Table 2.
The impacts on productivity: main results
| (1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | (9) | (10) | |
|---|---|---|---|---|---|---|---|---|---|---|
| Panel A | PubMed publications | PubMed citations | ||||||||
| Ties to China Post |
0.027 (0.003) |
0.020 (0.003) |
0.021 (0.005) |
0.024 (0.004) |
0.038 (0.005) |
0.192 (0.007) |
0.099 (0.008) |
0.101 (0.012) |
0.095 (0.009) |
0.095 (0.015) |
| Pretreatment avg. | 1.502 | 1.502 | 1.502 | 1.899 | 5.929 | 4.163 | 4.163 | 4.163 | 4.730 | 341.941 |
| R2 | 0.732 | 0.732 | 0.783 | 0.723 | 0.549 | 0.660 | 0.662 | 0.706 | 0.652 | 0.777 |
| Panel B | Non-PubMed publications | Non-PubMed citations | ||||||||
| Ties to China Post | 0.025 (0.003) |
0.015 (0.004) |
0.014 (0.008) |
0.020 (0.004) |
0.001 (0.010) |
0.079 (0.005) |
0.071 (0.006) |
0.057 (0.014) |
0.068 (0.007) |
0.023 (0.022) |
| Pretreatment avg. | 0.981 | 0.981 | 0.981 | 1.247 | 3.454 | 1.401 | 1.401 | 1.401 | 1.665 | 41.819 |
| R2 | 0.641 | 0.645 | 0.690 | 0.637 | 0.524 | 0.620 | 0.627 | 0.650 | 0.614 | 0.817 |
| Panel C | Total publications | Total citations | ||||||||
| Ties to China Post |
0.011 (0.003) |
0.008 (0.004) |
0.011 (0.006) |
0.010 (0.004) |
0.020 (0.005) |
0.180 (0.007) |
0.105 (0.008) |
0.105 (0.012) |
0.097 (0.009) |
0.082 (0.014) |
| Pretreatment avg. | 1.878 | 1.878 | 1.878 | 2.346 | 9.383 | 4.470 | 4.470 | 4.470 | 5.070 | 383.760 |
| R2 | 0.757 | 0.758 | 0.798 | 0.748 | 0.605 | 0.685 | 0.687 | 0.727 | 0.678 | 0.787 |
| No. of obs. | 792582 | 792582 | 792582 | 792582 | 792582 | 792582 | 792582 | 792582 | 792582 | 792582 |
| Scholar FE | Y | Y | Y | Y | Y | Y | Y | Y | Y | Y |
| Year FE | Y | Y | Y | Y | Y | Y | Y | Y | Y | Y |
| Baseline covariates*Year FE | Y | Y | Y | Y | Y | Y | ||||
| Entropy balancing | Y | Y | ||||||||
| Inverse hyperbolic sine | Y | Y | ||||||||
| Poisson | Y | Y | ||||||||
Note: For Columns (1)–(10), the models always control for scholar and year fixed effects. In Columns (2), (4), (5), (7), (9), and (10), we include the interactions between year dummies and four baseline covariates: 1) total number of publications in 2010–2014, 2) total citations in 2010–2014, 3) number of NIH-funded publications in 2010–2014, and 4) indicator for Asian researcher. Outcomes in columns (1)–(3) and (6)–(8) are log-transformed. Outcomes in columns (4) and (9) are transformed using inverse hyperbolic sine. Columns (3) and (6) use entropy balancing to balance all four covariates before running the regression. Columns (5) and (10) use Poisson regression. SE are clustered at the scholar level.