0
$\begingroup$

In my panel data I observe a quite large amount of different individuals for a fixed time duration. The time axis is fixed for every individual, i.e., I observe 20 time points for every individual.

The dependend variable is of kardinal scale but I choosed to transform it into ordinal scale. The reason for this transformation is due to the fact that I observe the case "-Inf" very often which can obviously not be included in any form of regression model. So I choose to split the axsis into intervals like (-Inf,-20000],(-20000,-10000],(-10000,0],(0,10000] and so on. Due to this transformation I actually observe a dependend variable of ordinal scale.

My first question in this context is: In what way a ordinal scaled dependend variable with time and individual specific variation can be modelled.

2nd: If I acutally do a "ordinal regression" how do I deal with a continuous explanatory variable. I already checked how small the cells get and I'll receive a very large multidimensional table with a lot of zeros and here and there a number from 4 upwards.

Kind regards

Improve this question
$\endgroup$
13
  • 1
    $\begingroup$ How is it possible to observe "-Inf"? $\endgroup$
    – Glen_b
    Commented Apr 9, 2013 at 23:45
  • 1
    $\begingroup$ I assume there's an $E$ goes in here: $MR = \frac{P}{MR-P}$ somewhere. In any case, that looks like one is observing P and Q, and then calculating E. $\endgroup$
    – Glen_b
    Commented Apr 10, 2013 at 7:17
  • 1
    $\begingroup$ One possibility would be to split into the cases where $\frac{\partial P}{\partial Q} = 0$ (I'm assuming Q isn't 0, let me know if that's wrong), and the cases where it isn't. The model may be somewhat different there, and then you don't need to categorize the whole thing. $\endgroup$
    – Glen_b
    Commented Apr 10, 2013 at 7:20
  • 1
    $\begingroup$ Thanks for the clarification. Why would that put all those cases together? $\endgroup$
    – Glen_b
    Commented Apr 10, 2013 at 7:40
  • 1
    $\begingroup$ Well, yes and no. I mean that when D=1, you'd have a suitable model for the original dependent variable as a continuous thing. When D=0, you'd do whatever was suitable there. Imagine writing a model for each of those rows. $\endgroup$
    – Glen_b
    Commented Apr 10, 2013 at 8:03

0

Reset to default

Your Answer

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