Re: Choosing type of regression

Posted by Bruce Weaver on Jan 24, 2011; 2:32pm
URL: http://spssx-discussion.165.s1.nabble.com/Choosing-type-of-regression-tp3354689p3354717.html

Ordinal logistic regression (PLUM procedure) would be a candidate.  But you'd need a fairly large sample to avoid over-fitting, given your list of explanatory variables.

HTH.

mortenuth wrote

I am making an regression analysis regarding the time perspective in global companies risk management, and have some problems choosing the right type of regression analysis (multiple regression, multinominal logistic regression etc.)

My dependent variable is the time horizon of companies foreign exchange rate exposure management, based on following question:

1) At the present time, what is the average time horizon that your company has covered its foreign  
    exchange rate exposures by use of financial means? (i.e. forward contracts, options, swaps and debt in
    foreign currency)
   
     a) The category of answers is: 0-1 months, 1-3 months, 3-6 months, 6-12 months, 1-2 years, 2-4
         years and finally above 4 years

The independt variables are various and e.g.
       
       a) Numbers production subsidiaries abroad (answers may be 0, 1-2, 3-5, 6-9 and >9)
       
       b) What percentage of your company's consolidated debt is in foreign currency (answers may be
           0%,1-20%, 21-40%, 41-60%, 61-80%, 81-99% and 100%)

       c) How would you characterize the general business environment in which your company primarily
           operates (answers by 5-point likert scale from very stable to very volatile)

       d) Total balance (from 0 dollars to an infinite number of dollars)

I hope someone may be help me choosing the best way of making an regression analysis?

Best regards Morten