Results That Do Not Make Headlines

Scientific discoveries often attract attention when they reveal a striking effect or suggest an unexpected explanation. Research that fails to support a prediction is less likely to make headlines. Nevertheless, the absence of an expected effect can itself be useful information. Imagine several laboratories independently testing the same promising idea. If only the laboratory that observes a favorable result publishes its work, readers may ( 1 ). They will see evidence from the apparent success without knowing that other attempts produced a different picture.

Sharing negative results can help correct this imbalance, but it does not follow that every unsuccessful experiment is equally informative. An experiment may be too small to detect the effect being investigated, or its measuring equipment may be unreliable. A report of no effect is much more useful when the methods are described clearly and relevant limitations are acknowledged. ( 2 ), the value of a negative finding depends partly on the quality of the process that produced it, just as the value of a positive finding does.

One approach is for researchers to describe their questions and methods before collecting data. This makes it easier to distinguish plans made in advance from decisions made after seeing the results. Researchers still need flexibility: unexpected observations may suggest worthwhile new questions. The important point is to identify such later analysis honestly, rather than presenting it as originally planned. The aim of greater openness is therefore not to ( 3 ). It is to let other researchers evaluate how a conclusion was reached and decide what should be tested next. Even an unexciting outcome can then contribute to scientific progress.

Results That Do Not Make Headlines

Scientific discoveries often attract attention when they reveal a striking effect or suggest an unexpected explanation. Research that fails to support a prediction is less likely to make headlines. Nevertheless, the absence of an expected effect can itself be useful information. Imagine several laboratories independently testing the same promising idea. If only the laboratory that observes a favorable result publishes its work, readers may ( 1 ). They will see evidence from the apparent success without knowing that other attempts produced a different picture.

Sharing negative results can help correct this imbalance, but it does not follow that every unsuccessful experiment is equally informative. An experiment may be too small to detect the effect being investigated, or its measuring equipment may be unreliable. A report of no effect is much more useful when the methods are described clearly and relevant limitations are acknowledged. ( 2 ), the value of a negative finding depends partly on the quality of the process that produced it, just as the value of a positive finding does.

One approach is for researchers to describe their questions and methods before collecting data. This makes it easier to distinguish plans made in advance from decisions made after seeing the results. Researchers still need flexibility: unexpected observations may suggest worthwhile new questions. The important point is to identify such later analysis honestly, rather than presenting it as originally planned. The aim of greater openness is therefore not to ( 3 ). It is to let other researchers evaluate how a conclusion was reached and decide what should be tested next. Even an unexciting outcome can then contribute to scientific progress.