on 6 August 2007
This book contains the essentials of scientific manuscript preparation. As many today obtain their first degree without ever having seen a primary publication, books like this one have become even more essential fare for any PhD student. Gustavii is right about many things. Most importantly, he is right to wonder why so many authors (and editors!) prefer summary data graphs over individual data graphs. He is right about grouped column charts. His discussion of line versus column graphs is also spot-on. It is surprising, then, that he seems to argue *against* box-and-whisker graphs; it becomes clear on the next page, however, that he is in favour of such graphs, but disapproves of the ridiculous width these boxes are given in standard software. I am afraid Gustavii's typical reader will come away from this advice thinking that such graphs are to be avoided. However, on the whole I wish more authors would heed what Gustavii has to say about graphs and tables.
The chapters on authorship, title, abstract and so on are standard fare. It is somewhat discouraging that these things need to be spelled out at all, but, alas, it is, and Gustavii does as good a job as any other number of books in this vein. New to me was his name for the nasty trick of denying earlier work importance or precedence by mentioning it fleetingly, once, deep in the Discussion section. Gustavii tells us that we should not do this, but I rather fear he is dispensing advice to tomorrow's charlatans. The discussion of slashes versus negative exponents is correct, but fails to explain why one is better than the other. Perhaps Gustavii felt that telling biomedical researchers that division is equivalent to subtraction in the exponent does little to enlighten them.
The discussion on common statistical errors is generally good. There is a section entitled "Using mean when median is meant" which misleadingly suggests that authors who make the mistake somehow intend to report the median but end up using the mean, when the problem is, of course, that they are simply not familiar with percentiles. The discussion of confidence intervals is also close but no cigar. Then again, the discussion of abuse of SEM is spot-on and there is a very good section on reporting the effects on risk. Gustavii mentions the public clamour which can result when the relative change in a risk is reported instead of the actual change of itself or, better yet, the Number Needed to Treat. In the vile scare-mongering tabloid environment of the UK, this advice will, sadly, continue to go unheeded.
The parts on manuscript preparation reflects current standard practice without a critical discussion of its grave flaws. I for one do not understand why manuscripts have to be double or triple spaced at the review stage (except that it perhaps discourages idiotic mark-up shenanigans at the other extreme). I also do not understand why people use Word instead of LaTeX, which does away with the need to control authors' typographic gormlessness through restrictive formatting rules. Gustavii tells us "The second most frequent fault in papers submitted is unnumbered pages." Why does this fault occur at all? Because we are hell-bent on remaining in the DTP Stone Age.
The chapter on the review process tries to strike a balance between training the next generation up to be good citizens and preparing them for harsh reality. Gustavii claims referees are honorable with few exceptions, but gives no hard data on how few that is, or what happens if you fall foul of one of these "extremely rare" evil-doers. Gustavii assures us that self-plagiarism will lead to a life sentence (presumably a publication ban). Well, no. I know an emiment mathematician who gets away with it again and again.
Some more niggles: New Scientist is not a prestigious magazine, but a popular publication, and different editorial rules apply (I disagree with the example, not the particular point Gustavii makes). Finally, spelling mistakes and wrong English usage are embarrassing in a book on how to get it right.
on 22 June 2009
I find this book very helpful. It is detailed, perceptive,and authoritative. It provides advice on questions and style and presentation that it is often just assumed that postgradute/academic authors know. I also like the way the book doesn't try to do much - it is a short, succinct, book that focuses on the nitty-gritty of writing and presentation. There's no flannel in this book.