Table 1 Generalized linear models of self-efficacy predicting performance in the bot-recognition task.

From: Exposure to social bots amplifies perceptual biases and regulation propensity

 

Before exposure

After exposure

Changes

(Intercept)

1.29***

1.28***

1.26***

High-ambiguity mixed botsa, c

− 0.47***

− 0.47***

− 0.47***

High-ambiguity political botsb, c

− 0.36***

− 0.35***

− 0.32***

Age

− 0.08***

− 0.09***

− 0.09***

Education

− 0.01

− 0.01

− 0.01

Independentd

− 0.02

− 0.02

− 0.02

Republicand

− 0.12*

− 0.12*

− 0.12*

Self-efficacy

 Before exposure

0.05*

  

 After exposure

 

− 0.01

 

 Change scores

  

− 0.04*

\(R^2\)

0.094

0.089

0.093

  1. The dependent variable is the proportion of accurate answers out of twenty trials. We also carried out the regression at the trial level, and the results are consistent except for the self-efficacy changes (the last column), where the association with the actual accuracy yields \(p=.051. \)
  2. \({^{*} } p<.05; {^{***} }p<.001\).
  3. We use beta distribution as the link function
  4. aThe second condition of Experiment I
  5. bExperiment II
  6. cLow-ambiguity mixed bots (i.e., the first condition of Experiment I) as the reference group
  7. dDemocrat as the reference group.
Back to article page