Skip to contents

Simulated balanced two-way ratings, twenty subjects by four raters

Source: R/data.R
ratings_twoway.Rd

Simulated data, not a real study. Twenty subjects are each scored once by the same four raters. The variance components are 0.6 (subject), 0.1 (rater), and 0.2 (residual), so the population ICC(A,1) is 0.667. The design is large enough for ci_method = "mpl", which the six-subject ratings data are not. The "Interval methods" article uses it.

Usage

ratings_twoway

Format

A data frame with 80 rows and 3 columns, as in ratings (subject, rater, score).

Source

Simulated by data-raw/make-vignette-data.R with set.seed(88). The score is a subject effect plus a rater effect plus noise, all normal draws with the variances above.

See also

ratings for the Shrout and Fleiss worked example.

Examples

str(ratings_twoway)
#> 'data.frame':	80 obs. of  3 variables:
#>  $ subject: Factor w/ 20 levels "1","2","3","4",..: 1 2 3 4 5 6 7 8 9 10 ...
#>  $ rater  : Factor w/ 4 levels "1","2","3","4": 1 1 1 1 1 1 1 1 1 1 ...
#>  $ score  : num  0.307 0.2 1.613 -1.985 0.283 ...
icc(ratings_twoway, score, subject, rater, type = "agreement",
  ci_method = "mpl")
#> ── Intraclass correlation: two-way random, absolute agreement ──────────────────
#> Subjects: 20 | Raters: 4 (random) | Observations: 80 of 80 cells (complete)
#> Engine: glmmTMB (REML) | CI: 95% modified profile likelihood (closed form)
#> 
#>   index     estimate   95% CI
#>   ICC(A,1)     0.709   [0.425, 0.865]
#>   ICC(A,k)     0.907   [0.747, 0.963]
#> 
#> Variance components: subject 0.652, rater 0.042, residual 0.226
#> Shrout & Fleiss equivalent: ICC(A,1) = ICC(2,1), ICC(A,k) = ICC(2,k)