IN science and medical publishing, everything is positive. Less than 4% of articles deal with negative results. There is a perception that negative results are non-results; only positive results are worth publishing. Why is it that showing that something does something is so much more important that showing that something doesn’t do something?
Obviously, I expect some common sense in this; I don’t very well expect that a paper should be published just because you have demonstrated that drinking water doesn’t cause sunburn, this would be a deeply unsurprising discovery. But what if it is a study that demonstrates that a particular drug doesn’t do what people expected it to do? What if it is a biotechnology that doesn’t work for a whole swathe of biological research?
Online science forums (or fora) are replete with anecdotal evidence describing how time, and time, and time again research scientists make the same mistakes, or encounter the same limitations, in particular techniques. This is because no-one ever publishes such limitations, or at least, not more than 4% of the time.
So what is the problem? Well, science is expensive. Very expensive. It is expensive in material cost, and it is expensive in research hours. To have discovered that you’ve wasted a year doing work that elsewhere in the world someone once wasted a similar amount of time doing, only, 3 years ago, is deeply frustrating.
In coffee breaks around the world, many scientists have discussed the idea of a Journal of Negative Results, a compendium that can be consulted at the outset of a research project to determine whether a technique or approach has already been taken toward a research problem, but has been found not to work. Sometimes such negative results a mentioned, but only in passing, and only after an alternative technique resulted in positive results, which resulted in the subsequent publication. They are rarely keyword searchable and thus inordinately difficult to find.
As I mentioned, science costs a lot of money, far more money than is necessary. This is largely because the money isn’t real, there is poor ownership of it, it is monopoly money. If it were coming out of our own pockets, we simply wouldn’t pay the price we do, we’d demand more competitive prices. Consumable companies are free to charge extortionate prices for items that they are producing by the million. I have tubes in my lab that cost £3.75 each; they can only be used once, and invariably one or two of them can be wasted due to one problem or another. Kits are all the rage in research; pre-fabricated methodologies with all the reagents and instructions one needs to perform a particular experiment. The reagents themselves cost practically nothing in most cases, yet the kits can cost anywhere between £300 – £1500, and in many circumstances, afford you between 5 – 20 experiments.
Now this combination of expensive research is part of what leads to negative results being unwanted. There’s no real money in debunking an idea, it must come along side a positive result if it is to come at all. In the pharmaceutical industry, it is part of the reason why any new drug being produced is just too much of an investment to allow to fail, so the pressure is on to ensure, by hook or by crook, that the drug is licensed. Ben Goldacre writes at length about this in his recent book, and blog of the same name, Bad Science; this is most definitely worth a read!
Expensive research also prevents investment into rarer diseases, or any medications that run the risk of having a short shelf-life. One class of drugs that have fallen foul of this economic equation are antibiotics, and this is a rather long pre-amble into what I wanted to say in this blog essay (or blessay, and Stephen Fry attests to horribly calling it).