Skip to main content
. 2021 Jan 21;23(1):e20457. doi: 10.2196/20457

Table 2.

Regression analyses for predicting the perception of high- and low-quality mental health websites and treatment choice.

Step Evaluation of high-quality website (linear regression) Evaluation of low-quality website (linear regression) Treatment choice (binary logistic regression)


β P value β P value Exp(B) P value
Step 1: Sociodemographics






Gender (0=males) –.056 .30 .035 .497 1.866 .02

Age .002 .98 .122 .02 0.990 .29

Education level (0=low) –.036 .50 –.179 <.001 0.998 .99

R2 (%) 0.3
4.1
4.8
Step 2: Experience






Health information–seeking .004 .95 .028 .63 1.042 .77

Depression-related information–seeking .023 .69 –.052 .36 1.108 .31

Past experience of depression –.018 .75 –.128 .02 1.306 .30

R2 (%) 0.7
5.1
6.2
Step 3: Health literacy






Objective, performance-based health literacy (NVSa) –.019 .73 –.099 .07 2.068 .004

Subjective, perception-based health literacy (eHEALSb) .144 .01 –.041 .47 0.697 .04

R2 (%) 2.4
5.6
10.4
Step 4: Perception of mental health websites






High-quality site N/Ac N/A N/A N/A 1.253 .03

Low-quality site N/A N/A N/A N/A 1.088 .74

R2 (%) N/A
N/A
18.6
Step 5: Interaction





Interaction between objective health literacy and perception of a low-quality site N/A
N/A
0.886 .035
R2 (%) N/A
N/A
20.1

aNVS: Newest Vital Sign.

beHEALS: eHealth Literacy Scale.

cN/A: not applicable.