5.2 Question 2: Which of the applied measures (phonological awareness, executive functioning, or phonological working memory) is the best predictor of numerical performance? Does the predictor change when looking at different age groups?

Since the BRIEF-P Working Memory and Planning and Organizing subscales were highly correlated, the two scales were combined according to the BRIEF-P protocol. For clarity of communication, we will refer to this combination as BRIEF-P Working Memory/Planning & Organizing subscales. Raw scores were used for all variables and age was corrected within the models.

## 5.2.1 Predictors of early numeracy performance

Regression analyses using the enter method were performed to determine which linguistic and working memory measures predicted TEMA-3 performance and to determine if care type was a significant predictor (see Table 3 for regression results). All four models indicated a significant effect: Model 1, including age, care type, BRIEF-P Working Memory/Planning & Organizing subscales, and real words repetition task; Model 2, which substituted non-words repetition task for the real words repetition task; Model 3, which added PPVT-III; and Model 4 which included PALS. The best fit was Model 4 and accounted for 76% of the variance in TEMA-3 performance, with age (t[71] = 4.60, p < 0.001), PPVT-III (t[71] = 2.45, p < 0.05), and PALS (t[71] = 4.95, p < 0.001) as significant predictors.
