Table 4. Impact of correct causal attribution on the urgency with which help would be sought.

Vignette	Kruskal-Wallis Test $\left({\chi }_2^2,p\right)$	2-way comparison	Dunn’s Test, (Z, p-adj)
A) Joint stiffness and pain	8.12, 0.02	Correct vs. Approximate	-0.97,0.99
		Approximate vs. Incorrect	2.23, 0.08
		Correct vs. Incorrect	2.49, 0.04
B) Joint stiffness, pain and swelling wrists and hands	6.18, 0.04	Correct vs. Approximate	-1.57,0.35
		Approximate vs. Incorrect	1.08, 0.84
		Correct vs. Incorrect	2.46, 0.04
C) Joint stiffness, pain and swelling feet and ankles	0.70, 0.70	Correct vs. Approximate	n/a
		Approximate vs. Incorrect	n/a
		Correct vs. Incorrect	n/a
D) Bowel Cancer	81.05, <.001	Correct vs. Approximate	-2.56,0.03
		Approximate vs. Incorrect	4.50,<.001
		Correct vs. Incorrect	8.99,<.001
E) Angina	76.11, <.001	Correct vs. Approximate	-0.69, 1.0
		Approximate vs. Incorrect	7.11,<.001
		Correct vs. Incorrect	7.82,<.001

Rank-based nonparametric Kruskal-Wallis Test was used to determine if there were significant differences in speed with which help would be sought between groupings of causal attributions; two-way comparisons were subsequently used to directly compare correct attribution with approximate and incorrect attributions. A Bonferroni adjustment has been applied to the p-adj values for the pairwise comparisons.