1
$\begingroup$

I performed an Interview with 44 questions Protocol. The structure of questions is based on 18 variables. Major variables are coming from theory. Every major variable consists of 3,4 or more question items. In fact, I want to compare every major variable with each other and compute a correlation coefficient. Now for analysis, when I want to build major variable with available data, I have a problem. For example: major variable(1) i.e. organizational Justice consists of two nominal data (two questions of interview protocol with nominal categories) and one ordinal (one question of interview protocol with ordinal categories). Similarly,major variable(2): organizational trust consist of three ordinal data and two nominal data. Now my available data are the combination of ordinal and nominal. I don't know how do I merge nominal with ordinal data to build a unique variable? The major variable are based on theory,

enter image description here

$\endgroup$
4
  • $\begingroup$ pl. indicate your questions and response categories. $\endgroup$
    – user10619
    Apr 20, 2016 at 8:40
  • $\begingroup$ I don't understand what you mean by "pl.". please describe it more. $\endgroup$ Apr 26, 2016 at 7:57
  • $\begingroup$ may be you need to write your qualification. $\endgroup$
    – user10619
    Apr 26, 2016 at 10:48
  • $\begingroup$ I am master student of Industrial Engineering. We read more about programming producing activity in bachelor degree, Human resources and strategic issues in master. So more description about statistical issues is needed... $\endgroup$ Apr 30, 2016 at 6:33

1 Answer 1

1
$\begingroup$

I think it would be possible to use factorial analysis (PCA, MCA, ...) on each of your varibles subset so as to create multiple variables (PCA dimensions) which are linear combinations of your initial variables. So if your PCA first dimension yields enough inertia you could use it as a single synthetic variable for all the other and test correlation with it.

The problem being that if your first dimension does not yield a sufficient amount of inertia then that dimension would not really be representative of your initial inertia. Your computed correlation coefficients would then have a weaker meaning.

EDIT : Note that using ordinal variables is a violation of the continuity assumption made for PCA but it is often used anyway.

You should also use MFA (used PCA as it is more renown) as you have both categorical and numeric variables.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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