Audit reporting coverage across many studies

Summarises how consistently a set of studies reported each framework item. This is the corpus-level view: instar_report() describes one study, instar_audit() describes a literature.

Usage

instar_audit(x, ...)

Arguments

x

What to audit. Any of:

• a path to a directory of sheets, or a vector of file paths, which are read with read_instar();

• an instar_corpus;

• a list of instar_report objects, or of items tables;

• a data frame of scores in wide form, passed to audit_from_matrix().

...

Passed to read_instar() when x is a path, or to audit_from_matrix() when x is a data frame.

Value

An object of class instar_audit: a list with

• items, one row per framework item with counts and the percentage of applicable studies that reported it;

• studies, one row per study with its own coverage (as summary.instar_corpus() returns);

• long, one row per study-item pair, the tidy form to compute on;

• n, the number of studies.

See also

read_instar() to build a corpus, audit_from_matrix() for retrospective scoring.

Examples

if (FALSE) { # \dontrun{
# Prospective: a folder of deposited sheets
audit <- instar_audit("supplements/")
audit
summary(audit)
plot(audit)

# Which items does the literature handle worst?
head(summary(audit)[order(summary(audit)$percent_reported), ])
} # }