Runs all available equity audits (geographic, gender, funding, OA status)
on a corpus and returns a combined summary. Individual audit results can
be passed as ... arguments to avoid re-computation; any missing audits
are run with default parameters, except that the gender audit uses
method = "manual" and audits that send web requests run only when
online = TRUE.
Usage
# S3 method for class 'sm_audit_summary'
print(x, ...)
sm_audit_summary(corpus, ..., online = FALSE, call = rlang::caller_env())Arguments
- x
An audit object to print.
- ...
Optional pre-computed audit objects (e.g.,
sm_audit_geographic,sm_audit_gender,sm_audit_funding,sm_audit_oa). If provided, these are included directly rather than being re-run.- corpus
An
sm_corpusobject.- online
Logical. Whether audits that send web requests may run. Only the funding audit does:
sm_audit_funding()looks up every DOI on Crossref. With the defaultFALSE, a funding audit is included only when one is passed in...; otherwise it is reported as not run. SetTRUEto run it with its default Crossref source.- call
Caller environment for error reporting.
Value
An sm_audit_summary S3 object containing:
- geographic
An
sm_audit_geographicresult.- gender
An
sm_audit_genderresult.- funding
An
sm_audit_fundingresult. A funding audit that was not run has no funders,source = "none"andcoverage = NA.- oa
An
sm_audit_oaresult.- overview
A tibble with one row per audit:
audit,coverage(NAfor an audit that was not run) andtop_finding.
See also
Other audit:
sm_audit_funding(),
sm_audit_gender(),
sm_audit_geographic(),
sm_audit_oa()
Examples
corpus <- sm_example_corpus(n_works = 50, n_authors = 20)
# The funding audit sends batched Crossref requests, so it is not run
# unless online = TRUE.
summary_audit <- sm_audit_summary(corpus)
#> ℹ Skipped the funding audit, which sends batched Crossref requests.
#> ℹ Use `online = TRUE` to run it, or pass a precomputed `sm_audit_funding()`
#> result.
summary_audit$overview
#> # A tibble: 4 × 3
#> audit coverage top_finding
#> <chr> <dbl> <chr>
#> 1 geographic 1 DE (15.54%)
#> 2 gender 0 NA (100%)
#> 3 funding NA not run (needs online = TRUE)
#> 4 open_access 1 54% open access
# A precomputed audit is used as it is. This funding audit sends no
# request because its copy of the corpus has no DOIs.
no_doi <- corpus
no_doi$works$doi <- NA_character_
funding <- sm_audit_funding(no_doi)
#> No funding data: works lack DOIs or identifiers.
summary_audit <- sm_audit_summary(corpus, funding)
print(summary_audit)
#>
#> ── <sm_audit_summary> ──────────────────────────────────────────────────────────
#>
#> ── Overview
#> geographic: coverage 100% | DE (15.54%)
#> gender: coverage 0% | NA (100%)
#> funding: coverage n/a | no work could be checked
#> open_access: coverage 100% | 54% open access
#>
#> Use `print(x$geographic)`, `print(x$gender)`, etc. for detailed reports.
#>
#>
#> ── Limitations
#> • All equity audits rely on metadata that may be incomplete, inaccurate, or
#> biased toward certain regions and publishers.
#> • Coverage varies across audit dimensions; low coverage means results should be
#> interpreted with extreme caution.
#> • These audits describe patterns in the available metadata and should not be
#> treated as definitive measures of equity.
#> • We encourage transparent reporting of method, coverage, and limitations
#> alongside any equity analysis.