Make predictions with standardized coefficients

[Seko1987]

Hello,

I've been struggling with a problem since a while, hopefully someone here could help me out.

The problem:

I want to use the coefficients from a previous research to make a prediction equation. The problem is that the paper only lists the standardized (beta) coefficients of the independent variables, the F-statistic and their p-value.

I don't have the unstandardized coefficients neither the intercept of the equation.

I think it is necessary to convert unstandardized the coefficients to standardized coefficients.

How do I calculate the intercept?
Is this possible? Is there any formula to convert it?

I would greatly appreciate any help.

Thanks

 

[hlsmith]

Could it be possible to get to the intercept if they provided estimates of the parameters (even though it may be in the both the numerator and denominator? Regardless, it would very beneficial if you cited the paper or upload the article, so we can see what data are available.

 

[Seko1987]

Please take a look on the paper if you have the time: paper

Statistics is not my field, so I'm unware of some procedures...

 

[hlsmith]

I just skimmed the article, whew. There may be plenty of work attempting to use their data. Just from a perusal, they collected quite a bit of data. Could not tell from first glance, but they seem to have a pretty saturated model, which controlled for many variables including interactions. Interpretation of the variables may become difficult when having to keep in mind and interprete their interactive affects. May be better off conducting your own study, up to you and your objectives.

 

[Seko1987]

Thanks for your relply and advise.

This other paper used a very simple model based on the beta coefficients of the paper we are talking about.

Assuming I still want to do this, is there a simple or naîve way of using this small ammount of beta coefficients and do it? I thought that with the beta coefficients I could replicate the equation even with just these coeeficients? It is statistically possible?