Four indices were significant predictors in the regression: USF, SemD, McD, and EAT_types.
In addition, two other indices (SemD and McD) demonstrated medium effect sizes (r > .30), while two (EAT_tokens and EAT_types) demonstrated small effect sizes (r > .10) (Cohen, 1988).
Interestingly, an inverse relationship between semantic ambiguity (SemD) and lexical variability (McD) was reported: As speakers become more lexically proficient, they use words that are more semantically ambiguous but also more lexically distinct.
The second strongest predictor of lexical proficiency was SemD, which explained another 10% of variance in ratings.
Table 5 and Figure 1 also indicate that Sample A had higher SemD and McD values than Sample B, as predicted by our model, suggesting that more proficient speakers produce words that are more semantically diverse at the same time that they are more lexically distinct.
In addition, none of the indices were highly inter-correlated (the highest correlation was between USF and SemD: r = -.337, p > .001).
Radiographs of these two affected individuals showed skeletal changes involving the spine, metaphyses and epiphyses of tubular bones descriptively termed SEMD (Figs 3 - 7).
(1) As the most common of the SEMD group of disorders, it is estimated to affect 1 in every 20 000 individuals.