Correlation and reliability

[Guest skalibaba]

I want to find the correlation between 2 sets of measurements. The range of the observed values is between -1.5 and 1.5. Now in theory, the values could be between -6 and 6 (but this is less likely to be).

 

I also know that the 2 sets should have a high correlation (both sets are measurements of the same thing, but it has been measured with 2 different machines). Right now, I don't get a very high correlation (around 0.4). I know that if I increase the range of the observed data, it should help raise the correlation. But it is difficult to do so experimentally. So I thought that a correction to the correlation might help. I read on the Internet that low reliability attenuates the correlation.

 

So I was wondering if the problem I have right now is related to low reliability. If so, do you think applying the correction for attenuation (ie correlation coefficient divided by Cronbach's alpha) would give me a better indication of the true correlation? Or maybe there is a better way?

 

Thank you very much.