Table 3. Several tests for assessing patterns within the collaboration network and its main component.
Collaboration network | Collaboration network MainComponent | |||
Geary | Moran | Geary | Moran | |
Academic title | 0.7* | 0.3** | 0.8 | 0.1 |
Gender | 0.8* | 0.2* | 0.7* | 0.2 |
Papers | 0.2*** | 0.7*** | 0.7 | 0.2* |
G-index | 0.5*** | 0.5*** | 0.5** | 0.2 |
H-index | 0.5*** | 0.7*** | 0.5* | 0.2 |
Citations | 0.5*** | 0.7*** | 0.5** | 0.2 |
Tenure | 0.3*** | 0.8*** | 0.4** | 0.3* |
Note. Moran’s I statistic of autocorrelation ranges from −1.0 (perfect negative autocorrelation) through 0 (no autocorrelation) to +1.0 (perfect positive autocorrelation). Geary’s C statistic of autocorrelation ranges from 0.0 (perfect positive autocorrelation) through +1.0 (no autocorrelation) to +2.0 (perfect negative autocorrelation). According to Hanneman and Riddle [42], Geary’s C statistic is more sensitive to “local” differences, while Moran’s I statistic is more sensitive to how similar or dissimilar is each pair to the overall average (i.e. “global” difference).