I have been poring over some manuscripts recently and realized that there are some words which are best avoided in any scientific paper. Hopefully I would not use them myself and I would find myself grimacing if someone else used them.
Probably the most verboten word is "prove". There is no proof in science, only in mathematics. Especially in science where almost everything we do consists in building a model; whether it is a protein-ligand interaction model, stereoselective organic reaction model, or transition state model. A model can never be "proven" to be "true". It can only be shown to correlate to experimental results. Thus anyone who says that such and such a piece of data "proves" his model should get the referees' noose right away.
So what would be a better word? "Validate"? Even that sounds too strong a word to me. So does "justify". How about "support"? Perhaps. I think about the best thing that all of us can say is that our model is consistent with the experimental data. This statement makes it clear that we aren't even proposing it as the sole model, only as a model that agrees with the data.
Even here the comparison is tricky since all pieces of data are not created equal. For instance one might have a model of a drug bound to a protein that's consistent with a lot of SAR data but somehow does not seem to agree with one key data point. The question to ask here is what the degree of disagreement is and what the quality of that data point is. If the disagreement is strong, this should be made clear in the presentation of the model. Often it is messy to tally the validity of a model with a plethora of diverse data points of differing quality. But quality of data and underreporting of errors in it is something we will leave for some other time.
For now we can try to keep the proofs out of the manuscripts.
Field of Science
-
-
-
Political pollsters are pretending they know what's happening. They don't.4 weeks ago in Genomics, Medicine, and Pseudoscience
-
-
Course Corrections5 months ago in Angry by Choice
5 comments:
Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="http://www.fieldofscience.com/">FoS</a> = FoS
Subscribe to:
Post Comments (Atom)
Yes, prove is almost always a bad word. But, on the other hand, you don't want to go so far in the other direction that your paper has so much waffle that you want to pour syrup on it.
ReplyDeleteOT, but I *think* that quote originates with Scott Fahlman not Hal Abelson.
ReplyDeleteI could be wrong; it was pretty popular way back when.
What about "show"/"shows"? It may seem a little colloquial but it drops all pretense. It's a very effective way of saying, "This appears to be self-evident to anyone versed in the field." This, without saying anything in the least that would make it appear as if you are not at all certain about your results. It also has the added benefit of keeping a certain level of stylistic clarity. If you promise to "show" you'll tend to be a lot more inclined to draw clear lines from one piece of reasoning to the next. I'm not certain because I've been taking it for granted, but I believe that this word is often used- or perhaps that's just a quirk of my memory now that I've mentioned it.
ReplyDeleteExcellent points.
ReplyDeleteI have recently started using "...results are most readily explained by the XXX model", particularly when I'm comparing one particular model with others in the literature. This way I'm making it clear that all I can claim is that of available models/explanations, this one is the most consistent with the data. Future data might go against it, and future models might surpass it.
@TheChemist: "Show", while colloquial, does indeed seem to be an appropriate word. Unfortunately my past advisor often eschewed useful colloquialism like this, a practice with which I sometimes disagreed.
ReplyDelete@swheele2: That seems like the right way to say it