Transforming scales, such as the POES, to interval scaling in a valid way is essential to support the evaluation of evidence-based mental health care practice.
A complication of transformations is possible re-transformation bias in switching from the transformed scale to the original dollar-based scale.
Therefore, heritability on this transformed scale depended on the particular transformation used and thus, did not provide robust heritability estimates.
These highly skewed graphs straighten out when graphed on a log-log transform scale.
A routine practice in the analysis of repeated measurement data is to represent individual responses by a mixed effects model on some transformed scale.
As with the mixed models fit with lme4, the Bayesian models were run on transformed data, so the means and standard deviations are on the transformed scale.
However, the use of the transformed scale did not make a meaningful difference to the pattern of results, therefore the untransformed scale is presented for ease of interpretation.
A new transformed scale based on rasch analytical principles is promising combining features of both tools for evaluation of functional recovery of critically ill.
One of these drawbacks is that the predictions are not robust enough to detect the heteroscedasticity in the transformed scale [1,4].
Variance testing (Levene) revealed that the variances were similar in the transformed scale.
