Embedded Files
Here's something most people outside of science don't know: a huge proportion of research that gets conducted never makes it into a published paper. Plenty of studies sit in file drawers, exist only as conference posters, or simply get abandoned halfway through. This phenomenon, sometimes called the "file drawer problem," is one of the most under-discussed issues in science.
Why research goes unpublished
The most common reason research doesn't get published is because of negative or null results: studies that show no significant effect, no meaningful difference between groups, or outcomes that don't confirm the hypothesis. Journals have historically been much more likely to accept and publish studies with dramatic, positive, statistically significant findings. A study that says "we found nothing" is a hard sell, even if "nothing" is the right answer.
There are other reasons too: the study may have been abandoned due to funding running out, the researcher may have moved to a different institution or left the field, the data may have been considered too preliminary to write up, or the results may have been published by a competing group first.
Why this is a serious problem
When only positive results make it into the published literature, the picture of scientific evidence becomes distorted. For instance, if ten research groups independently test the same drug, eight of which find no effect and two find a positive effect, those two are most likely to be the only ones getting published.
If only those two publish, the literature suggests the drug works even though the vast majority of the evidence indicates otherwise.
This distortion has real-world consequences. It can lead to treatments being adopted based on incomplete evidence, to public health guidelines being built on a skewed evidence base, and to scientists wasting resources investigating phenomena that might already have been ruled out.
What's being done about it
The scientific community has increasingly recognized this problem and world towards addressing it. Preregistration requires researchers to publicly register their hypotheses and methods before conducting a study, making it harder to selectively report only favorable results later. Journals dedicated to publishing null results, like PLOS ONE, which evaluates papers based on methodological rigor rather than the significance of the findings have grown in influence.
Preprint servers like bioRxiv and arXiv allow researchers to share their work publicly before formal peer review, which means that even unpublished or preliminary findings can enter the scientific conversation without needing a journal's stamp of approval first.
Page updated
Google Sites
Report abuse