Table 3: Univariable and multivariable regression analysis results

Participant characteristic	Univariable analysis			Multivariable analysis
Participant characteristic	B	p	95% CI	B		p	95% CI
Age category, yr
18-29	Ref			Ref
30-39	0.159	0.7	-0.549 to 0.868	0.205		0.6	-0.499 to 0.909
40-49	-0.034	0.9	-0.737 to 0.669	-0.017		1.0	-0.716 to 0.681
50-59	0.969	0.005	0.289 to 1.649	1.001		0.004‡	0.323 to 1.680
60-69	1.424	< 0.001	0.631 to 2.216	1.469		< 0.001‡	0.681 to 2.257
70-79	2.312	0.001	0.971 to 3.653	2.448		< 0.001‡	1.117 to 3.779
≥ 80	3.796	0.03	0.338 to 7.255	4.457		0.01‡	1.032 to 7.881
Sex
Female	Ref			Ref
Male	-2.034	< 0.001	-2.490 to -1.579	-2.363	< 0.001‡		-2.825 to -1.900
Education
Up to high school level	Ref			Ref
Some postsecondary school	0.372	0.1	-0.138 to 0.882	0.301		0.2	-0.211 to 0.813
Undergraduate/postgraduate level	0.570	0.07	-0.054 to 1.195	0.581		0.08	-0.073 to 1.235
Income, $
< 15 000	Ref			Ref
15 000-30 000	0.739	0.2	-0.338 to 1.817	0.456		0.4	-0.613 to 1.524
30 000-50 000	0.668	0.2	-0.343 to 1.678	0.294		0.6	-0.713 to 1.300
50 000-80 000	0.839	0.1	-0.153 to 1.832	0.486		0.3	-0.508 to 1.479
80 000-150 000	0.911	0.08	-0.105 to 1.928	0.666		0.2	-0.362 to 1.693
≥ 150 000 or more	1.729	0.009	0.435 to 3.022	1.258		0.06	-0.030 to 2.547
Prefer not to answer	2.028	0.001	0.875 to 3.182	1.940		0.001‡	0.788 to 3.092
Language
English	Ref			Ref
French	0.703	0.02	0.124 to 1.281	1.307		0.0‡	0.703 to 1.912
Subjective numeracy	1.271	< 0.001	0.812 to 1.731	0.858		< 0.001	0.630 to 1.085
Risk gist knowledge*
Incorrect	Ref			Ref
Correct	0.298	0.4	-0.336 to 0.932	0.168		0.7	-0.599 to 0.935
Unsure	-0.148	0.7	-0.994 to 0.697	-0.321		0.4	-1.158 to 0.516
Risk verbatim knowledge†
Incorrect	Ref			Ref
Correct	0.384	0.233	-0.073 to 0.841	-0.305		0.3	-0.929 to 0.320
Constant				54.632		0.0	53.243 to 56.020

Note: B = beta coefficient, CI = confidence interval.

*In the gist question, a visual representation (100 people) is used to describe the risk of withdrawal for 2 treatments, treatment A (5 out of 100 people will have to stop taking the treatment) and treatment B (10 out of 100 people will have to stop taking the treatment). The respondent was asked to choose if they preferred treatment A, treatment B or were unsure.

†The verbatim question then asked how much more likely the risk of treatment A/B (irrespective if they answered this correctly) was compared with the risk in treatment B/A by selecting a number between 1 and 10 from a drop down box. A correct answer was considered to be treatment A, followed by 2 times more likely than treatment B.