Table 2.
Hierarchical ordinary least squares regression of predictors for the number of retweets (based on Texas public health agencies’ COVID-19–related tweets that were posted between January 1 and June 30, 2020; N=7269).
| Variables | Number of retweets | ||||||||
|
|
Model 1a | Model 2b | Model 3c | ||||||
|
|
βd (SE) | P value | β (SE) | P value | β (SE) | P value | |||
| Control | |||||||||
|
|
Followers | .43 (.01) | <.001 | .43 (.01) | <.001 | .50 (.01) | <.001 | ||
| Functions | |||||||||
|
|
Information | .10 (.03) | <.001 | .09 (.02) | <.001 | .04 (.03) | <.001 | ||
|
|
Action | .10 (.01) | <.001 | .09 (.01) | <.001 | .12 (.02) | <.001 | ||
|
|
Community | .01 (.02) | .35 | .02 (.01) | .07 | .04 (.02) | .001 | ||
| Types of actions proposed | |||||||||
|
|
Handwashing | N/Ae | N/A | −.01 (.03) | .36 | −.05 (.03) | <.001 | ||
|
|
Social distancing | N/A | N/A | .02 (.03) | .13 | .02 (.03) | .16 | ||
|
|
Mask wearing or face covering | N/A | N/A | .01 (.03) | .64 | .03 (.03) | .01 | ||
|
|
Staying at home | N/A | N/A | .05 (.02) | <.001 | .04 (.02) | .001 | ||
|
|
Getting tested | N/A | N/A | .03 (.02) | .004 | .11 (.02) | <.001 | ||
|
|
Learning more information | N/A | N/A | .05 (.01) | <.001 | .06 (.01) | <.001 | ||
|
|
Other behaviors | N/A | N/A | −.04 (.02) | .01 | .03 (.02) | .08 | ||
| Health Belief Model variables | |||||||||
|
|
Severity | N/A | N/A | N/A | N/A | .10 (.01) | <.001 | ||
|
|
Susceptibility | N/A | N/A | N/A | N/A | .20 (.01) | <.001 | ||
|
|
Benefits | N/A | N/A | N/A | N/A | −.002 (.01) | .91 | ||
| Model of f values | 494.52 | <.001 | 186.74 | <.001 | 207.11 | <.001 | |||
aModel 1 had a change in R2 of 0.21 and a total R2 of 0.21.
bModel 2 had a change in R2 of 0.01 and a total R2 of 0.22.
cModel 3 had a change in R2 of 0.07 and a total R2 of 0.21.
dβ is a standardized coefficient.
eN/A: not applicable.