Comparing different units of measurement

#1
Hiya,
Im wondering if anyone has any views on how to compare change in variables that are measured in different units? I cant normalise the data becasue that defeats the object of looking at the sensitivity of each of the measures, but some variables are measured as ratios, some in units below 10, others in units above 100 so even using percent change has limitations. Any suggestions?
 

TheEcologist

Global Moderator
#2
Hiya,
Im wondering if anyone has any views on how to compare change in variables that are measured in different units? I cant normalise the data becasue that defeats the object of looking at the sensitivity of each of the measures, but some variables are measured as ratios, some in units below 10, others in units above 100 so even using percent change has limitations. Any suggestions?
You will have to find a fair comparison meassure (so you need to transform) otherwise no test is valid.

see
http://www.talkstats.com/showthread.php?t=4673

Maybe ranking might help, not sure though not enough info to work with.

goodluck
 
#3
Thanks for your response.
What I am tryring to do is look at a series of measurements over time to try and find out if some change more than others. So cant normalise or transform (i dont think) as Im interested in the sensitivity of each measure. But there lies the problem as each measure is in different units ie mm, ratios, arbitary numbers, some on a big scale (100+), others on smaller scale (below 10) so if try to use percent change some will show greater/less changes that aren't really true.

Any other ideas or suggestions would be most welcome....