[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Unbalanced classification problem

From   "Fardad Zand" <[email protected]>
To   [email protected]
Subject   st: Unbalanced classification problem
Date   Fri, 17 Oct 2008 15:47:43 +0200

Hi All,

In my sample of firm level data, I'm trying to investigate the
moderating role of a dummy variable by introducing interaction terms
(with my main variables of interest) in my specification models.
However, there is the problem of unbalanced classification. The number
of records with a zero value for this dummy variable is far less than
those with a positive or missing value; you can think of an order of
100, 4000 and 6000 for the 0, 1 and missing values respectively. I
believe, if I don't correct for this problem, the minority group will
be underrepresented in calculating the estimates and this will at
least reduce the precision of the estimates.

Q1: Is this problem results in "biased" estimates as well or it only
reduces the efficiency of my estimates?

Q2: How can I correct for or mitigate this problem in Stata?

Thank you so much for taking your time reading and answering my post.
Wish you all success in your own research.

Kind regards from Holland,
*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index