Skip to main content

Table 3 Multilevel regression on learning persistence and performance in experiment 2

From: Compared to what? Effects of social and temporal comparison standards of feedback in an e-learning context

 

Persistence

Performance

Model 1 B(SE)

Model 2 B(SE)

Model 3 B(SE)

Model 1 B(SE)

Model 2 B(SE)

Model 3 B(SE)

Intercept

− 1.236*** (0.229)

− 1.239*** (0.227)

− 1.239*** (0.229)

− 0.227 (0.143)

− 0.228 (0.144)

− 0.233 (0.144)

Feedback level (L1)

 Time to exam

− 0.110** (0.045)

− 0.109** (0.045)

− 0.110** (0.045)

0.071* (0.043)

0.070 (0.043)

0.069 (0.043)

 Correct exercisesa

− 1.241*** (0.454)

− 1.232*** (0.454)

− 1.262*** (0.456)

1.885*** (0.185)

1.887*** (0.186)

1.894*** (0.186)

 Session length

0.065*** (0.011)

0.065*** (0.011)

0.066*** (0.011)

-0.051 (0.034)

− 0.050 (0.034)

− 0.049 (0.034)

 Feedback

0.015 (0.037)

0.012 (0.037)

0.008 (0.037)

0.031 (0.036)

0.032 (0.036)

0.032 (0.036)

 Fluency

0.037 (0.037)

0.037 (0.037)

0.036 (0.037)

-0.037 (0.036)

-0.037 (0.036)

-0.036 (0.036)

Learner level (L2)

 Temporal RNO

0.003 (0.075)

0.006 (0.075)

− 0.001 (0.076)

− 0.039 (0.057)

− 0.041 (0.057)

− 0.044 (0.057)

 Social RNO

− 0.089 (0.081)

− 0.093 (0.081)

− 0.090 (0.082)

0.030 (0.061)

0.036 (0.061)

0.036 (0.061)

Cross-level-interactions

 Temporal RNO × Feedback

 

0.042 (0.038)

0.039 (0.039)

 

− 0.071* (0.038)

− 0.074* (0.038)

 Social RNO × Feedback

 

− 0.074* (0.040)

− 0.070* (0.040)

 

0.082** (0.037)

0.080** (0.038)

 Fluency × Feedback

  

− 0.013 (0.037)

  

0.006 (0.036)

 Temporal RNO × Fluency

  

0.057 (0.039)

  

− 0.002 (0.038)

 Social RNO × Fluency

  

− 0.025 (0.040)

  

− 0.004 (0.037)

 Temporal RNO × Feedback × Fluency

  

0.038 (0.038)

  

0.045 (0.038)

 Social RNO × Feedback × Fluency

  

− 0.111*** (0.040)

  

0.028 (0.037)

  1. N = 4690. RNO = reference norm orientation (preferences for comparison standards). Persistence measured as session abort after feedback (L1, binary, 0 = no abort, 1 = abort), performance as performance on items presented with feedback (L1, binary, 0 = wrong, 1 = correct). Effect coding of feedback (− 1 temporal, 1: social) and fluency (− 1: fluent, 1: disfluent). aProportion of correct answers p for performance models and proportion of correct answers times proportion of wrong answers p × (1 − p) for persistence models
  2. *p < 0.1, **p < 0.05, ***p < 0.01