Plan for 0.1.0

[LukaszChrostowski]

Models and methods that will be available in the 0.1.0 version of the nonprobsvy package

1. Population mean estimators (standard approach)

• Doubly robust estimator
• Inverse probability weighted estimator
• Mass imputation estimator using model based approach
• Mass imputation estimator using predictive mean matching (k-nearest neighbor)

2. Population mean estimation with minimization of asymptotic bias

• Doubly robust with gaussian, binomial and poisson distribution for outcome variables
• Mass imputation with gaussian, binomial and poisson distribution for outcome variables
• Inverse probability weighted with gaussian, binomial and poisson distribution for outcome variables

3. Propensity score estimation using Maximum likelihood approach

• logit regression
• probit regression
• complementary log-log regression

4. estimation of population mean variance for each of the estimators

• analytic
• bootstrap

5. Control parameters for:

• selection equation
• outcome equation
• inference method