litsearchr

an R package to facilitate quasi-automatic search strategy development for systematic reviews

2020 Commendation from the Society for the Improvement of Psychological Science

Systematic review, meta-analysis, and other forms of evidence synthesis require researchers to identify the body of evidence relevant to their research question. Partially automating evidence synthesis would reduce the amount of human time and effort needed to conduct a systematic review and reduce bias in keyword selection when researchers develop search strategies.

The litsearchr R package facilitates quick, objective, reproducible search strategy development using text-mining and keyword co-occurrence networks to identify important terms to include in a search strategy as described in Grames et al. (2019). It can automatically write Boolean search strings in up to 53 different languages, with stemming support for English. To assess the quality of a search, it can also check the results of a search against a set of known, relevant articles to get performance metrics.

Getting Started

litsearchr is designed for users of any level, whether you have no experience coding in R, have never done a systematic review before, or are a seasoned expert.

New! Graphical user interface

If you want to use litsearchr but without any coding involved, there is now a GUI! It is still in development, so there may be bugs, but you can execute most of the core litsearchr functions by clicking and pointing.
https://elizagrames.shinyapps.io/litsearchr/

Getting started in R and RStudio

litsearchr is an R package, which means you will need to have R and RStudio installed on your computer to use it. Get help installing R and RStudio, learn about working in the R environment, and navigate RStudio with these resources:

The Comprehensive R Archive Network (CRAN)

RStudio

RStudio Education

Installing litsearchr

Install the remotes package in R using install.packages("remotes")

Load the remotes package by running library(remotes)

Install litsearchr by running install_github("elizagrames/litsearchr", ref="main")

Alternatively, you can download the litsearchr source code archive and build the package yourself locally.

Library Carpentry

If you are starting from the beginning, you may want to attend community-taught Software Carpentry and Library Carpentry workshops or work through some of these lesson materials at your own pace:

Library Carpentry: Introduction to R

Library Carpentry: Introduction to R and litsearchr (in development)

Software Carpentry: Introduction to R and RStudio

Tutorials and Documentation

Identifying Search Terms for a Systematic Review: A Demonstration of the litsearchr Package

Note: The video demonstration is for an OLD version (v0.1.0) of litsearchr and some of the content and specific functions are outdated though the general approach to using the package remains the same. Please refer to the vignette for the current version of litsearchr and the function help files for the most recent instructions.

Current Vignette: Search term selection with litsearchr v1.0.0 for an example systematic review of the effects of fire on black-backed woodpeckers
Useful litsearchr tutorial created by Luke Tudge on the subject of medication and cognitive behavioral therapy for phobias
This in-development Library Carpentry lesson introduces people working in library- and information-related roles to using R and RStudio when working on systematic reviews. In addition to the basics of R and RStudio, you will learn to use litsearchr.
Archived: Introduction to litsearchr v0.1.0 with an example of writing a systematic review search strategy for Black-backed Woodpecker occupancy of post-fire forest systems (corresponds to Grames et al. 2019, Methods in Ecology and Evolutionary Biology)

Report a Bug or Make a Feature Request

If you run into a problem where a function in litsearchr is not behaving as expected, please open an Issue on the litsearchr GitHub repository or email the developer, Eliza, at eliza.grames@uconn.edu. When describing the problem or requesting a feature, please make sure to:

Check that you do not have any typos that could be creating the error, especially extra spaces, commas, or periods, and that the error is actually being produced by litsearchr or one of its dependencies, not your computer.
Include the error message or describe why the output was not what was expected, and copy the code snippet that creates the error including all the options you used (i.e. not just the function name). We need to be able to reproduce the error in order to fix it.
If possible and there is no sensitive information in your data files, include a copy of your environment when you got the error (i.e. save.image() and attach the .RData file to your email or issue) since that can be really helpful for troubleshooting problems with text processing.

Report a Bug

If you think of a feature that would be particularly useful to you, and probably also to other users of litsearchr, feel free to suggest it. We can't promise to implement all suggestions right away since all litsearchr updates and upkeep are done by one PhD student in her spare time, but we will try!

The First "Bug"
At litsearchr, we love bugs and insects of all kinds (check out the EntoGEM project!). The first programming "bug" was a moth found in Mark II, a machine being developed by one of the first computer programmers, Grace Hopper.

Contribute

If you have an idea for a feature request, you can either initiate a request (see section above), or you can try to come up with a fix or new feature yourself. You can also work on fixes to open Issues on the litsearchr repository that others have created.

Feel free to fork the litsearchr repository, add or modify functions to make litsearchr capable of doing more, and then once you've checked that the changes (probably) don't break any other functions, you can make a pull request and once Eliza has double-checked that the feature increases the functionality of litsearchr and works well with the other functions, she will merge it into the development version of the package for the next release.

Related Tools and Resources

synthesisr
Import, assemble, and deduplicate bibliographic data stored as RIS or BibTex formatted text files.

revtools
Article screening for evidence synthesis using manual or visual methods.

topictagger
Tag documents with meta-data and based on user-defined ontologies or topic groups.

metaDigitise
High-throughput, reproducible extraction of data from figures.

robvis
Create publication quality risk-of-bias assessment figures.

Evidence Synthesis Hackathon
A series to bring together interested researchers, practitioners and coders to discuss and develop new Open Source technologies for evidence synthesis applications.

About the Developer

Eliza Grames is a PhD Candidate in Ecology and Evolutionary Biology at the University of Connecticut. She works on developing partially automated methods and software to reduce the time and effort needed for evidence synthesis projects and applies these approaches to bird and insect conservation.

Her dissertation research combines field work observations of bird behaviors, nest predation, and food availability with meta-analytic structural equation modeling and Bayesian path analysis to understand why small forest fragments are less densely populated by breeding birds than larger fragments and how the underlying mechanisms interact.