Statistical transformations of indices of gingivitis measured non-invasively

An index (PMGI) that combines the Massler-Schour (PMA) and Löe-Silness (GI) indices for non-invasive scoring of gingivitis was applied in 2 independent comparative clinical studies. 4 statistical transformations of the scores of individual papillary and marginal units were analyzed: the whole-mouth mean, its square root, the proportion of units judged to be inflamed, and its arc sin transformation. According to the criteria of how closely the transformations produced approximately normal distributions, and of how successful they were in detecting treatment differences, the square root transformation performed the best.