Conducting and Visualizing Specification Curve Analyses

The goal of specr is to facilitate specification curve analyses (Simonsohn, Simmons & Nelson, 2019; also known as multiverse analyses, see Steegen, Tuerlinckx, Gelman & Vanpaemel, 2016). It can be used to investigate how different (theoretically plausible) analytical choices affect outcome statistics within the universe of one single data set.

It provides functions to setup, run, evaluate, and plot the multiverse of specifications. A simple usage example is provided below. For more information about the various functions and specific use cases, visit the documentation.

There are also some vignettes that exemplify and explain specific aspects and functions of the package:

Getting started: A comprehensive example. This vignette illustrates the major functions of the package.
Customizing specification curve plots: This vignette exemplifies various ways to plot the specification curve.
Decomposing the variance of the specification curve: An example of how to investigate variance components of the specification curve.
Visualizing progress during estimation: This vignette explains how to create a progress bar for longer computations.

Disclaimer

We do see a lot of value in investigating how analytical choices affect a statistical outcome of interest. However, we strongly caution against using specr as a tool to somehow arrive at a better estimate. Running a specification curve analysis does not make your findings any more reliable, valid or generalizable than a single analysis. The method is only meant to inform about the effects of analytical choices on results, and not a better way to estimate a correlation or effect.

Installation

Install specr from CRAN:

install.packages("specr")

Or install the most recent development version from GitHub with:

# install.packages("devtools")
devtools::install_github("masurp/specr")

Usage

Using specr is comparatively simple. The main function is run_specs() in which analytical choices are specified as arguments. The function plot_specs() can then be used to visualize the results.

library(specr)

# Run specs
results <- run_specs(df = example_data, 
                     y = c("y1", "y2"), 
                     x = c("x1", "x2"), 
                     model = c("lm"), 
                     controls = c("c1", "c2"), 
                     subsets = list(group1 = unique(example_data$group1),
                                    group2 = unique(example_data$group2)))
# Result frame
head(results)
#> # A tibble: 6 x 12
#>   x     y     model controls estimate std.error statistic  p.value conf.low
#>   <chr> <chr> <chr> <chr>       <dbl>     <dbl>     <dbl>    <dbl>    <dbl>
#> 1 x1    y1    lm    c1 + c2     4.95      0.525     9.43  3.11e-18    3.92 
#> 2 x2    y1    lm    c1 + c2     6.83      0.321    21.3   1.20e-57    6.20 
#> 3 x1    y2    lm    c1 + c2    -0.227     0.373    -0.607 5.44e- 1   -0.961
#> 4 x2    y2    lm    c1 + c2     0.985     0.324     3.04  2.62e- 3    0.347
#> 5 x1    y1    lm    c1          5.53      0.794     6.97  2.95e-11    3.96 
#> 6 x2    y1    lm    c1          8.07      0.557    14.5   6.90e-35    6.98 
#> # … with 3 more variables: conf.high <dbl>, obs <int>, subsets <chr>

# Plot
plot_specs(results, choices = c("x", "y", "controls", "subsets"))

How to cite this package

citation("specr")
#> 
#> To cite 'specr' in publications use:
#> 
#>   Masur, Philipp K. & Scharkow, M. (2019). specr: Statistical functions
#>   for conducting specification curve analyses. Available from
#>   https://github.com/masurp/specr.
#> 
#> A BibTeX entry for LaTeX users is
#> 
#>   @Misc{,
#>     title = {specr: Statistical functions for conducting specification curve analyses (Version 0.2.1.9000)},
#>     author = {Philipp K. Masur and Michael Scharkow},
#>     year = {2019},
#>     url = {https://github.com/masurp/specr},
#>   }

References

Simonsohn, U., Simmons, J. P., & Nelson, L. D. (2019). Specification Curve: Descriptive and Inferential Statistics for all Plausible Specifications. Available at: https://doi.org/10.2139/ssrn.2694998

Steegen, S., Tuerlinckx, F., Gelman, A., & Vanpaemel, W. (2016). Increasing Transparency Through a Multiverse Analysis. Perspectives on Psychological Science, 11(5), 702-712. https://doi.org/10.1177/1745691616658637

specr