Re: Really Dumb Question...

Not just in your opinion.

A fantastic resource for this issue (and power and sample size in general,
with online calculators) is

http://www.stat.uiowa.edu/~rlenth/Power/index.html
(Russ Lenth's power and sample-size page).

I strongly recommend it to my students. It does not take too much time to
read the references that are downloadable as PDF (so you can even do without
the author's main article in TAS), and they are not mathematically demanding
at all (so perfect not only for the physicians I teach in the grad courses,
but also for me :)

I tell them to try whichever of the online caluclators and chooose
Options -> Post Hoc Power ...

and select either "Yes" or "No" when asked whether the test was
"significant" -- and see what Prof. Lenth tells them ...

.... and they should never ask such a question again :)

Regards,

Gaj Vidmar, PhD
Univ. of Ljubljana, Fac. of Medicine, Inst. of Biomedical Informatics
http://www.mf.uni-lj.si/ibmi-english [/biostat-center]

"Kevin E. Thorpe" <kevin.thorpe@xxxxxxxxxxx> wrote in message
news:1181820378.042622.243250@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
On Jun 13, 5:50 pm, mcap <mca...@xxxxxxxxxxx> wrote:
Dear All:

I know this is not a great question but I am curious. Here
goes.......

If power is the ability to detect a significant difference given
that one exits, then why is it important to evaluate it in studies
where a significant difference was actually found (null was
rejected). Is it because the rejection was based on the sample data
but but the hypothesis refers to a theoretical population where the
null is not true - two different things?

I know there is a bias towards type I errors in publication and
not enough attention paid to type II errors......

Thanks,
Marc

What you are describing is sometimes called "post-hoc power."
In my opinion, it is a pointless exercise no matter what
the study outcome.


.

Relevant Pages

  • Re: OT - Whats in your CD player / on your iPOD?
    ... the population or less to have any power at all. ... Majority rules in the US. ... opinion. ... to assist in making a decision that could affect the lives of any ...
    (rec.roller-coaster)
  • Re: Collapse of Left and Right, Rise of Populism
    ... Bert, you are allowing your own cynicism to give power to the very ... Your opinion, and the opinion of EVERY CITIZEN is of value, of ... You can speak just for yourself or you can ...
    (alt.politics)
  • Re: Benazir Bhutto - the reality.
    ... Difference with my 'opinion' is that I've been saying it was ... manages best a transition to some other power source. ... The Scottish Government announced they would not allow ANY new nuclear power stations on Scottish soil and the Crown Estate service announced, yesterday, that an undersea power line to carry renewable and sustainable power from Scotland to London, England and even the main continent of Europe was a feasible project. ... Now that we have our own parliament which is now also ready to buck Westminster, and do what we Scots do so very well, we should soon get ahead. ...
    (soc.culture.scottish)
  • Re: Kicking the Dragon
    ... Hey idiot, Larken was not jailed for expressing an opinion. ... Larken Rose spent a year in prison, not for hurting anyone, not for lying about anything, not for hiding anything, not for cheating or stealing, but simply for voicing an opinion which those in positions of power didn't approve of. ... a corrupt and dishonest government bent on harassing, demonizing, persecuting, silencing and imprisoning those ...
    (misc.taxes)
  • solely yearly furnished
    ... The pleasure of the great is the power to make people happy. ... Thus opinion is the queen of the world, ... Sound opinions of the people.--Civil wars are the greatest of evils. ... You can apply it to God, ...
    (sci.crypt)