How to know the correlation between the T level, the C level and the score audiometry?

[Lucas Bordure]

Good afternoon.
I have 5 columns. I would like to know the correlation between the T level and C level on the score audiometry.
Can you share me ideas?

[George Ford]

I don't know what you variable labels are (Stata won't allow spaces), but this is the basic command:

correl Tlevel Clevel

[Leonardo Guizzetti]

It's not clear exactly what *kind* of correlation you want. Since you have repeated measures on each subject across electrodes, a simple correlation in the full T and C variables will be some mix of between and within correlation. Perhaps the simplest starting point will be to examine correlations by electrode, or else you may know of a way to to combine electrode measures into a single score per patient.

The former can by done by

bysort electrode : corr T C

[Lucas Bordure]

Sorry for the delay of my answer. I've been very sick this week. Thank you a lot for your help Leonardo and George.

[Carlo Lazzaro]

Lucas:
first of all I do hope you've fully recovered.
The issue is potentially trickier here, as many people have two ears (hence, the observations are not independent at person/patient level, too).
I was not able to spot whether this is the case with your dataset but, if it were, I would probably consider regression vs. correlation.

[Lucas Bordure]

I wrote this:

corr scoreaudiometry tlevel clevel

What kind of interpretation is possible?

[Carlo Lazzaro]

Lucas:
see Example 1: Principal component analysis of audiometric data, -pcs- entry, Stata .pdf manual.

[Lucas Bordure]

Thank you Carlo Lazzaro. I'm not recovered, I think the remission will be long. Principal component analysis is not the concept to use in my example. But as you said, I will use the regression concept.

[Carlo Lazzaro]

Lucas:
all the best for a full and shorter than expected recovery.
Researching and all the rest will follow.
I'm happy with helping you as long as I can.