Solved – How to convert logistic coefficients into correlations

correlationregression

I am new to statistics so I have no clue about all these. I have to identify correlations for my work. Here is where I am stuck. Few papers have direct correlations and few papers do not have. So, I am confused with the papers where there is no direct correlations given.

This is the picture I have doubt in

In the above figure there is coefficient and is i same as regression coefficient?

Is the regression coefficient same as the coefficient as mentioned above picture

My only doubt is can I use the Table 4 coefficients to convert into correlations?
And what are the variables x,y in the formula 4?

Thanks

Best Answer

The x and y in equation 4 are the predictor and response variables so it's the ratio of standard deviations in predictor and response.

That doesn't help you much though because these formulas are for an ordinary regression analysis and do not apply to logistic regression. There is no equation for logistic regression to derive the correlations directly from coefficients because you don't calculate Pearson's r for logistic response data (just 0's and 1's). You can calculate things that look like it from the data—usually using a modified t-test where you flip dependent and independent variables—but it's probably best to just interpret the t-ratio directly since that's already a standardized value.

Related Question