Although I have taught statistics for almost 20 years, I still don't understand
the relevance of directional alternate hypotheses. The critical region of the
test statistic is determined by the statistic's distribution given that the null
hypothesis is true. This conditional distribution has little to do with the
investigator's opinion of which direction it should (or more accurately, could)
be false.
--
"Karl L. Wuensch" wrote:
> I have referred to this procedure as being a "half-tailed test." The
> one-tailed probability for this application of F (and for the typical
> applications of Chi-Square as well) are, indeed, nondirectional, since the
> numerator is representing the effect as a squared quantity. Consider your
> guessing the direction of effect as event A and your getting a sample as
> discrepent with a nondirectional null as event B. These events should be
> independent, so, applying the multiplication rule under independence, the
> directional p is .5*P(B). Taking this a bit further, if you had three
> groups and could convince us that you predicted the ordering of the means of
> the three, a one-sixth tailed probability would be appropriate.
> ----- Original Message -----
> From: "jim clark" <[EMAIL PROTECTED]>
> To: "TIPS" <[EMAIL PROTECTED]>
> Sent: Friday, March 02, 2001 8:33 PM
> Subject: Re: Question about one-tailed tests
>
> > Another term that is sometimes used is directional vs.
> > non-directional to refer to the alternative hypothesis. You can
> > also think of the F distribution as the t-test folded over so
> > that both the "negative" and "positive" tails are in the upper
> > end of the F distribution. In any case you are correct that the
> > p for F can be halved to get the one-tail probability. This
> > causes no end of consternation when teaching the use of tables,
> > because it means that you have to use area=.10 to do a
> > directional test with F when alpha=.05 (i.e., alpha = area/2).
> > Although you are correct on this issue, I am reluctant to say
> > definitively whether or not you are off your rocker in a more
> > general sense because (1) you don't provide enough information
> > about yourself, (2) I don't have your MMPI or PAI profile
> > available, and (3) I'm not a clinical psychologist
> > anyway. However, your postings over the years haven't shown any
> > obvious delusional signs.
> >
> > Best wishes
> > Jim
> >
> >
> ============================================================================
> > James M. Clark (204) 786-9757
> > Department of Psychology (204) 774-4134 Fax
> > University of Winnipeg 4L05D
> > Winnipeg, Manitoba R3B 2E9 [EMAIL PROTECTED]
> > CANADA http://www.uwinnipeg.ca/~clark
> >
> ============================================================================
> >
*****************************************************************
* Mike Scoles * [EMAIL PROTECTED] *
* Department of Psychology * voice: (501) 450-5418 *
* University of Central Arkansas * fax: (501) 450-5424 *
* Conway, AR 72035-0001 * *
********* http://www.coe.uca.edu/psych/scoles/index.html ********
