Bremsstrahlung

Recently at the lab where I work we analyzed the last year or so worth of data, and found a result which is interesting, but unexplained and not at a high enough significance level to be a big discovery. All of the senior scientists are debating what the collaboration should do: Should we hide the result for now until we know whether our new data will support it, should we publish it, and if so how should we describe it, or should we do something else. I thought that Dr H. phrased the problem interestingly in today's group meeting. To paraphrase:

Is it a[n experimental] scientist's duty to produce data or finished scientific results?

The former would entail being bound to say "We did an experiment and this is what we saw," while the latter would require a statement of the form "We did an experiment, here is what we saw, and this is why we think we saw it."

I feel that I lean toward the former view, in that I see no reason that an experimentalist must be required to explain everything he sees, and indeed that someone else may be better prepared to have the insight to understand a new phenomenon. After all, if I discovered an anti-gravity field naturally occurring in my apartment, I think that the scientific community could accomplish far more if I published detailed records of my observations than if I sat on those data while I tried to figure out the mechanism of anti-gravity. Also historically many major physics problem seem to me to have been solved by the work of many people doing their own research based on each other's findings. I think a good example is the discovery of cosmic rays. Had Pacini chosen to keep his data to himself because he didn't know the source of the additional radiation rather than sharing them with the other physicists of the day the progress of modern astrophysics could have been held back significantly. As it was many leading scientists debated the topic, making numerous wrong guesses, until they homed in on the true explanation. I think it unlikely that my group's measurement will have much scientific impact, but there is a great value in early and thorough sharing of results.

On the other hand, it does seem to me that the first viewpoint must not be taken to an extreme where the experimentalist makes measurements, shrugs, writes them in a paper, and wanders off to do something else. One can hardly call oneself a scientists if one doesn't make a serious effort to wring all of the insight possible out of ones findings. Even if you can't fully explain what you see, you want to rule out unlikely explanations, establish limits, and test any applicable theories as much as possible. That isto say, the task of interpretation should be equally divided between theorists who generate theories and experimentalists who generate data to test theories.