Package {psreplicate}

psreplicate: Access the Political Science Replication Index from R

Description

Search and browse the Political Science Replication Index (https://jsakowuah.github.io/polisci-replication/) without leaving R. The index aggregates replication-package metadata crawled monthly from flagship political science journals' Harvard Dataverse collections and tags each record by method and data type.

Getting started

• list_functions() lists every function this package exports.

• search_replications() searches and filters the whole index at once; search_by_title(), search_by_abstract(), search_by_author(), search_by_journal(), search_by_method(), search_by_data_type(), and search_by_year() do the same one field at a time.

• list_journals(), list_methods(), and list_data_types() show what values are actually valid to search on.

• dataset_info() and browse_dataset() look up or open one record by DOI.

• list_dataset_files() and download_dataset() get the actual replication files onto disk; load_dataset_file() loads one file straight into R with no disk step (all three require the dataverse package).

• refresh_index() downloads (or re-downloads) the index.

Author(s)

Maintainer: Joseph Akowuah siawjoe@gmail.com [copyright holder]

See Also

Useful links:

• https://github.com/jsakowuah/psreplicate

• https://jsakowuah.github.io/polisci-replication/

• Report bugs at https://github.com/jsakowuah/psreplicate/issues

Open a dataset's page in your web browser

Description

Looks up the record by DOI and opens its url with utils::browseURL(), so you can go from a search result straight to the dataset itself without copying a link.

Usage

browse_dataset(doi)

Arguments

Value

Invisibly, the URL that was opened.

Examples

browse_dataset("doi:10.7910/DVN/PRYYZ5")

Look up a single record by DOI

Description

Look up a single record by DOI

Usage

dataset_info(doi)

Arguments

Value

A one-row tibble::tibble(), or a zero-row tibble if no record matches (see search_replications() for the column definitions).

Examples

dataset_info("doi:10.7910/DVN/PRYYZ5")

Download a dataset's files

Description

Downloads a replication package's files to a local directory, without leaving R. Requires the dataverse package (install.packages("dataverse")); this is a thin convenience wrapper around it, scoped to this index's data and DOI format.

Usage

download_dataset(doi, dest = tempfile("psreplicate_"), files = NULL)

Arguments

Value

Invisibly, a character vector of the local file paths written. A file that fails to download (e.g. access-restricted) is skipped with a warning rather than aborting the rest of the download.

Examples

hit <- search_by_title("Rebel Victory and Authoritarian")
download_dataset(hit$doi[1])

List data-type tags used in the index

Description

The valid values to pass to search_by_data_type() or search_replications()'s data_type argument aren't documented anywhere else - this lists exactly what's actually in the data, most common first.

Usage

list_data_types()

Value

A tibble::tibble() with columns data_type and n_records.

Examples

list_data_types()

List the files in a dataset

Description

Previews what's available before downloading anything with download_dataset(). Requires the dataverse package (install.packages("dataverse")), which does the actual work of talking to Harvard Dataverse's file API.

Usage

list_dataset_files(doi)

Arguments

Value

A tibble::tibble() with columns filename, filesize (bytes), and content_type.

Examples

hit <- search_by_title("Rebel Victory and Authoritarian")
list_dataset_files(hit$doi[1])

List every function this package exports

Description

A quick reference, since the API is spread across several small, single-purpose functions rather than one do-everything call. Equivalent information is available via library(help = "psreplicate"), but this is meant to be more skimmable.

Usage

list_functions()

Value

A tibble::tibble() with columns function_name and description, in the order you'd typically reach for them.

Examples

list_functions()

List journals covered by the index

Description

List journals covered by the index

Usage

list_journals()

Value

A tibble::tibble() with one row per journal currently represented in the index: journal_short, journal_name, and n_records (the number of indexed replication packages).

Examples

list_journals()

List method tags used in the index

Description

The valid values to pass to search_by_method() or search_replications()'s method argument aren't documented anywhere else - this lists exactly what's actually in the data, most common first.

