DATE
|
TOPIC
|
HANDOUTS (data & do-files)
|
ASSIGNMENT(S)
|
Week 1
|
Useful Information: Replication, do-files and log files, CLARIFY and simulation
|
Week 2
|
Maximum Likelihood Estimation: Theory and mechanics
|
Weeks 3 & 4
|
Binary Dependent Variables: Logit, probit, scobit, heteroskedastic probit, rare events logit
|
Weeks 5 & 6
|
Discrete Choice Models: A general framework for discrete choice models. Multinomial logit, conditional logit, nested logit, multinomial probit, and mixed logit. Simulated maximum likelihood.
|
Week 7
|
Ordered Response Models: Ordered probit and logit, generalized ordered logit, heteroskedastic ordered probit
|
Week 8
|
Event Count Models: Poisson model, negative binomial model, generalized event count model, hurdle model, zero-inflated models
|
Week 9
|
Continuous Time Duration Models: Basic components of duration analysis. Nonparametric models, parametric models (exponential, weibull, log-logistic, generalized gamma, etc.), semi-parametric models (Cox)
Note: For notes and exercises engaging discrete time duration models (logit, probit, clog log) and advanced duration models (frailty, competing risks, repeated events, split population, selection) see weeks 10 and 11 on Matt Golder's Methods IV page. |
Week 10
|
Tobit Models: Truncated and censored data
|
Week 11
|
Selection Models: Heckman model (two-step and MLE), bivariate probit, bivariate probit with partial observability, bivariate probit with sample selection
|
Weeks 12 & 13
|
Time Series: Stationary and non-stationary data, Durbin-Watson and LM tests for serial correlation, Cochrane-Orcutt and Prais-Winsten procedures, lags (finite distributed, ADL, etc.), error correction models, impulse and unit response functions, AR and MA error processes, tests for stationarity and unit roots, drifts and trends, cointegration, structural stability, spurious regression
|