MD_IRR: Percent Inter-Rater Reliabilitiy
In trinker/metaDAT: Tools for Data Management in Meta-Analysis

Description Usage Arguments Value Note Examples

Percent inter-rater reliabilitiy by variable.

  MD_IRR(meta.obj.1, meta.obj.2,
    exclude.vars = c("author", "ID", "time"),
    cor.var = NULL, comp.var = "coder", file = NULL,
    order.by.comp.var = TRUE)

`meta.obj.1`	An MDlist (an object from `read.MD`) for time 1 (primary coding).
`meta.obj.2`	An MDlist (an object from `read.MD`) for time 2 (secondary coding).
`exclude.vars`	A character vector of variables to exclude from reliability check.
`cor.var`	An optional argument for the MDlist variable that is a list of correlation matrices (can detect and handle correlation variables that are in character correlation triangle format).
`comp.var`	The variable to compare for discrepencies between the two MDlists. Default is "coder". This argument is particularly useful when there are more than two coders.
`file`	Optional connection, or a character string naming the file to print to.
`order.by.comp.var`	logical. If TRUE orders the study disagreements by the comparison variable.

Returns a list:

`Variable_Agreement_Rate`	A dataframe of percentage agreement/disagreement for each variable. Final row is overall agreement rate.
`Disagreements`	A list of disagreements of study variables between comparison groups (time 1 and time 2) per study. If correlations are being compared the function outputs an agreement matrix. Agreement is marked as `-` and disagreement as `X`.

This function assumes double coding, that is all studies have been coded by two different coders.

path <- system.file("extdata/bibTest.bib", package = "metaDAT")
dat1 <- read.MD(path)
path2 <- system.file("extdata/bibTest2.bib", package = "metaDAT")
dat2 <- read.MD(path2)
MD_IRR(dat1, dat2, cor.var = "correlation_triangle")
# MD_IRR(dat1, dat2, cor.var = "correlation_triangle", file = "foo.txt")
# unlink("foo.txt", recursive = TRUE, force = FALSE) #delete previous file