerning one-tailed tests, I like the following idea, which I ran
across somewhere I cannot remember where. One should only use a one-tailed
test when one is making a prediction the opposite of which would not be
interesting or useful. For example, I am trying to decide whether to use an
expensive plant food, and I do a study to see whether my plant food helps my
plants grow. This would be a one-tailed test because, for my purposes, my
decision will be the same if the plants are not benefited, i.e., not to
continue using the fertilizer. While it may be interesting if the
fertilizer actually hinders growth, my decision in this case would be the
same as if it merely doesn't promote growth, not to use it. I like this as
a "hard-nosed" requirement for a one-tail test to be used.
Joe Hatcher
Ripon College
Ripon, WI USA
> ----------
>
