RE: Poisson and Overdispersion

For econometric discussions not necessarily related to EViews.

Moderators: EViews Gareth, EViews Moderator

eviewsuser10
Posts: 16
Joined: Thu Oct 15, 2020 3:27 am

RE: Poisson and Overdispersion

Postby eviewsuser10 » Thu Oct 22, 2020 12:20 pm

Hello,

I am going to conduct interrupted time series analyses with a 'count' dependent variable. I ran poisson regression of fatalc on independent variables. Then, I obtained the fitted values of the dependent variable and conducted a regression based test of the Poisson restriction using the equation: equation testeq.ls (fatalc-fatalc_f)^2-fatalc fatalc_f^2

As seen below, the t-statistic of the coefficient (-0.065989) is significant. Is my dependent variable overdispersed? The coefficient is negative. What does being negative mean? Would you please help me interpret the table below. Thank you.

Dependent Variable: (FATALC-FATALC_F)^2-FATALC
Method: Least Squares
Date: 10/22/20 Time: 14:53
Sample: 2009M01 2020M09
Included observations: 141

Variable Coefficient Std. Error t-Statistic Prob.

FATALC_F^2 -0.065989 0.017866 -3.693427 0.0003

R-squared -0.137073 Mean dependent var -0.760818
Adjusted R-squared -0.137073 S.D. dependent var 1.533614
S.E. of regression 1.635348 Akaike info criterion 3.828655
Sum squared resid 374.4108 Schwarz criterion 3.849568
Log likelihood -268.9202 Hannan-Quinn criter. 3.837154
Durbin-Watson stat 1.742558

Return to “Econometric Discussions”

Who is online

Users browsing this forum: No registered users and 26 guests