View Full Version : Summary: Minimizing group differences at baseline in controlledtrials

05-14-2010, 09:23 AM
Two months ago I sent a message to the BIOMCH-L and Sportscience lists
asking for volunteers to try out the spreadsheets I devised to allocate
subjects to groups in a manner that minimizes differences in group means of
subject characteristics. I ended up with feedback from several people,
thank you. The article is now written and peer-reviewed. Go to
http://sportsci.org/2010 to view the article explaining the gains in
precision with minimization and to download the spreadsheets.

There are bound to be a few bugs in the spreadsheets and errors in the
article. Please get back to me if you find anything worth fixing.

By the way, in the process of checking this article on the Web I discovered
that Firefox produces pixelly images from the html conversions I do from the
original Word doc. There is also a problem with some symbols of the symbol
font with Firefox 3. I have also been playing with the screen resolution
since I got a new laptop this week, and I found that Firefox does not render
pages within frames correctly when I change the DPI setting from normal (96
DPI) to the large (120 DPI) or a custom setting. None of these problems
occur with Internet Explorer. I don't know about other browsers. I found
also that Dreamweaver does not display its windows correctly with anything
other than the normal DPI setting. In fact it's practically unusable. Bottom
line for me is that I HAVE to use a screen resolution (1440 x 900) that is
less the optimum (1680 x 1050) in order to be able to read the screen, but
the result is a degraded fuzzy display. Does anyone have a solution? I have
played with larger font sizes with 1680 x 1050 resolution, but there are
many windows that don't respond to the new fonts and I end up with a
disgusting mess. My IT people aren't any help.

Will G Hopkins, PhD FACSM
Contact info: http://sportsci.org/will
Sportscience: http://sportsci.org
Statistics: http://newstats.org
Be creative: break rules.