Extracting, Computing and Exploring the Parameters of Statistical Models using R

Abstract

The recent growth of data science is partly fueled by the ever-growing amount of data and the joint important developments in statistical modeling, with new and powerful models and frameworks becoming accessible to users. Although there exist some generic functions to obtain model summaries and parameters, many package specific modeling functions do not provide such methods to allow users to access such valuable information.

Full text

Extracting, Computing and Exploring the Parameters of
Statistical Models using R
Daniel Lüdecke1, Mattan S. Ben-Shachar2, Indrajeet Patil3, and
Dominique Makowski4
1University Medical Center Hamburg-Eppendorf, Germany 2Ben-Gurion University of the Negev,
Israel 3Max Planck Institute for Human Development, Germany 4Nanyang Technological
University, Singapore
DOI: 10.21105/joss.02445
Software
•Review
•Repository
•Archive
Editor: Sebastian Benthall
Reviewers:
•@cmaimone
•@Ohlsen
Submitted: 01 July 2020
Published: 09 September 2020
License
Authors of papers retain
copyright and release the work
under a Creative Commons
Attribution 4.0 International
License (CC BY 4.0).
Summary
The recent growth of data science is partly fueled by the ever-growing amount of data and
the joint important developments in statistical modeling, with new and powerful models and
frameworks becoming accessible to users. Although there exist some generic functions to
obtain model summaries and parameters, many package-specic modeling functions do not
provide such methods to allow users to access such valuable information.
Aims of the Package
parameters is an R-package (R Core Team, 2020) that lls this important gap. Its primary
goal is to provide utilities for processing the parameters of various statistical models. Beyond
computing p-values, standard errors, condence intervals (CI), Bayesian indices and other
measures for a wide variety of models, this package implements features like parameters
bootstrapping and engineering (such as variables reduction and/or selection), as well as tools
for data reduction like functions to perform cluster, factor or principal component analysis.
Another important goal of the parameters package is to facilitate and streamline the process
of reporting results of statistical models, which includes the easy and intuitive calculation
of standardized estimates in addition to robust standard errors and p-values. parameters
therefor oers a simple and unied syntax to process a large variety of (model) objects from
many dierent packages.
parameters is part of the easystats ecosystem, a collaborative project created to facilitate
the usage of R for statistical analyses.
Comparison to other Packages
parameters functionality is in part comparable to packages like broom (Robinson, Hayes,
& Couch, 2020), nalt (Harrison, Drake, & Ots, 2020) or stargazer (Hlavac, 2018) (and
maybe some more). Yet, there are some notable dierences, e.g.:
•broom (via glance()), nalt (via ff_metrics()) and stargazer (via stargazer())
report t indices (such as R2 or AIC) by default, while parameters does not. However,
there is a dedicated package in the easystats project for assessing regression model
quality and t indices, performance (Lüdecke et al., 2020b).
Lüdecke et al., (2020). Extracting, Computing and Exploring the Parameters of Statistical Models using R. Journal of Open Source Software,
5(53), 2445. https://doi.org/10.21105/joss.02445
1

