Time Counts - PowerPoint PPT Presentation

1 / 11
About This Presentation
Title:

Time Counts

Description:

a roadmap for applied users. the use of EHA. key decisions and how to make them. Why not OLS? the positive nature of duration data. censoring Daschle v. McCain ... – PowerPoint PPT presentation

Number of Views:18
Avg rating:3.0/5.0
Slides: 12
Provided by: facul97
Learn more at: https://gwipp.gwu.edu
Category:
Tags: counts | time

less

Transcript and Presenter's Notes

Title: Time Counts


1
Time Counts
2
Delay in politics and everyday life
  • reopening of Laurel Park
  • Iraqi elections

3
Event history analysis...
  • a roadmap for applied users
  • ?the use of EHA
  • ?key decisions and how to make them
  • ?interpretation

4
Why not OLS?
  • the positive nature of duration data
  • censoringDaschle v. McCain
  • time-varying covariatesDaschle as party leader

5
The basics of EHA
  • the hazard ratethe rate at which units fail by t
    given that the unit had survived until t
  • absolutely continuous v. discrete treatments
  • given that Daschle has been in the Senate for 18
    years, what are the chances hell be defeated in
    2004?

6
Parametric v. non-parametric models
  • what is duration dependence?
  • the s-curve in state innovation
  • consider parametric models when
  • ? theory for time dependency
  • ? substantive interest in time dependency

7
Some popular parametric models
  • Weibull modelmonotonic baseline hazard
  • non-monotonic modelslog-logistic and log-normal
  • theory and data in choosing parametric models

8
The Cox proportional hazards model
  • duration dependency is left unspecified
  • so what does this mean in practical terms?

9
Time and rulemaking
  • Cox Model
  • Variable Est.(s.e.)
  • Constant n.a.
  • Adcomm -1.00(.25)
  • Pages .0008(.004)
  • Indep -.33(.29)
  • p n.a.
  • N 170
  • Log-likel. -696.72
  • Weibull Model
  • Variable Est.(s.e.)
  • Constant 5.89(.30)
  • Adcomm .80(.25)
  • Pages -.0007(.004)
  • Indep .27(.30)
  • p 1.23(.15)
  • N 170
  • Log-likel. -227.90

10
Topics for another day
  • interpreting for substantive meaning
  • models for discrete duration data
  • time-varying covariates
  • repeated eventsmultiple and competing
  • models for unobserved heterogeneity
  • just scratching the surface

11
If you want more
  • ? Janet M. Box-Steffensmeier and Bradford S.
    Jones, Event History Modeling A Guide for Social
    Scientists
  • ?Mario A. Cleves, William W. Gould, and Roberto
    G. Gutierrez, An Introduction to Survival
    Analysis Using Stata
Write a Comment
User Comments (0)
About PowerShow.com