Table 5. Average scores of how effective respondents think each suggestion will be (Likert scale where 1 = not at all effective and 5 = extremely effective) and number of respondents who selected each suggestion as “most effective” at reducing publication bias.
| Editors (n = 73) | Academics (n = 160) | Academics < 10 y experience (n = 98) | Academics > 10 y experience (n = 62) | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
|
Suggestion |
Mean ± SD scorea | Chosen as most effective (%) | Mean ± SD scorea | Chosen as most effective (%) | pdiff mean scoresb | Mean ± SD scorea | Chosen as most effective (%) | Mean ± SD scorea | Chosen as most effective (%) | pdiff mean scoresb |
| Research registration | 3.3 ± 1.5 | 21 | 2.9 ± 1.2 | 6 | 0.064 | 2.9 ± 1.2 | 6 | 2.9 ± 1.3 | 7 | 0.831 |
| Mandatory publication | 3.1 ± 1.6 | 25 | 3.0 ± 1.4 | 14 | 0.564 | 3.0 ± 1.4 | 11 | 2.9 ± 1.4 | 18 | 0.420 |
| Negative results journals/articles | 3.1 ± 1.3 | 16 | 3.6 ± 1.3 | 21 | 0.002 | 3.9 ± 1.2 | 24 | 3.2 ± 1.4 | 18 | 0.003 |
| Pre-study publication of methodology | 3.0 ± 1.4 | 8 | 3.1 ± 1.3 | 6 | 0.606 | 3.2 ± 1.3 | 7 | 2.9 ± 1.3 | 5 | 0.197 |
| Two-stage review | 2.7 ± 1.4 | 11 | 3.4 ± 1.3 | 26 | 0.001 | 3.5 ± 1.3 | 28 | 3.1 ± 1.4 | 23 | 0.070 |
| Peer-review training and accreditation | 2.6 ± 1.1 | 11 | 3.3 ± 1.3 | 10 | < 0.001 | 3.4 ± 1.3 | 11 | 3.0 ± 1.2 | 8 | 0.048 |
| Post-publication review | 2.5 ± 1.2 | 6 | 2.7 ± 1.2 | 3 | 0.201 | 2.8 ± 1.2 | 1 | 2.7 ± 1.2 | 5 | 0.521 |
| Published rejection lists | 2.2 ± 1.1 | 1 | 3.0 ± 1.2 | 4 | < 0.001 | 3.0 ± 1.2 | 0 | 2.9 ± 1.3 | 10 | 0.630 |
| Open reviewing | 1.9 ± 1.0 | 1 | 2.9 ± 1.2 | 11 | < 0.001 | 3.0 ± 1.2 | 12 | 2.9 ± 1.3 | 8 | 0.844 |
Abbreviations: SD, standard deviation
aMean scores provided by the whole sample, not just those who selected the suggestion as the most effective
bIndependent samples t-test