Table 2 Fixed-effects panel regression results. Dependent variable is the log number of citations per paper, and the “specialization” variable is standardized. Standard errors are in parentheses. All control variables described in the main manuscript are included. Based on 22,589 unique biomedical researchers with between 75 and 99 career publications career publications, for a total of 145,143 researcher-time window observations
From: Scientific rewards for biomedical specialization are large and persistent
(1) | (2) | (3) | (4) | |
|---|---|---|---|---|
Specialization | 0.265*** | 0.359*** | 0.410*** | 0.727*** |
(0.004) | (0.006) | (0.008) | (0.014) | |
Specialization × career age | − 0.007*** | − 0.025*** | ||
(0.0003) | (0.001) | |||
Specialization × papers | − 0.055*** | − 0.098*** | ||
(0.002) | (0.005) | |||
Career age × papers | 0.021*** | |||
(0.001) | ||||
Specialization × career age × papers | 0.004*** | |||
(0.0003) | ||||
Observations | 145,143 | 145,143 | 145,143 | 145,143 |
R 2 | 0.174 | 0.178 | 0.178 | 0.186 |
Adjusted R2 | 0.022 | 0.026 | 0.026 | 0.036 |
Control variables | Yes | Yes | Yes | Yes |
Year fixed-effects | Yes | Yes | Yes | Yes |
Researcher fixed-effects | Yes | Yes | Yes | Yes |