>>Unless I am interpreting the data wrong, this indicates to me that the
>>values obtained for NCSP and RCSP by "inexperienced clinicians" is a
>>totally random process and not within 2 standard deviations of the mean 95%
>>of the time.
>
>>From my research of students measuring RCSP and NCSP and every other piece
>bar one I have read would totally agree with this statement Craig. I am
>sure Anne Marie Keenan would support this and show that experience
>examiners are only marginally better.
I am not really talking about measurement reliability/repeatability (which
is bad anyway), but about the shape of the curve when a large group of
people take a measurement.
If you take a large group of incompetent people (from any discipline) and
get to measure something (could be anything), even if the measurement
technique has poor relibility/repeatability, you should be able to assume
that there will be a bell shaped distribution clustered about what could
assumed to be a true mean. If very competent people are used, the mean
should still be the same, BUT the standard deviation a lot smaller.
Rather than a bell shaped distribution, I got a unifrom distribution
indicating that it was random. You can not interpret the mean of a uniform
distribution as having any relationship to the true population mean - there
just is not one.
Kind Regards
CP
______
\\_ _//
( @ @ )
****************************************o00o-()-o00o**********
Craig Payne email: [log in to unmask]
Department of Podiatry phone: (+61)(3) 9479 5820
School of Human Biosciences fax: (+61)(3) 9479 5784
Faculty of Health mobile: (+61)(0419) 103327
LaTrobe University
Bundoora, Vic 3083
Australia http://www.health.latrobe.edu.au/hs/schools/pod/home
*************************************************0ooo.********
.ooo0 ( )
( ) ) /
\ ( (_/
\_)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|