Answer to Question #232632 in Economics for KAT

Question

Explain the underlying ideas behind the logit model. Explain on what grounds logit model is an improvement over linear probability model?

Answer 1

The generalized linear model extends the general linear model by associating the dependent variable with factors and covariates through a specified function. Moreover, the model assumes that the dependent variable has a distribution that differs from normal. Commonly used statistical models are covered, such as linear regression for normally-distributed responses, logistic models for binary data, log-linear models for counting data, double-logged logarithm models for interval-censored survival data, plus many other statistical models, up to and including very general editions.

Example. A shipowner can use generalized linear models to fit Poisson regression to the number of damage to ships of several types built at different times, and the final model can help determine which types of ships are most susceptible to damage.

An auto insurance company can use generalized linear models to fit a gamma regression to insurance claims for vehicle damage, and the final model can help identify the factors most relevant to claims.

Physicians can use generalized linear models to fit a complementary log-log regression to interval-censored survival data to predict the time of disease recurrence.

Features of these generalized linear models

Data. Response data can be scale, quantities, binary data, or test events. Factors are assumed to be categorical variables. Covariates, scale weights, and biases are assumed to be scale variables.

Assumptions. Observations are assumed to be independent.

To get a generalized linear model

This feature requires the Custom Tables module and advanced statistics.