•parameters easily allows to compute standardized estimates, robust estimation, small-
sample-size corrections for degrees of freedom (like Satterthwaite or Kenward-Roger),
bootstrapping or simulating parameters, and feature reduction. Furthermore, param-
eters provides functions to test for the presence or absence of an eect (equivalence
testing, see Lakens, Scheel, & Isager, 2018).
• For most functions, easy-to-use plot()-methods exist to quickly create nice looking
plots (powered by the see package (Lüdecke et al., 2020a)).
•parameters is a very lightweight package. Its main functionality only relies on the
insight, the bayestestR, and the eectsize packages (Ben-Shachar, Makowski, &
Lüdecke, 2020; Lüdecke, Waggoner, & Makowski, 2019; Makowski, Ben-Shachar, &
Lüdecke, 2019) to access and process information contained in models, and these pack-
ages in turn only depend on R core packages. However, additional features that do not
belong to the core functions of parameters require the installation of other packages,
such as sandwich (Zeileis, 2006) for robust estimation, psych (Revelle, 2019) for factor
analysis or PCA or cAIC4 (Saefken, Ruegamer, Kneib, & Greven, 2018) for parameter
selection for mixed models.
Examples of Features
As stated above, parameters creates summary tables of many dierent statistical models.
The workow is simple: t a model and pass it to the model_parameters() function (or its
shortcut, parameters()) to obtain information about the model’s parameters.
In the following, we show some brief examples. However, a comprehensive overview includ-
ing in-depth examples are accessible via the dedicated website (https://easystats.github.io/
parameters/).
Summary of Model Parameters
model_parameters() allows you to extract the parameters and their characteristics from
various models in a consistent way.
library(parameters)
model <- lm(Sepal.Length ~Species, data = iris)
parameters(model)
Lüdecke et al., (2020). Extracting, Computing and Exploring the Parameters of Statistical Models using R. Journal of Open Source Software,
5(53), 2445. https://doi.org/10.21105/joss.02445
2

#> Parameter | Coefficient | SE | 95% CI | p
#> -----------------------------------------------------------------
#> (Intercept) | 5.01 | 0.07 | [4.86, 5.15] | < .001
#> Species [versicolor] | 0.93 | 0.10 | [0.73, 1.13] | < .001
#> Species [virginica] | 1.58 | 0.10 | [1.38, 1.79] | < .001
Extraction of robust indices is possible for many models, in particular models supported by
the sandwich (Zeileis, 2006) and clubSandwich (Pustejovsky, 2020) packages.
parameters(model, robust = TRUE)
#> Parameter | Coefficient | SE | 95% CI | p
#> -----------------------------------------------------------------
#> (Intercept) | 5.01 | 0.05 | [4.91, 5.11] | < .001
#> Species [versicolor] | 0.93 | 0.09 | [0.75, 1.11] | < .001
#> Species [virginica] | 1.58 | 0.10 | [1.38, 1.79] | < .001
For linear mixed models, parameters() also allows to specify the method for approximating
degrees of freedom, which may improve the accurracy for calculated standard errors or p-
values.
library(lme4)
model <- lmer(
Sepal.Length ~Sepal.Width *Petal.Length +(1|Species),
data = iris
)
parameters(model, digits = 3)
#> Parameter | Coefficient | SE | 95% CI | p
#> --------------------------------------------------------------------------
#> (Intercept) | 0.707 | 0.652 | [-0.57, 1.98] | 0.278
#> Sepal.Width | 0.731 | 0.156 | [ 0.43, 1.04] | < .001
#> Petal.Length | 1.023 | 0.143 | [ 0.74, 1.30] | < .001
#> Sepal.Width * Petal.Length | -0.084 | 0.040 | [-0.16, -0.01] | 0.035
parameters(model, digits = 3,df_method = "kenward")
#> Parameter | Coefficient | SE | 95% CI | p
#> --------------------------------------------------------------------------
#> (Intercept) | 0.707 | 0.654 | [-0.70, 2.11] | 0.298
#> Sepal.Width | 0.731 | 0.157 | [ 0.42, 1.04] | < .001
#> Petal.Length | 1.023 | 0.145 | [ 0.74, 1.31] | < .001
#> Sepal.Width * Petal.Length | -0.084 | 0.040 | [-0.16, -0.01] | 0.037
Visualisation
parameters functions also include plotting capabilities via the see package (Lüdecke et al.,
2020a). A complete overview of plotting functions is available at the see website (https:
//easystats.github.io/see/articles/parameters.html).
Lüdecke et al., (2020). Extracting, Computing and Exploring the Parameters of Statistical Models using R. Journal of Open Source Software,
5(53), 2445. https://doi.org/10.21105/joss.02445
3

