Displaying the data

Communicating science can be difficult, even to and by scientists.

We’ve recently commented on one aspect of the digital/analog divide, to the effect that a digital display is better if you’re interested in an exact number, and an analog one for getting across how accurate the number is.  We phrased it in terms of the advent of digital computers in the second half of the twentieth century, but now we realize it went back much further than that.

In the old days, much more than today, a scientist or an engineer might be sent to a graph to read off a number for processing.  Opening a book almost at random, we find “Drag coefficient vs. Reynolds’ number for two-dimensional flow round a circular cylinder.”  One would calculate the Reynolds’ number, read off the drag coefficient, and use it in a calculation.  (Nowadays the data are hidden somewhere in software, though we suspect in many cases that they are no more accurate than this printed graph.)  If one needed accuracy beyond the three figures that is the general limit of a graph, one went to tables.  The dinosaurs among us recall tables of logarithms and trigonometric functions; four figures meant a pamphlet, five a book, and six a serious question as to why one would desire such a thing.  In any case, we realize, tables are a digital display.  They’re more difficult to use, but necessary sometimes.

Graphs are much better for showing what’s happening overall (as one commentator of a previous post pointed out).  You might not even notice the numbers, but concentrate on the fact that (say) the luminosity of high-mass stars runs in a different way from that of low-mass stars.

But communicating science involves more than tables and graphs.  There is also the scientific paper, a written accomplishment that may contain both tables and graphs, and has a framework of printed words.  Its purpose is to get across a possibly complicated and dense mass of information and explanation.  For this it’s important that the reader be able to stop, reread a section, look carefully at a plot, possibly look up from the document and think something through carefully before going on.  The reader’s control is vital to its success.

Meanwhile, the author of the paper is also presenting a summary of the work at a conference.  It’s easy enough to show plots and tables as necessary; but the framework is verbal, not written.  The listeners cannot stop and refer back to a previous paragraph, and the plot they might want to see again has already been replaced by another.  The density and complication must necessarily be much more limited than that of a written paper.

Unfortunately, our astronomer reports that some scientists seem unaware of this, and simply read what they’ve already written.  The fact that oral and visual means of communication are different hasn’t occurred to them.  (As well as the fact that laser pointers require skill and care to be of any use at all.  But that’s another subject.)

Digital, analog; oral, written; we have barely scratched the surface of how best to get science across.  It’s unfortunate that lessons here are still to be learned.