Table 3. Typical ratings and features of prediction tasks proposed under the three instruction types (the no-example baseline, the obesity example, and the savings example).

Rating or feature (variable type)	Mean x baseline	Mean x obesity	Mean x savings
Perceived task importance (x = 1–5 Likert; 5: Strongly agree)	3.09	3.16	3.50
Inputs are useful (x = 1–5 Likert; 5: Strongly agree)	3.36	3.91	3.67
Questions are answerable (x = 1) or unanswerable (x = 0)	0.82	0.92	0.92
Questions are objective (x = 1) or subjective (x = 0)	0.59	0.67	0.68
Questions are numeric (x = 1) or Boolean (x = 0)	0.25	0.35	0.60

Note:

Bold treatment quantities show a significant difference (p < 0.05) from the baseline (Chi-square tests for Likert x; Fisher exact tests for binary x).