Print

Print


Douglas (& Teresa);

 

CATmaker does indeed allow the entry of descriptive statistics to compute a confidence interval, but does not yield the correct confidence interval for the difference between two means.  The equation used to compute the standard error of the difference between means has the pooled variance term outside the radical; therefore, the standard error is too large, resulting in confidence intervals that are much wider than they should be.

 

I use (and teach using) CATmaker, but make sure to recalculate the confidence interval for the difference between means using an Excel spreadsheet (Confidence Interval Calculator) available at http://www.pedro.org.au/english/downloads/.

 

Frank Underwood

 

From: Evidence based health (EBH) [mailto:[log in to unmask]] On Behalf Of Douglas Badenoch
Sent: Friday, June 08, 2012 10:16 AM
To: [log in to unmask]
Subject: Re: Online calculator sought

 

Hi Teresa

CATmaker does that:  http://www.cebm.net/index.aspx?o=1216

Caveat:  It's really old:  15 years or so, which is about 105 in human years.  You have to download it and install it on your PC.  There's no Mac version.  So I would imagine that no guarantees are given nor implied!

However, it does what you require.

cheers

Douglas

On 08/06/2012 14:54, Benson, Teresa wrote:

I love the online calculator at KT Clearinghouse (Canada), it’s great for newbies who just need to figure out the ARR and NNT when an article reports only the odds ratio, or the PPV when the article only gives specificity, etc.  However, it only works with categorical data.  I’m wondering if there’s anything similarly basic for continuous data?  For example, so that when an article gives the standard mean difference but doesn’t report confidence intervals, the user can calculate the latter (e.g., to more adequately assess the power of an equivalence trial or the precision of a reported effect size).  I’ve found lots of calculators where you can put in the raw data, but I really need one where they can just plug in the means and standard deviations of the two groups, which articles almost always provide.  Any ideas?  (Dimension Research used to have a great one, but they recently took the site down.)

 

Teresa Benson, MA, LP

Clinical Lead, Evidence-Based Medicine

McKesson Health Solutions
18211 Yorkshire Ave

Prior Lake, MN  55372

[log in to unmask]
Phone: 952-226-4033


 

 




-- 
Mr Douglas Badenoch
Director, Minervation Ltd
-------------------------
23 Bonaly Grove
Edinburgh
EH13 0QB
-------------------------
Tel:  +44 131 441 4699
Web:  www.minervation.com
-------------------------
Minervation is a limited company registered in England and Wales 
Registered number: 4135916 
VAT number: 792674384
Registered Office: Salter's Boat Yard, Folly Bridge, Abingdon Road, Oxford, OX1 4LB  
United Kingdom