Improved testing inferences for beta regressions with parametric mean link function

This code code can be used to perform data analysis and Monte Carlo simulations.

Authors

Cristine Rauber, Francisco Cribari-Neto, Fábio M. Bayer

How to cite this work

Rauber, C., Cribari-Neto, F., & Bayer, F. M. (2020). Improved testing inferences for beta regressions with parametric mean link function. AStA Advances in Statistical Analysis, 104(4), 687-717.

Programming language

Ox The console version of Ox is freely distributed for academic use. It is currently available for Windows, Linux, macOS, and several Unix platforms.

Description

Data analysis

The program is used to estimate the beta regression model with parametric mean link function presented in Section 5.1.

Monte Carlo simulation 1

The program is used to compute the null rejection rates of the tests of H0: \gamma_2 = \gamma_3 = 0 (fixed precision) in the varying precision beta regression model with parametric mean link function; \gamma_1 = \log(30).

Monte Carlo simulation 2

The program is used to compute the null rejection rates of the tests of H0: \lambda = 1 in the fixed precision beta regression model with parametric mean link function; \phi = 30.

Monte Carlo simulation 3

The program is used to compute the null rejection rates of the tests of H0: \beta_4 = 0 in the fixed precision beta regression model with parametric mean link function; \phi = 30.

Files

Data analysis

• gasoline.mat: data
• appPrater.ox: Ox code for reproducing some of the results in Section 5.1
• appPrater.out: Ox output

Monte Carlo simulation 1

• log30fixedprec.ox: Ox code
• log30fixedprec.out: Ox output

Monte Carlo simulation 2

• phi30lambda.ox: Ox code
• phi30lambda.out: Ox output

Monte Carlo simulation 3

• phi30beta4.ox: Ox code
• phi30beta4.out: Ox output

Execution

Data analysis

1. Command line: oxl appPrater.ox > appPrater.out (On some systems, one must use oxl64 instead of oxl).

2. IDE, via OxEdit: Open the file appPrater.ox in OxEdit (Ox IDE) and then click on: Run -> Ox

Monte Carlo simulation 1

1. Command line: oxl log30fixedprec.ox > log30fixedprec.out (On some systems, one must use oxl64 instead of oxl).

2. IDE, via OxEdit: Open the file log30fixedprec.ox in OxEdit (Ox IDE) and then click on: Run -> Ox

Monte Carlo simulation 2

1. Command line: oxl phi30lambda.ox > phi30lambda.out (On some systems, one must use oxl64 instead of oxl).

2. IDE, via OxEdit: Open the file phi30lambda.ox in OxEdit (Ox IDE) and then click on: Run -> Ox

Monte Carlo simulation 3

1. Command line: oxl phi30beta4.ox > phi30beta4.out (On some systems, one must use oxl64 instead of oxl).

2. IDE, via OxEdit: Open the file phi30beta4.ox in OxEdit (Ox IDE) and then click on: Run -> Ox

Contact

Cristine Rauber (rauberoliveira at gmail.com)

License

MIT