DOCSEARCHbyAlgolia
Submit Docs

metadat: Meta-Analysis Datasets for R

website·Indexed 29 days ago

https://wviechtb.github.io/metafor/reference/index.html

A collection of meta-analysis datasets for use in R, providing structured data from various studies across medical, psychological, and social science domains. The package includes technical metadata, study descriptions, and a data preparation function (prep_dat), with practical examples for computing effect sizes and fitting models using the metafor package.

Tokens
213.7K
Snippets
870
Records
1.5K
Agent score
74%

What's inside metadat

  1. Overview of metadat R package datasets

    The metadat package for R provides a collection of meta-analysis datasets used for technical demonstrations and study descriptions. Available datasets include:
    # Example: Accessing a dataset (assuming the package is loaded)
# data(dat.aloe2013) # Studies on Supervision Quality
# data(dat.bcg)       # Studies on BCG Vaccine Effectiveness

  2. Overview of the metadat R package

    The metadat package provides a comprehensive collection of meta-analysis datasets for use in R. These datasets include various study samples and associated descriptions across multiple domains such as medicine, psychology, and social sciences.

  3. Overview of the Generation Effect dataset (dat.mccurdy2020)

    The dat.mccurdy2020 dataset contains results from 126 articles, 310 experiments, and 1,653 mean recall estimates. It is used to study the 'generation effect'—the phenomenon where self-generated information is better remembered than read or experimenter-provided information. The data is structured to support multilevel meta-regression and cluster-robust inference, accounting for the complex hierarchy of articles, experiments, and samples.

  4. Overview of metafor package capabilities

    The metafor package is a comprehensive R library for conducting meta-analyses. Its core capabilities include:

    • Calculating various effect sizes and outcome measures.
    • Fitting equal-, fixed-, random-, and mixed-effects models.
    • Performing moderator and meta-regression analyses.
    • Creating meta-analytical plots such as forest, funnel, radial, L’Abbé, Baujat, bubble, and GOSH plots.
    • Specialized methods for binomial and person-time data, including Mantel-Haenszel, Peto’s method, and generalized linear mixed-effects models (logistic and Poisson regression).
    • Fitting multivariate/multilevel models to account for non-independent sampling errors or true effects (e.g., clustering, multiple endpoints).
    • Conducting network meta-analyses and analyses accounting for known correlation structures (e.g., phylogenetic relatedness).

  5. Overview of the metafor R package

    The metafor package is a comprehensive collection of functions for conducting meta-analyses in R. It supports calculating various effect size measures and fitting equal-effects, fixed-effects, and random-effects models. Key capabilities include:

    • Meta-regression: Fitting mixed-effects models by including study-level variables (moderators) as predictors.
    • Specialized Methods: Support for 2x2 tables, proportions, and incidence rates using Mantel-Haenszel, Peto's method, and generalized linear mixed-effects models (logistic and Poisson regression).
    • Complex Data: Multilevel and multivariate models for non-independent estimates (e.g., clustering or correlated sampling errors).
    • Diagnostics: Tools for assessing model fit, identifying outliers/influential studies (Cook's distances, standardized residuals), and sensitivity analyses (leave-one-out).
    • Inference: Implementation of Knapp and Hartung method, permutation tests, and robust variance estimation.
    • Visualization: Functions for forest, funnel, radial, normal Q-Q, L'Abbé, Baujat, bubble, and GOSH plots.
    • Publication Bias: Tools for examining small-study effects via Egger's regression, rank correlation, trim and fill, and selection models.

  6. Overview of metafor transformation functions

    The metafor package provides a suite of transformation functions to prepare data for meta-analysis. These functions handle common conversions such as Fisher's r-to-z, logit, probit, and arcsin transformations, as well as conversions between different effect size metrics (e.g., converting standardized mean differences to correlations).
    # Examples of available transformation functions:
transf.rtoz(xi)      # Correlation to Fisher's z
transf.ztor(xi)      # Fisher's z back to correlation
transf.logit(xi)    # Logit transformation
transf.ilogit(xi)    # Inverse logit
transf.probit(xi)   # Probit transformation
transf.iprobit(xi)   # Inverse probit

  7. Overview of metafor package capabilities

    The metafor package is a comprehensive R library for conducting meta-analyses. Its core capabilities include:

    • Effect Size Calculation: Functions to calculate various outcome measures.
    • Model Fitting: Support for equal-effects, fixed-effects, random-effects, and mixed-effects models.
    • Analysis: Tools for moderator and meta-regression analyses.
    • Visualization: Creation of forest, funnel, radial, L’Abbé, Baujat, bubble, and GOSH plots.
    • Specialized Methods: Implementation of Mantel-Haenszel and Peto’s methods, as well as generalized linear mixed-effects models (logistic and Poisson regression) for binomial and person-time data.
    • Advanced Modeling: Support for multivariate/multilevel models (accounting for non-independent sampling errors or clustering), network meta-analyses, and analyses accounting for known correlation structures (e.g., phylogenetic relatedness).

  8. Overview of the dat.knapp2017 dataset

    The dat.knapp2017 dataset contains results from 31 studies comparing planning performance (typically using the Tower of London test or similar cognitive tasks) between schizophrenia patients and healthy controls. The primary outcome measure is the standardized mean difference (SMD), where positive values indicate better performance by the healthy control group.

  9. Browse and search for meta-analysis datasets

    You can list all available datasets in the metadat package using the help() function. To find specific datasets based on concept terms (e.g., research field, outcome measure, or model type) or via full-text search of help files, use the datsearch() function.
    # List all datasets in the package
help(package = "metadat")

# Search for datasets using concept terms or full-text search
metadat::datsearch()

  10. Use the pignon2000 dataset for head and neck cancer meta-analysis

    The dat.pignon2000 dataset contains results from 65 trials examining mortality risk in patients with nonmetastatic head and neck squamous-cell carcinoma. It compares locoregional treatment plus chemotherapy against locoregional treatment alone. The data is structured to allow the estimation of log hazard ratios and sampling variances.

  11. Search for datasets using datsearch()

    The datsearch() function allows you to find datasets within the metadat package. You can search by 'concept terms' (tags related to research fields, outcome measures, or models) or perform a full-text search of the dataset help files. If a single match is found, the help file is displayed automatically; if multiple matches are found, the user is prompted to select one.
    # Start in interactive mode to be prompted for search terms
datsearch()

# Search for a specific concept term
datsearch("standardized mean differences")

# Search for multiple concept terms (AND logic by default)
datsearch("odds ratio, multilevel")

# Perform a full-text search instead of a concept search
datsearch("infarct", concept = FALSE)

  12. Understand the St. John's Wort depression dataset (dat.linde2005)

    The dat.linde2005 dataset contains results from 26 double-blind placebo-controlled trials. It is used to analyze the effectiveness of Hypericum perforatum extracts for treating depression. The dataset includes information on sample sizes, HRSD (Hamilton Depression Rating Scale) scores, dosage, and response rates for both treatment and placebo groups.