method.WILS: Weighted and Iterated Least Squares (WILS) Method

In PublicationBiasBenchmark: Benchmark for Publication Bias Correction Methods

Weighted and Iterated Least Squares (WILS) Method

Description

Implements the weighted and iterated least squares (WILS) method for publication bias correction in meta-analysis. The method is based on the idea of using excess statistical significance (ESS) to identify how many underpowered studies should be removed to reduce publication selection bias. See \insertCitestanley2024harnessing;textualPublicationBiasBenchmark for details.

Usage

## S3 method for class 'WILS'
method(method_name, data, settings = NULL)

Arguments

method_name	Method name (automatically passed)
data	Data frame with yi (effect sizes) and sei (standard errors)
settings	List of method settings (see Details)

Details

The WILS method has two implementation versions based on Stanley & Doucouliagos (2024). The following settings are implemented

"default"

The simulation version (default) uses residuals from the t ~ Precision regression for the first iteration, then switches to individual excess statistical significance (ESS) for subsequent iterations.

"example"

The example version consistently uses residuals from the t ~ Precision regression to identify studies to remove across all iterations.

Value

Data frame with WILS results

Author(s)

František Bartoš [e-mail]

References

\insertAllCited

Examples

# Generate some example data
data <- data.frame(
  yi = c(0.2, 0.3, 0.1, 0.4, 0.25),
  sei = c(0.1, 0.15, 0.08, 0.12, 0.09)
)

# Apply WILS method
result <- run_method("WILS", data)
print(result)

PublicationBiasBenchmark documentation built on March 16, 2026, 5:07 p.m.