library(see)
model <- lm(Sepal.Length ~Petal.Width *Species, data=iris)
plot(parameters(model))
Licensing and Availability
parameters is licensed under the GNU General Public License (v3.0), with all source code
stored at GitHub (https://github.com/easystats/parameters), and with a corresponding issue
tracker for bug reporting and feature enhancements. In the spirit of honest and open science,
we encourage requests/tips for xes, feature updates, as well as general questions and concerns
via direct interaction with contributors and developers.
Acknowledgments
parameters is part of the collaborative easystats ecosystem. Thus, we would like to thank
the members of easystats as well as the users.
References
Ben-Shachar, M. S., Makowski, D., & Lüdecke, D. (2020). eectsize: Compute and interpret
indices of eect size. CRAN. doi:10.5281/zenodo.3952214
Harrison, E., Drake, T., & Ots, R. (2020). nalt: Quickly create elegant regression results
tables and plots when modelling. Retrieved from https://CRAN.R-project.org/package=
finalfit
Lüdecke et al., (2020). Extracting, Computing and Exploring the Parameters of Statistical Models using R. Journal of Open Source Software,
5(53), 2445. https://doi.org/10.21105/joss.02445
4

Hlavac, M. (2018). stargazer: Well-formatted regression and summary statistics tables.
Bratislava, Slovakia: Central European Labour Studies Institute (CELSI). Retrieved from
https://CRAN.R-project.org/package=stargazer
Lakens, D., Scheel, A. M., & Isager, P. M. (2018). Equivalence testing for psychological
research: A tutorial. Advances in Methods and Practices in Psychological Science,1(2),
259–269. doi:10.1177/2515245918770963
Lüdecke, D., Ben-Shachar, M. S., Waggoner, P., & Makowski, D. (2020a). see: Visualisation
toolbox for ’easystats’ and extra geoms, themes and color palettes for ’ggplot2’. CRAN.
doi:10.5281/zenodo.3952153
Lüdecke, D., Makowski, D., Waggoner, P., & Patil, I. (2020b). performance: Assessment of
regression models performance. CRAN. doi:10.5281/zenodo.3952174
Lüdecke, D., Waggoner, P., & Makowski, D. (2019). insight: A unied interface to access
information from model objects in R. Journal of Open Source Software,4(38), 1412.
doi:10.21105/joss.01412
Makowski, D., Ben-Shachar, M., & Lüdecke, D. (2019). bayestestR: Describing eects and
their uncertainty, existence and signicance within the bayesian framework. Journal of
Open Source Software,4(40), 1541. doi:10.21105/joss.01541
Pustejovsky, J. (2020). clubSandwich: Cluster-robust (sandwich) variance estimators
with small-sample corrections. Retrieved from https://CRAN.R-project.org/package=
clubSandwich
R Core Team. (2020). R: A language and environment for statistical computing. Vienna,
Austria: R Foundation for Statistical Computing. Retrieved from https://www.R-project.
org/
Revelle, W. (2019). psych: Procedures for psychological, psychometric, and personality
research. Evanston, Illinois: Northwestern University. Retrieved from https://CRAN.
R-project.org/package=psych
Robinson, D., Hayes, A., & Couch, S. (2020). broom: Convert statistical objects into tidy
tibbles. Retrieved from https://CRAN.R-project.org/package=broom
Saefken, B., Ruegamer, D., Kneib, T., & Greven, S. (2018). Conditional model selection in
mixed-eects models with cAIC4. ArXiv e-prints. Retrieved from https://arxiv.org/abs/
1803.05664
Zeileis, A. (2006). Object-oriented computation of sandwich estimators. Journal of Statistical
Software,16(9), 1–16. doi:10.18637/jss.v016.i09
Lüdecke et al., (2020). Extracting, Computing and Exploring the Parameters of Statistical Models using R. Journal of Open Source Software,
5(53), 2445. https://doi.org/10.21105/joss.02445
5