Rushing in where angles fear to tread......
you mean angEls of the halo type?
The reported/calculated p-values are valid (if at all) _only_ inasmuch as the core assumptions regarding data are maintained. Critical evaluation of the validity of those assumptions spell the difference between simplistic and useful applications of p-values.
rather than fuss too much about p values ... perhaps the more crucial matters are:
1. how useful REALLY ... is null hypothesis testing? ... as we generally practice it
2. why haven't we focused much more systematically on REPLICATION studies?
as to the first item ... take correlations ... when we form the typical null hypothesis of rho being 0 in the target population ... and if we are interested in the relationship between ability and achievement (or lots of other cases) ... aren't we wasting our time testing such a null? i find most of the nulls that we routinely test ... to be rather silly and again, we know they are not true
the important questions are: what IS the correlation likely to be ... not what it isn't ... or what kinds of effects do we get? not that they are NOT 0
as to the second item ... we have for far too long ... made an assumption that if our p values are good in our studies ... that this means that if we did this again ... we would find this same kind of result ... but p values are NOT EVIDENCE FOR REPLICATIONS ...
if there are important treatments we want to evaluate ... then we need to forget p values and, replicate (or at least try) the treatments a number of times ... with different Ss in different contexts ... to see if we consistently get the same trend of results
replication studies and results would go so much further towards our understanding of important effects but ... academic rewards are rarely garnered for this ... in fact, we systematically discourage such research when it comes to p and t
too bad
. . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
