In phmpacheco-ufjf/clm: A package to deal with complex longitudinal survey data

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

clm

A package to deal with complex longitudinal survey data. It combines longitudinal methodology models with complex sampling design. It fits fixed and random effects models and covariance structured models so far. It also provides tools to perform statistical tests considering these specifications.

Installation

The clm package is in its initial development stage and does not yet have a version available on CRAN.

However, the development version of the package is now available for download and can be installed via the repository on GitHub with:

# install.packages("devtools")
devtools::install_github("phmpacheco-ufjf/clm")

Avaliable Functions

There are a few major functions available for usage. Here's an example using BHPS survey data.

First, we fit the fixed and random effects model considering the survey sampling design. The sigma argument can be: identity, exchangeable, autorregressive or a custom square matrix.

library(clm)

fit <- clm(
  score ~ wave + ageg + ecacg + qualifg,
  waves = wave, ids = id,
  weights = weight, stratum = strata, cluster = cluster,
  data = example_data, sigma = "exchangeable"
)
summary(fit)

Then, we fit the covariance structured model. But first, we must choose the covariance structure we want to fit. Here's an example for the Uniform Correlation Model - UCM:

sigmaThetaExpr_viewer("UCM", 5)

Finally, we fit the model using our previous random effects fitted model.

fitTheta <- cov_clm(fit,
  fittingType = "PML", sigmaThetaExpr = "UCM",
  optimParams = list(par = c(7, 5))
)
summary(fitTheta)

phmpacheco-ufjf/clm documentation built on Dec. 22, 2021, 8:40 a.m.