At 11:32 PM 11/17/2002, Donald Burrill wrote:
I usually added, in introducing a C.I. in the first place, that sometimes one has an obvious null hypothesis (that is, an obvious null- hypothetical value of a parameter) to test, and in that circumstance an hypothesis test is clearly appropriate. But sometimes there isn't an obvious value to specify for mu (or sigma, or rho, or beta, or ...), and then one might be interested in knowing what (potential) values of mu (or whatever) would be consistent with the data in hand. Is this any help? -- Don.
and i contend that the above ... the last sentence ... is REALLY what we want to infer about ... what mu or sigma or rho ... or beta might be ... NOT what the null isn't
being in this business far too long ... i have yet to see a null, in theory, that was really worth testing (and as herman rubin has oft argued ... no null is really true anyway) ... but, day in and day out, i see questions of the form: i wonder what the mean in the population might be ... or, i wonder what the correlation between X and Y is in the population ... to be commonplace and sensible
while a CI can be used to test a hypothesis ... IMHO ... THE value of a CI is not that ...
of course, i don't see much (if any) value to typical null hypothesis testing either ... we place FAR too much emphasis on it during instruction ... and, far too much emphasis on it in our research publications
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
