Per wonderful Wikipedia, which is not everything and not always right or balanced, but anyway – thank you Wikipedia! – here’s a definition of ‘null result’: “In science, a null result is a result without the expected content: that is, the proposed result is absent. It is an experimental outcome which does not show an otherwise expected effect. This does not imply a result of zero or nothing, simply a result that does not support the hypothesis.”
But null results aren’t published enough, so we often get a skewed idea of the range of scientific findings regarding the subject of interest. Almost 80% of null results – at least in the social sciences – are unwritten and/or unpublished. This per Publication bias in the social sciences: Unlocking the file drawer.
What to do? It’s easy to say “publish the null results, you numskulls!” Or, “do more replication research”. Try having a career that way. Try attracting money. The forces are against it.
Funding organizations (or some far-seeing billionaires) need to expand their missions to include research dedicated to sniffing out false positives. And academic journals need to be more welcoming of papers based on such research. Of course, the research would have to be on the up-and-up, where design is sound and commitment to the scientific method trumps hoped-for findings.
Ideally, journals would have sections comprised of critiques of previously published papers summarizing research findings. But this is unlikely to happen, since it’s not exactly in a journal’s self-interest to expose weaknesses in what it had previously agreed to publish. Perhaps there should be journals dedicated to critiquing papers that were published elsewhere. To avoid a lot of sloppy, mean-spirited writing, standards for critiques would have to be high and the original researchers would be invited to respond.