One experiment too far
Today I got the results of an experiment I really wish I hadn’t done. Before today, the theory all made sense. Today, this new information has muddied the waters.
Of course I won’t pretend it never happened, sweep it under the carpet, go on as if nothing had changed. Even though it’s a rather minor point to the story, I will still have to regroup a little bit.
Makes you wonder, sometimes, about all the published scientific papers out there. How many of these might be cast into doubt by the next experiment, had it only been performed? When we draw a line under a story and pronounce it “finished”, can we sometimes be cutting off an eventual truth? Yet if we never stopped, how could we ever publish anything? How could we spin our narrative, focus our questions, find our answers?
The more years I am in research, the more provisional it all seems. If I go on like this, will I end up believing in nothing?
Hi Jenny
I think the simple truth is there will always be more you could do. If we all had armies of minions and huge grants then we might be justified in feeling bad about not doing every minor experiment that might or might not strengthen our project. However, for most of us time and money are real limitations, as is the need to get published or graduate sooner rather than later. Unless it is simple to do, and/or a crucial piece of data then a particular experiment may not be a justified investment of your time and effort.
I know that “awkward” data does sometimes get brushed under the carpet, and this is unethical. It is also not ethical to avoid an experiment because it might mess up your nice little story. However, because you cannot do every conceivable experiment, you can pick and choose to do the ones that make the most sense at the time.
A fact of life for scientists is they are judged on their publication record, not how good or throughout their unpublished research is. I also think that if it is important it is better to get out what you already know rather than wait and wait until you know even more. Even without all the bad data and poorly done experiments, science is provisional and never ending.
Besides, there are also only so many figures you can include in any one paper and it makes sense to divide a research project up into publishable units. Ideally a second or third paper (by you or someone else) will build upon the first, but if further data contradicts or confuses your initial findings it will only reflect badly on you if you (or others) cannot repeat what you did in the first place!
You could not tell anyone, and then agonise over it for months. The affair with a senior member of staff is optional, I guess.
Dr Girlfriend – all of what you say is true. But in my advancing years, I still feel a visceral sense that my own sense of certainty is steadily eroding away. I suspect this is a good thing, and might go some way to explaining why the scientific apprentice structure works so well: young hungry bucks, eager to bolt forward, but tempered by cynical older codgers like me who don’t quite believe a word of it. The compromise becomes the paper.
Bob, I get the feeling you’re reviewing The Honest Look one tweet and comment at a time. It’s a great narrative structure for a book review – keep it up!
Dear Jenny
I think it goes more deeply and has to do with what science and research are on a fundamental level, rather than just the publish-or-perish and grant money pressure. As I see it, and many others have said, research is a constant search for the truth and we generally get to it by designing experiments that will prove us wrong. So there is always something more we can do to check an hypothesis/prove a theory/provide an explanation. And that’s the beauty of science and research: the constant quest for knowledge.
And of course it is frustrating when that extra experiment you do muddles things up. But it can happen either before or after publishing the paper – that’s how knowledge evolves.
If “believing in nothing” means not taking anything for granted and questioning all known “wisdom”, then, I think that is exactly what a scientist should aim for. All we have are models, theories and possible explanations – all there to be disproven or strengthened by the next experiment we perform.
Good luck!
In scientia, nescimus.
Oh yes…