Assistance
The multinomial regression model for tester assistance during task 1 was not significant (ꭓ2 (2) = 1.27, p = 0.53). The model for task 2 (ꭓ2 (2) = 8.62, p = 0.01) was statistically significant and explained 32.3% (Nagelkerke R2) of the variance in tester assistance across groups. During task 2, it was 6.67 times more likely for children to seek partial assistance than no assistance if they were in the DS group compared to the TD group (Wald’s ꭓ2 = 3.04, p = 0.08). Similarly, it was 20 times more likely to seek total assistance than no assistance for children in the DS group compared to the TD group (Wald’s ꭓ2 = 5.61, p < 0.05). The model for task 3 (ꭓ2 (2) = 9.24, p = 0.01) was also statistically significant and explained 33.9% (Nagelkerke R2) of the variance in tester assistance across groups. During task 3, children were 21.33 times more likely to seek partial assistance than no assistance if they are in the DS group compared to TD group (Wald’s ꭓ2 = 5.92, p< 0.05). The percent data for tester assistance during the three tasks is provided in Figure 3.