Table 7.
Zero-inflated negative binomial regression of absolute citation counts as shown in Figure 6.
| Predictor | Estimate | Standard error | Z-Score | P-value |
|---|---|---|---|---|
| Reference | 1.902 | 0.186 | 10.242 | <0.0001 |
| Journal of Pediatric Surgery | 0.617 | 0.179 | 3.445 | 0.0006 |
| Pediatric Surgery International | 0.208 | 0.218 | 0.956 | 0.3391 |
| Israel | 0.023 | 0.253 | 0.091 | 0.9273 |
| Netherlands | 0.434 | 0.186 | 2.328 | 0.0199 |
| 2000–2003 | 0.664 | 0.205 | 3.239 | 0.0012 |
| 2015–2018 | −0.889 | 0.197 | −4.513 | <0.0001 |
| Log (θ) | −0.226 | 0.113 | −2.002 | 0.0453 |
Analyses of Goodness-of-fit: Nagelkerke's pseudo-R2 = 0.308 (Likelihood-ratio test: Log likelihood difference: −47.55, P <0.0001). Akaike information criterion = 1637.788. Bayesian information criterion = 1691.14. Number of iterations in Broyden-Fletcher-Goldfarb-Shanno (BFGS) optimization: 63 and log likelihood = −803.9 with 15 degrees of freedom. θ represents the dispersion parameter used to calculate the variance together with the mean and therefore represents the difference to the Poisson-distribution in which mean and variance have to be equal.