0
$\begingroup$

For my master thesis I'm preparing an individual patient data meta analysis based on data of four studies. In those studies parental psychopathology is measured by three different, but highly correlated, instruments. I'm looking for a way to compose a new variable based on the information of the three instruments measuring parental psychopathology.The new variable will be used in a regression analysis to study the predictive value of parental psychopathology on the outcome of treatment with Cognitive Behavioral therapy. Can anyone recall any discussion or examples in the literature that address the question of how to create a single measure when different instruments have been used across different studies?

$\endgroup$
  • $\begingroup$ Have you considered dimensionality reduction techniques, e.g. PCA? You can create one regressor out of several correlated ones this way. $\endgroup$ – Dmitry Smirnov Nov 24 '15 at 12:15
0
$\begingroup$

I think you're asking how to construct a composite index. I was looking into this a while ago and found this handbook by OECD very helpful, which discussed in detail the steps and different methods you should consider (PCA is one of many methods) I hope it is helpful for you!

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.