Working Paper: CEPR ID: DP17331
Authors: Felix Chopra; Ingar Haaland; Christopher Roth; Andreas Stegmann
Abstract: In experiments with economists, we measure how the evaluation of research studies depends on whether the study yielded a null result. Studies with null results are perceived to be less publishable, of lower quality, less important, and less precisely estimated than studies with statistically significant results, even when holding constant all other study features, including the precision of estimates. The penalty for null results is of similar magnitude for various subgroups of researchers, from PhD students to editors. The null result penalty is larger whenexperts predict a non-null result and when statistical uncertainty is communicated in terms of p-values rather than standard errors. Our findings have implications for understanding mechanisms underlying publication bias and the communication of research findings.
Keywords: null results; publication bias; learning; information; scientific communication
JEL Codes: No JEL codes provided
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| null results (C29) | perceived publishability (Y90) |
| null results (C29) | perceived quality (L15) |
| null results (C29) | perceived importance (D46) |
| expert prediction of non-null result + null results (C52) | perceived publishability (Y90) |
| p-values (C12) | perceived quality (L15) |