The literature search

We describe a necessary but sometimes tedious step in scientific research.

Our astronomer has finally gotten his paper accepted, the one he’d been working on for a long time and which we’ve mentioned now and then.  It’s not going to win him a Nobel, or even much notice among astronomers in general, though some of them will be interested.  But it’s good to finally get it finished.  What next?

There is a sort of momentum, a tendency to keep working on astronomy research after the final effort on the paper, rather than (say) changing course and helping our photographer organize his latest work and plan his next project.  And indeed growing directly out of this paper our astronomer came up with two questions, and after scribbling some equations, an idea.  But rather than immediately sitting down at the computer and crunching some data, he started the necessary first step: the literature search.  Has it already been done?  Has anyone asked the questions before (one of them, almost certainly) and, most important, answered them?  It would be a waste of time, indeed rather pointless, to recreate a mathematical procedure that already exists.

Well, some literature searches are easy.  If we were dealing with, say, pulsation periods of RR Lyrae stars the keywords are tabulated and the latest results would be in our hands in moments.  But a mathematical procedure to deal with data could be called almost anything, and appear in almost any journal.  What our astronomer is thinking of could have been published fifty years ago in a book or paper he’s never heard of, one that has no obvious connection to astronomy.  And what he calls the “expected Fourier amplitude due to noise” could be well-known as “Wellington coefficients” or some such name, making any kind of keyword search difficult or impossible.

Well, he started with textbooks.  He found some similar topics, expressed in unfamiliar ways.  He will have to spend more time going through them carefully, to work out how far they apply to his idea, and indeed if they have any relevance at all.  He is a little resentful at the prospect of learning a great deal that is not really of interest; but we suppose it’s no different from the feeling of many of our tutor’s students toward, say, Chemistry or Calculus.

Then there was the scan of recent abstracts from astronomy-related journals.  This is sort of the other end of the process: the textbooks are a beginning, recent papers are the end.  A paper on time-series analysis may not directly apply to the question at hand, but may refer to something that does.  It may be the end of a string that can be followed back to a useful point.  But a recent paper may include a great deal of unfamiliar material, some of it not relevant to the question at hand.

There’s a lot of careful study to do yet.  But that’s all part of the game.