If you are indeed interested in both positive and negative differences
between the two groups, I would:
- report the stats for the F test with FWE = 0.05
- report which areas are positive and which negative by computing the T
tests corresponding to the F test, one the negative of the other (e.g.,
T tests = [1 1 -1 -1] and [-1 -1 1 1]), using FWE = 0.025.
Please somebody correct me if I'm wrong.
Bruno
MCLAREN, Donald wrote:
> To get the two-tailed p-values simply multiply the p-value by 2,
> although I'm not sure about doing this with FWE corrected values. By
> default, I don't think there is a way to get SPM to do this unless you
> use and F-test at least for simple models.
>
> The issue raised about two-tailed tests isn't necessarily valid. For
> example, if I have a hypothesis that the hippocampus is smaller in A
> compared to B AND the visual cortex will be bigger in group A than B,
> then I don't need to divide p<.05 by 2 because both would be considered
> one-sided tests. The fact that they are in different directions doesn't
> justify the use of a two-tailed test. Two-tailed tests should be used if
> you don't have an a priori hypothesis about the regions you are
> investigating (e.g. what regions are different without considering the
> direction of change).
>
> Best Regards, Donald McLaren
> =================
> D.G. McLaren
> University of Wisconsin - Madison
> Neuroscience Training Program
> Office: (608) 520-0586
> =====================
> This e-mail contains CONFIDENTIAL INFORMATION which may contain
> PROTECTED HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and
> which is intended only for the use of the individual or entity named
> above. If the reader of the e-mail is not the intended recipient or the
> employee or agent responsible for delivering it to the intended
> recipient, you are hereby notified that you are in possession of
> confidential and privileged information. Any unauthorized use,
> disclosure, copying or the taking of any action in reliance on the
> contents of this information is strictly prohibited and may be unlawful.
> If you have received this e-mail unintentionally, please immediately
> notify the sender via telephone at (608) 520-0586 or email.
>
>
> On Wed, Mar 17, 2010 at 3:14 PM, Romy <[log in to unmask]
> <mailto:[log in to unmask]>> wrote:
>
> Hello,
>
> I have a question about a comment I recently got from a journal
> reviewer:
>
> 'The cross-sectional comparison is conducted at a threshold of p<0.05,
> corrected for multiple comparisons. However, this threshold is
> applied for
> both patients > controls and then again for controls > patients.
> Thus, the
> test is two-tailed and so thresholds must be halved (i.e. p<0.025)
> for each
> direction of the test.'
>
> The trouble (if I understand correctly) is that I have FWE-corrected p
> values listed in a table for areas of both decreased and increased gray
> matter. These are all significant at levels of p<0.002 and under,
> but I'm
> not sure if it is correct to report these values in this way because
> they
> reflect one-tailed tests (defined by the relevant contrast), when I was
> actually looking for differences in either direction. If this is
> indeed a
> problem, how can SPM generate p values for a two-tailed comparison?
>
> Thanks a lot in advance!
>
> Kind regards,
> Romy
>
>
|