Quartile average question

#1
Hi,

This might be a basic question, but I have a dataset that measures items on 3 distinct measures (say time, speed, and distance). I’m attempting to find a combined ranking of these. Is it acceptable to take the quartile ranking of each measure and then find the average quartile ranking across the three measures?

In other words one item might have quartile rankings of 1 (time), 2 (speed), and 3 (distance). The average quartile ranking would then be 2. I’m just not sure if that’s something that’s done or shouldn’t be done.

Thanks for any help!
 

katxt

Well-Known Member
#2
Quartiles may be a little coarse, if you have actual measurements.
This looks like the same sort of problem as scoring the decathlon. For that, the various events are scaled by a relatively simple formula based (I assume) on historical data so that no event can dominate the others.
I only know this because I just read https://www.sportcalculators.com/decathlon-calculator which may give you some ideas for your project.
Or a simpler version of the same idea. Find the mean and SD of each measure. Transform each score using (value - mean)/SD or (mean - value)/SD depending on whether a large number is good or bad. Average the new scores.
 
Last edited: