Correlation and range

tbb

New Member
#1
Background: I am working on this project that examines the relationship between battery life and several variables. However my specific question pertains to the variable of the temperature of the environment they are in.

I have 44 different batteries as my sample size. Their lifespan (in days) ranges from 4 days to 538 days.

The temperature variable ranges from 45 to 72 degrees (f), a significantly smaller range.

I have never taken a stats class, so this may be a simple question but here it is. Would the difference in range sizes impact the correlation coefficient? I feel that because the difference from 45 to 72 is not particularly notable, value wise, in comparison to 4 to 538.

Thoughts?, thanks
 

rogojel

TS Contributor
#2
hi,
probably not. The range is only a matter of chosing the scale of measurements and you could get any range bv using a linear zransformation. It would make the correlation coeficient completely useless if its value depended on the measurement ubits chosen. so, in short, it should not have an effect.

regards
rogojel