Table 3.
Summary of hierarchical negative binomial regression model predicting count of shared misinformation tweets, negative binomial regression model predicting count of shared control tweets, and linear regression model predicting discernment with pooled data of studies 2 and 3.
|
|
Misinformation | Control | Discernment | |||||
|
|
Estimate (SE) | P value | Estimate (SE) | P value | Estimate (SE) | P value | ||
| Intercept | −1.70 (0.26) | <.001 a | −5.74 (0.62) | <.001 | 0.55 (0.11) | <.001 | ||
| Trial | 0.00 (0.08) | >.99 | 0.68 (0.17) | <.001 | 0.00 (0.04) | .96 | ||
| Reference: personal network | −0.76 (0.25) | .003 | −0.95 (0.52) | .06 | −0.24 (0.11) | .02 | ||
| Reference: combined | −1.10 (0.26) | <.001 | −0.11 (0.41) | .79 | −0.31 (0.11) | .005 | ||
| Reference: all users | −0.29 (0.22) | .19 | 0.02 (0.36) | .95 | 0.03 (0.10) | .78 | ||
aPredictors are italicized if significant at P<.05.