Skip to main content
. 2024 Jul 9;8:e52503. doi: 10.2196/52503

Table 3.

Logistic regression: scenario 1 (social networking site users’ assessments of accuracy).


Odds ratio (95% CI) SE (robust) P value
Experimental group (1=treatment) 1.089 (0.747-1.588) 0.209 .65
Reliance on social media for health information (reference: not at all)

Not very much 0.964 (0.578-1.607) 0.252 .88

A little 1.627 (0.969-2.733) 0.431 .06

A great deal 1.873 (1.016-3.453) 0.125 .04
Trust in public health officials (reference: don’t trust at all)

Don’t really trust 2.794 (1.377-5.669) 1.009 .004

Trust to a degree 5.496 (2.883-10.478) 1.809 <.001

Trust a lot 14.222 (6.291-32.151) 5.919 <.001

Unsure 2.364 (0.777-7.195) 1.342 .13
Political affiliation (reference: Democrat)

Independent or Other 0.808 (0.475-1.373) 0.219 .43

Republican 0.570 (0.333-0.978) 0.157 .04

Nonvoter 0.566 (0.296-1.081) 0.187 .08
Age (log-transformed) 0.899 (0.701-1.154) 0.114 .40
Sex (reference: male)

Female 0.952 (0.640-1.416) 0.193 .80
Race (reference: White)

African American or Black 0.605 (0.352-1.042) 0.168 .07

Other 0.409 (0.210-0.796) 0.139 .008
Hispanic (1=yes) 1.176 (0.742-1.863) 0.276 .49
4-year degree (1=yes) 1.262 (0.834-1.908) 0.266 .27
Constant 0.442 (0.131-1.494) 0.266 .18
−2 log likelihood –318.920 (—a)
Pseudo R2 0.124 (—)

aNot applicable.