What do you do when your plotted data looks more like a sneeze-splatter than a straight line? Publish it anyway! The current dearth of Stuff That Doesn’t Work in the scientific literature dooms researchers to repeatedly try and fail over and over again. It doesn’t have to be this way. Publishing null results is possible. It just takes a bit of creativity.
The scientific literature is full of papers with the same format: “Based on the current understanding of this topic, we formed a hypothesis, tested it and it worked”.
Anyone deviating from this formula would be shot down by the peer review process. This leaves many a researcher abandoning reams of data just because the magical statistics program didn’t find the expected correlation.
There is nothing wrong with burying a sentence or two in another, related, paper that says “by the way, tried this and it didn’t work”. It has bought me great joy when I’ve happened upon these little gems of knowledge so I am fully supportive of this strategy.
And yet it can be gut-wrenching that, after months of data collection and interpretation, your whole project is condensed into a single sentence just because A did not lead to B after all.
Things not behaving as expected yield amazingly useful data and as such can and should be published as a stand-alone paper. What these sorts of results need is re-wording the story that the data actually tells.
Firstly revisit the knowledge gained from your experiments. Why have the data not produced the expected result? Maybe current knowledge is based only on model systems and your results are from complex real-world samples. Or perhaps you used a new and improved technique that shows something in more detail than has been previously possible.
Then, with that story in mind, it is easy to alter the wording of the project title to give it a more positive spin. So “Does A cause B?” where the response is “Actually no. It doesn’t.” becomes “Investigation of real-world samples using state-of-the-art technology.”
Writing these sorts of papers can be much more difficult than writing the standard “Oh look, it did exactly what we expected” paper but it can still be rewarding. At least then your data can contribute to the current state of knowledge on a topic and that’s really what research is all about.
And if, after months of work, your results aren’t showing what you expected, then undoubtedly you are operating at the cutting edge of science where the literature is insufficient to properly explain the phenomena. At least that’s what I tell myself.