I have a data frame which has non-aggregated data and the another dataframe which has the same information as the first one but in an aggregated manner. When I run the regression analysis for both of these models, I expected them to be same. But to my surprise they differed. Can you please explain why this is happening.
$\begingroup$
$\endgroup$
2
asked Oct 23, 2016 at 17:37
Rohit Venkat Gandhi MendadhalaRohit Venkat Gandhi Mendadhala
3131 gold badge2 silver badges8 bronze badges
-
1$\begingroup$ I am surprised that you are surprised! $\endgroup$– kjetil b halvorsen ♦Commented Oct 23, 2016 at 18:21
-
$\begingroup$ Without knowing how they differed it is almost impossible for anyone to clarify what is going on here. $\endgroup$– mdeweyCommented Oct 23, 2016 at 21:02
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
If aggregation means that you are combining data originating from several groups together, you might be experiencing what is sometimes called Simpson's Paradox: https://en.wikipedia.org/wiki/Simpson%27s_paradox
