Ad Lagendijk
23 March 2010
Should scientists make their own drawings and illustrations?
Tags: commercial, graphs, plots, salesmanPosted in Getting published, High-impact journals
To my disgust I recently discovered that scientist hire the services of commercial art bureau’s to have them make their figures. The resulting glossy, shining figures are supposed to increase the chances of acceptance in high-impact journals as Science and Nature. Nowadays with every submission to these journals the authors include a couple of – professionally made – illustrations, as suggestions to the editors to use them for the cover in case the article gets accepted. What is next? A ghost writer who produces the text, but who is not part of the list of authors?
Kids like illustrations
To keep the attention of children books for kids are full of illustrations, and in many cases they represent animals. Adults are supposed to be able to digest heavier texts. The French quality 
newspaper Le Monde, founded in 1944, waited until 1984 to publish photographs.
Figures in scientific articles
Scientific articles are not published with the intention that they are to be read by children only, but they nevertheless do contain graphic work, referred to as figures. An article has text and possibly figures and tables. Figures come in a variety. Graphs are included to represent tabular numerical data. Diagrams are in the paper to visually represent ideas, arguments, or to show elements of a hierarchical collection. Furthermore all kind of sketches can be found in scientific articles. To show the experimental set-up or an engineering drawing of essential part, or to show the structure of a sample.
The draftsman was on vacation
It is nowadays much and much easier to generate the needed graphic material for an article than it was in the 1980’s. At that
time I was still spending hours on making a single plot on graph paper. Technical drawings had to made by a draftsman and all the figures had to be photographed. As either the draftsman or the photographer was on vacation this was time consuming. Having made slides for a presentation was a nightmare.
Plotting software
At the present time, with reasonably priced, if not open source, software every scientist can make graphs quickly. The software allows for a large collection of numerical mathematical routines to be applied to the data.
Color and 3D extensions
More sand more scientific journals are stopping to publish their paper version. For this and other technical reasons the fees journal ask for the inclusion of colored artwork is going down. So the figures in the scientific journals become colored. Using colors and gray scales opens the way to represent data in three-dimensional plots 
and drawings. Every scientist can make them now himself. The figure on the left was made by me in less than an hour. Using MathCad. Why would I hire a professional agency to beautify my figures?
Beautifying
Articles in a scientific journal are written to communicate with your colleagues. It is not meant for the public at large. Science is all about content and about communicating the content. Form is and should always be secondary to content. Scientists are not marketeers and their papers are not commercials. Although I think due attention should be paid to the presentation of a scientific result, we do not need the commercial world of advertising agencies to sell or science. Let them sell soap.
[CR_show_voting_stars_handler]
Follow
24 Mar 2010 19:19, Mirjam
I think the real question is not who makes the figures, but what criteria journals use to accept papers. If journals accept papers based on content, as they should, it shouldn’t matter how flashy the graphics are. And luckily for you, the people that waste their money on commercial bureaus will then have less money left to do good science and will have fewer papers as a result. But then the crux is not in the commercial bureaus either, because I know enough scientists that have perfected their fancy 3D color graphic fabrication skills and are happy to sacrifice content for form. It is probably normal animal behavior to go for what looks good… (look at male birds).
25 Mar 2010 5:00, David Stern
A lot of authors should get editors to improve their text!
25 Mar 2010 15:41, Philip Chimento
I disagree with this essay completely. So many papers published today have confusing, unintuitive, or just plain ugly figures.
In my opinion, making figures shouldn’t be left to scientists, because we don’t know how! The very ease with which we can make figures nowadays means that nobody is forced to think about the most intuitive way to present their data. People who have had training in, say, information design, know more about how the human brain processes visual information effectively. As scientists, it’s not our job, so we generally don’t know how to design effective figures. And often we are also arrogant enough to assume that it can’t be all that hard, so we slap something together with MATLAB, click submit, and call it a day.
Take a look at any of Edward Tufte‘s books, especially The Visual Display of Quantitative Information, or for a little more “arty” example: http://www.informationisbeautiful.net.
In short: having well-designed, intiuitive, clear, and dare I say nice-looking, figures does not automatically mean that the paper is content-free, or that the author is only trying to impress his/her readers.
By the way, I would guess that most newspapers mainly didn’t print photographs in the past because of technical limitations, not because they thought it was childish. If anybody can provide me with a reference for or against this, I’d be interested to know.
29 Mar 2010 19:14, Ad Lagendijk
@Philip
Articles in scientific journals are written by scientists for scientist. We do not need other professionals to help us to communicate with each other. It adds an extra layer of communication that will only lead to confusion as the artist does not understand the paper and lead to delay.
–
If the paper or book is written for non-specialists I can imagine it could help.
–
I am not much impressed by many of the graphic specialists. Look at the cover of glossy magazines. The text on it is often difficult to read, because the designer has sacrificed legibility for appeal.
2 Apr 2010 14:41, Geert Altena
@Ad
Indeed, adding an extra layer of communication doesn’t help. That said, having the graph or sketch reviewed by a colleague in terms of clarity and legibility surely can’t hurt. Stuffing too much info into a single graph can be killing for an article. If you need 6 line colours or line styles, a reconsideration of the representation might be in order.
@Philip
The Informationisbeautiful site is a very nice collection of well thought-out representations of a wide range of data. Although I have yet to find an example where I could use one of their approaches to one of my graphs.
12 Aug 2010 1:56, Wolfgang
There is some need for most scientists to get at least some advice on making good graphics. Good in the sense of transporting the scientific facts while aesthetically pleasing, too. You can see it in two many journals that mostly the first graph or picture is plotted without any thoughts about how to improve it in any way. On the other side I agree also with those who say that content should be prior to form. But then you first have to blame Nature and Science that they discourage mostly the presentation of plain facts, in favour of some nicely written but often enough superficial text much like advertisement indeed. They even say that readers are usually not interested in too much detail, but science is in the details, that is what is all about!