Usage

list_methods()

Value

A tibble::tibble() with columns method and n_records.

Examples

list_methods()

Load one file from a dataset directly into R

Description

Unlike download_dataset(), this never writes anything to disk. Tabular files (.tab/.dta/.csv/etc. that Dataverse has ingested) come back as a tibble; .rds files are deserialized directly with readRDS() as whatever R object they contain; other plain-text files (scripts, READMEs) come back as a character vector of lines; anything else (PDFs, images, other binaries) comes back as a raw vector, since there's no sensible in-memory representation otherwise. Requires the dataverse package (install.packages("dataverse")).

Usage

load_dataset_file(doi, filename)

Arguments

Value

A tibble::tibble() for tabular files, whatever R object was saved for .rds files, a character vector for plain-text files, or a raw vector otherwise.

Examples

hit <- search_by_title("Rebel Victory and Authoritarian")
load_dataset_file(hit$doi[1], "area.tab")

Download or refresh the local copy of the replication index

Description

Downloads the current Political Science Replication Index and stores it in a per-session-independent, per-user cache directory (see tools::R_user_dir()), so subsequent calls to search_replications() and list_journals() don't need network access. The cache is refreshed automatically once it is more than a day old; call this directly to force an update sooner.

Usage

refresh_index(force = FALSE)

Arguments

Value

Invisibly, the path to the cached index file.

Examples

refresh_index()

Search by abstract only

Description

Matches only the dataset's abstract/description text (the description column), not title or authors.

Usage

search_by_abstract(query)

Arguments

Value

A tibble::tibble() of matching records (see search_replications() for the column definitions).

Examples

search_by_abstract("difference-in-differences")

Search by author name

Description

Matches only the authors column (each a "Last, First; Last, First" string), not title or abstract.

Usage

search_by_author(query)

Arguments

Value

A tibble::tibble() of matching records (see search_replications() for the column definitions).

Examples

search_by_author("Wantchekon")

Get all records tagged with a given data type

Description

Get all records tagged with a given data type

Usage

search_by_data_type(data_type)

Arguments

Value

A tibble::tibble() of matching records (see search_replications() for the column definitions).

Examples

search_by_data_type("Elections / Voting")

Get all records from a journal

Description

Get all records from a journal

Usage

search_by_journal(journal)

Arguments

Value

A tibble::tibble() of matching records (see search_replications() for the column definitions).

Examples

search_by_journal("AJPS")
search_by_journal(c("AJPS", "APSR"))

Get all records tagged with a given method

Description

Get all records tagged with a given method

Usage

search_by_method(method)

Arguments

Value

A tibble::tibble() of matching records (see search_replications() for the column definitions).

Examples

search_by_method("Survey Experiment")

Search by title only

Description

Unlike search_replications()'s query argument, which matches across title, abstract, and authors together, this matches only the title.

Usage

search_by_title(query)

Arguments

Value

A tibble::tibble() of matching records (see search_replications() for the column definitions).

Examples

search_by_title("regression discontinuity")

Get all records published in given year(s)

Description

Get all records published in given year(s)

Usage

search_by_year(year)

Arguments

Value

A tibble::tibble() of matching records (see search_replications() for the column definitions).

Examples

search_by_year(2020:2024)

Search the Political Science Replication Index

Description

Search the Political Science Replication Index

Usage

search_replications(
  query = NULL,
  journal = NULL,
  method = NULL,
  data_type = NULL,
  year = NULL
)

Arguments

Value

A tibble::tibble() of matching records, with one row per replication package and columns journal_short, title, doi, url, authors (a single "Last, First; Last, First" string), year, description, method_tags, data_type_tags.

Examples

search_replications(query = "regression discontinuity", journal = "AJPS")
search_replications(method = "Survey Experiment", year = 2020:2024)