Table 1.
Results of eight sets of regression models, corresponding to each of the observed attitudes and reactions from Weibo posts, with each set containing a linear model and an exponential model.
| Dependent variable | Independent variable | Posts, n | Linear | Exponentiala | |||||
|
|
|
|
β | P value | R2 | b | P value | R2 | |
| Uncertain (%) | Date | 21 | –0.0145 | .12 | 0.12 | –0.2260 | .004 b | 0.32 | |
| Scared (%) | Date | 21 | 0.0104 | .11 | 0.12 | 0.0286 | .13 | 0.59 | |
| Cautious (%) | Date | 21 | 0.2819 | <.001 | 0.50 | 0.0596 | <.001 | 0.54 | |
| Mask (%) | Date | 21 | 0.0168 | .07 | 0.16 | 0.0372 | .04 | 0.18 | |
| Cancel (%) | Date | 21 | –0.0001 | .98 | <0.01 | –0.0034 | .95 | <0.01 | |
| Self-treatment (%) | Date | 21 | –0.0003 | .20 | 0.08 | –0.1135 | .42 | <0.01 | |
| Normal (%) | Date | 21 | –1.2999 | .99 | <0.01 | –0.0006 | >.99 | <0.01 | |
| Evacuate (%) | Date | 21 | –0.0006 | .44 | 0.03 | –0.1647 | .09 | 0.12 | |
aThe exponential coefficient and P values were provided for the b coefficient in the Y=a+bx equation, and exponential R2 values were computed as Cox and Snell R2.
bItalics represent statistically significant P values.