Heard from the principal again today, complaining about my lack of published results. While she is technically correct, in fact she misses the point. I have published a lot of results, but they have all been negative.
One of the realities of science is that all results are important, whether they support a theory, deny the theory, or are neutral to the theory. Interestingly, only results that are definitively positive or negative tend to get actively pushed, and those which are positive tend to get a strong reception. It's a simple matter of human nature. We're positive-bias machines -- we like things that confirm what we already believe. Everyone knows that; it's freshman psychology to find a result that confirms that we like to confirm the results we believe should be confirmed.
The real business of science works better when we know everything that was tried and failed. We should know that of all plausible attempts, nine times out of ten experiments to confirm a hypothesis failed to produce a statistically significant result. That would be knowledge -- indeed it would be evidence against the hypothesis. And yet only one time in four (as we know from empirical scientific research) do negative results get into the shared record. All those so-called "file drawer" results, especially if they showed that some popular idea failed, are relegated to the dustbin of science.
For all that the so-called "hard science" empiricists belittle the research, meta-knowledge about how we do science is incredibly important. The extent that results that are negative or neutral get unreported is a major issue and one that needs to be addressed. Especially in situations like ours where the pressure to find significant new results is almost unbearable.
In my case it's particularly difficult. I'm saddled with a research program that is based on the false premise that the Tanzen have a hidden means of communication. All the evidence so far indicates that they do not. Where it's definitive I can publish something that says "no", which only raises more questions -- about the methodology, about the data, about me. Where it's not definitive, which is most of the time, I can only report a negative. It's about time to take the meta-program and deduce that the concept was flawed.
I wish I could redefine my own program, but I cannot. I'm a proxy for thousands of other researchers, and I have to answer their questions as well as my own. However hard it is to be definitive, and however much I might have to argue with Maija, the crypto-linguistic hypothesis is wrong. Ultimately that's the story here, not how many popular results I publish or unpopular ones I fail to publish.
Comments