Agreement up to a constant offset per rater

The consistency form does not penalise a rater who is systematically high, because the column mean square is left out of the denominator. Right when only the ordering matters, wrong when the absolute number is the point.

Usage

IccC(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(C,1) = (MS_R - MS_E)/(MS_R + (k-1) MS_E).

References

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

Examples

IccC(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] -1.44898
#> 
#> $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(C,1)"
#> 
#> attr(,"class")
#> [1] "morie_rich_result" "list"