calculating sample size

[ravi rohilla]

For a thesis study regarding Diabetes using scale "Diabetes Risk Scale" how to calculate the sample size so as to validate the study. Kindly please help.

--

Ravi

Dr Ravi Rohilla
Junior Resident, Community Medicine
Postgraduate Institute of Medical Sciences
Rohtak (Haryana) - 124001
Mo: +91 93153 80656

[Mehwish Hussain]

Dear Dr. Rohilla,

Sample size calculation is strictly depends on objective of the study. It also requires certain level of confident about the outcome of the study and margin of error or power of the test for rejecting the null hypothesis. Also, type of study and population size vary the sample size. You will have to consider these factors first.

--
To post a new thread to MedStats, send email to MedS...@googlegroups.com .
MedStats' home page is http://groups.google.com/group/MedStats .
Rules: http://groups.google.com/group/MedStats/web/medstats-rules

--
Regards

Mehwish Hussain, PhD*
Senior Lecturer (DUHS)
Coordinator, MSBE Program (DUHS)
Manager, ORIC (HEC)
Pakistan

[ציפי שוחט‎]

Dear Dr. Rohilla,

It seems to me that you may be asking about Post-Hoc Sample Size. I.e you are trying to judge a finished study, and you wish to know if the Sample Size was "Large Enough".

If this is indeed the case, I'm sure that the esteemed members of Medstats would agree that Post-Hoc calculations are not very informative and it is better to examine the width of Confidence Intervals and whether these CI's contain clinically meaningful values in order to assess study quality.

Tzippy Shochat  

2012/5/14 Mehwish Hussain <mehvish....@gmail.com>

[Steve Simon, P.Mean Consulting]

As several people have already noted, your question lacks sufficient
detail to answer it. Depending on what you want, you may find one of
these pages helpful.

www.pmean.com/01/power.html

www.childrensmercy.org/stats/weblog2007/ValidityReliabilitySampleSize.aspx

(Watch the word wrap on this one).

www.childrensmercy.org/stats/weblog2007/DoYourOwnPower.aspx

www.childrensmercy.org/stats/weblog2005/CorrelationCoefficient.aspx

One last comment. If your goal is to validate the scale, then often the
measure you use to establish validity is a correlation of some sort. In
this case, the last of the above links is relevant.

Feel free to offer additional details if you want additional help.

Steve Simon, n...@pmean.com, Standard Disclaimer.
Sign up for the Monthly Mean, the newsletter that
dares to call itself average at www.pmean.com/news

[mohan]

Ravi, I think you want to find out the validity of "diabetes risk score" in certain population so try find answer for that.