[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

From |
agostino@unical.it |

To |
statalist@hsphsun2.harvard.edu |

Subject |
st: Poisson model with interaction term |

Date |
Mon, 25 Feb 2008 14:41:29 +0100 |

Dear all, I’m estimating a Poisson model, which includes an interaction term as follows: (ln)Y=a+b1x1+b2x2+b12x1*x2 and I need to compute the impact (marginal effect) of x1 on lnY. I have found on SJ an article “Computing interaction effects and standard errors in Logit and Probit models”, by Norton, Wang and Ai (2004), who warn that for nonlinear models:“ First, the interaction effect could be nonzero even if b12 = 0. (…) Second, the statistical significance of the interaction effect cannot be tested with a simple t-test on the coefficient of the interaction term b12. Thirdly, the interaction effect is conditional on the independent variables, unlike the interaction effect in linear models. (It is well known that the marginal effect of a single uninteracted variable in a nonlinear model is conditional on the independent variables). Fourthly, the interaction effect may have different signs for different values of covariates. Therefore, the sign of b12 does not necessarily indicate the sign of the interaction effect” (page 156). Please notice that my interest is computing the effect of x1 on lnY , I’m not interested in the marginal effect of the interaction term, nor in the effect of x1 on E(Y), because my dependent variable is not a count. Since there is the interaction I cannot conclude that the impact of x1 on lnY is just b1. In a linear model I could apply the following formula: marginal effect of x1= b1+ b12*x2, and compute the relative standard errors sqrt(varb1+var b12*x22+2x2covb1b12) However, according to Norton et al. (2004), I cannot employ the STATA estimates in this way, as the interaction effect may have different signs for different values of covariates. How can I then compute (and possibly graph) the impact of x1 across the values of x2 using STATA? I would really appreciate any assistance with this issue Maria ------------------------------------------------- This mail sent through IMP: http://horde.org/imp/ * * 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/

**Follow-Ups**:**Re: st: Poisson model with interaction term***From:*Maarten buis <maartenbuis@yahoo.co.uk>

- Prev by Date:
**Re: st: RE: Out of Sample forecasting** - Next by Date:
**Re: st: Poisson model with interaction term** - Previous by thread:
**[no subject]** - Next by thread:
**Re: st: Poisson model with interaction term** - Index(es):

© Copyright 1996–2014 StataCorp LP | Terms of use | Privacy | Contact us | What's new | Site index |