Computes several collaboration indicators for the corpus, including the Collaboration Index (CI), mean authors per paper, proportion of multi-authored papers, and international collaboration rate.
Usage
sm_metric_collab_index(corpus, call = rlang::caller_env())Arguments
- corpus
An sm_corpus object with a populated
authorshipstable.- call
Caller environment for error reporting.
Value
A tibble with one row per work and columns:
work_idThe work identifier.
n_authorsNumber of authors on the work.
n_countriesDistinct known countries;
NAif none are known.n_institutionsDistinct known
institution_idvalues;NAif none are known. Affiliation match labels are not counted.is_internationalLogical;
TRUEif more than one country is known;FALSEif exactly one is known, andNAif none are known. Missing codes can hide international collaboration.is_multi_authoredLogical;
TRUEif more than one author.
Details
The Collaboration Index (CI) for a set of works is typically defined as
the mean number of authors per paper. This function returns per-work
data, from which aggregate CI can be computed by the user.
Country and institution counts are NA when no non-blank value is known.
With one known country, is_international is FALSE, even if some
affiliations are missing. Institutions are counted by institution_id
only; names and affiliation text are not matched.
Examples
corpus <- sm_example_corpus()
collab <- sm_metric_collab_index(corpus)
head(collab)
#> # A tibble: 6 × 6
#> work_id n_authors n_countries n_institutions is_international
#> <chr> <int> <int> <int> <lgl>
#> 1 W000000001 1 1 NA FALSE
#> 2 W000000002 6 4 NA TRUE
#> 3 W000000003 4 4 NA TRUE
#> 4 W000000004 5 5 NA TRUE
#> 5 W000000005 4 3 NA TRUE
#> 6 W000000006 1 1 NA FALSE
#> # ℹ 1 more variable: is_multi_authored <lgl>
# Aggregate collaboration index:
mean(collab$n_authors)
#> [1] 3.775