Agreement that counts a systematic rater offset as disagreement

The extra k (MS_C - MS_E)/n term in the denominator is the whole difference from the consistency form: it charges for between-rater bias. Two raters who agree on the ordering but differ by a constant score high on consistency and low here.

Usage

IccA(y, subject, rater)

Arguments

y: Ratings.
subject: Subject label.
rater: Rater label.

Value

List with estimate, ms_r, ms_c, ms_e, k, n_subjects.

Details

Formula: \(ICC(A,1) = (MS_R - MS_E) / [MS_R + (k-1) MS_E + k (MS_C - MS_E)/n]\).

References

McGraw & Wong (1996) Psychol Methods 1:30-46, table 4; Shrout & Fleiss (1979) Psychol Bull 86:420-428.

Examples

IccA(y = c(1, 2, 3, 4, 5, 6, 7, 8), subject = c(1, 2, 3, 4, 5, 6, 7, 8),
  rater = c(1, 2, 3, 4, 5, 6, 7, 8))
#> $estimate
#> [1] 3.227273
#> 
#> $ms_r
#> [1] 48
#> 
#> $ms_c
#> [1] 48
#> 
#> $ms_e
#> [1] -12.85714
#> 
#> $k
#> [1] 8
#> 
#> $n_subjects
#> [1] 8
#> 
#> $method
#> [1] "Intraclass correlation ICC(A,1)"
#> 
#> attr(,"class")
#> [1] "morie_rich_result" "list"