-chi2fit- is a user-written program by Stas Kolenikov.
The problem indicated is the lack of a variable -poisson-,
which is needed to hold your theoretically expected cdf.
This is indicated in the help file.
In this case it seems to me that
. poisson locn
would provide a direct approach.
Nick
n.j.cox@durham.ac.uk
Rajiv Mallick
> I am trying to use -chi2fit- test to compare distribution of
> my variable
> LOCN against Poisson distribution. I am looking for some
> guidance on how to
> use -chi2fit- test, i.e.,
> - correct Stata syntax for the test,
> - do I need to first compute cdf of the variable,
> - in the Stata command how do I specify the
> theoretical distribution
> I am comparing the distribution of my variable against,
> - do I need to -collapse- the variable? If I do, then
> is it group
> wise or for any point.
>
> You can see, any help in implementing this test will be appreciated.
>
> My variable looks and feels like the following, and my
> attempt on -chi2fit-
> is below that.
>
>
> Variable LOCN
> ----------------------------------------------------------------
> . tabulate locn
> LOCN | Freq. Percent Cum.
> ------------+-----------------------------------
> 0 | 2,984 64.35 64.35
> 1 | 1,409 30.39 94.74
> 2 | 185 3.99 98.73
> 3 | 59 1.27 100.00
> ------------+-----------------------------------
> Total | 4,637 100.00
>
> . summ locn
> Variable | Obs Mean Std. Dev. Min
> Max
> -------------+--------------------------------------------------------
> locn | 4637 .4218245 .6325427 0
> 3
> ----------------------------------------------------------------------
>
>
> My Attempt on -chi2test-
> ----------------------------------------------------------------
> . chi2fit locn poisson if loyald==1 pw(4) fw(4), bins(4) df(0)
> variable poisson not found
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/