The intraclass correlations (ICC) of consistency for rater reliability using
the variance estimates from a linear mixed model. The function returns the
ICC, standard error of measurment (sem) and confidence intervals for ICC.

`icc_consistency(data, cols = colnames(data), alpha = 0.05, twoway = FALSE)`

## Arguments

- data
data.frame with a column for each observer/rater and a row per
rated subject.

- cols
character vector with the column names to be used as observers.
Default is `cols = colnames(data)`.

- alpha
confidence interval level, default `alpha = 0.05`.

- twoway
logical indicator if the variance components are estimated from
the two-way model default: `twoway = FALSE`.

## Details

The ICC type consistency is the variance between the subjects divided by the sum
of the subject variance and the residual variance. The subject variance and
error variance are adjusted for the fixed rater effect, accordingly the rater
variance is not used to calculate the ICC. The ICC for consistency generalizes
only to the fixed set of raters in the data (Shrout & Fleiss,
1979). The `icc_model()` function is used to compute the variances.
This is a `lmer` model with a random slope for the subjects as well as for
the raters. The sem is the square root of the error variance.
The confidence are computed with the exact F method. F = (k * subject variance +
error variance)/ error variance, with df1 = n - 1 and df2 = (n - 1) * (k - 1)
(Shrout & Fleiss, 1979).

## References

Fleiss, J. L., & Shrout, P. E. Approximate interval estimation for a certain
intraclass correlation coefficient. Psychometrika, 1978, 43, 259-262.