home > epc > autumn 2015 > lost in translation
European Pharmaceutical Contractor

Lost in Translation

Communication is an integral and essential part of science, for without the dissemination of results, science is little more than a hobby. Indeed, the progressive nature of science is dependent on the communication of results. Yet, despite this, many people in science seem to think that data presentation is unimportant and that the data ‘speaks for itself’.

It sometimes appears as if many scientists believe that the more complex and incomprehensible the data presentations are, the more scientific they are – just take a glance at the majority of scientific posters on the walls of the next meeting you attend. You will often see tables and graphs which fail to articulate any story, but rather simply act as repositories for the data collected.

Big Data, Big Problems

According to Calvin Schimd, author of Statistical Graphics: Design Principles and Practices, “It is indeed strange – perhaps even incredible – that social scientists and statisticians who would be outraged by ungrammatical sentences, misspelled words, or even improper punctuation, are so indifferent, if not actually receptive of, crude, clumsy, and amateurish charts”(1). This is very unfortunate, as science involves data, data means numbers, and numbers are best understood using tables or graphs. A well-designed data presentation can explain the main theme or story of even very large and complex data sets.

The advent of the big data era will only exacerbate this problem as the huge amounts of data that will become available will make interpretation even more challenging. As poorly conceived and executed tables and graphs can actually prevent an understanding of the aims and conclusions of a publication, we will need to pay greater attention to the presentation of data to avoid being swamped under by the incoming deluge promised by big data.

Guiding Principles

In a departure from the misled motivations already expressed, there are undoubtedly principles of good data presentation which can, in turn, help us utilise them to communicate. First up: tables. There are of course, lots of small things that can improve a tabular presentation, but the main tenets can be summarised as follows:
  • Every table should have a message
  • Do not add data that is not relevant to the table just because you have it
  • Put numbers that are to be compared next to each other (to show the message) and not at opposite ends of the table
Of course, all three of these principles require one to know the main message of the data. The simple next step is to ask: what question do these data answer, or why were these data collected in the first place? Thus, safety data in a clinical trial answers the question of whether a treatment or intervention is safe, compared to a placebo or an active comparator. Patient demographic and baseline characteristic data in a comparative trial answer the question of whether the treatment groups are comparable for demographics. As a consequence, any discrepancies observed in the trial can be attributed to the difference in treatments. From experience, when presented with data from a trial, people often begin looking at it in detail before determining what the question actually is – needlessly complicating the message, if not completely missing it.

Visual Masterpiece

Graphic presentations (line graphs, bar charts, scatter plots, pie charts), as a result of their intrinsic diversity, are more complex, and yet the same three principles of good presentation still apply. Many graphs become hopeless masses of lines, data points, or a blizzard of colours in an attempt to present every single data point, rather than the actual message itself. A particularly good example of this is the near ubiquitous scatter plots seen in publications and scientific posters. Many appear as simple clouds of coloured dots, often with several colours representing different study groups all in the same plot, with no way to compare or make sense of the data. All that is needed to make a scatter plot comprehensible and convey a message is to add regression lines or the mean averages. The aim should always be to produce a clear display that quickly communicates the answer to the fundamental question of why the data was collected in the first place.

A classic and oft-cited example of this is the Charles Minard graphic of Napoleon’s invasion of Russia in 1812, dubbed by Professor Edward Tufte of Yale University, “[as] probably the best statistical graphic ever drawn”(2). Although this statistical graph contains thousands of data points (numbers of troops superimposed on a two-dimensional map of Russia), the message is clear within an instant; it beautifully conveys the magnitude of what is widely considered to be one of the greatest military catastrophies of all time.

Graphical Abstracts

A recent development in scientific publishing even suggests that good data presentations which clearly display the main message may soon be required for publication. A number of scientific journals have begun requesting that graphical abstracts be included when a manuscript is submitted to a journal. According to the Elsevier For Authors website,“A graphical abstract should allow readers to quickly gain an understanding of the main take-home message of the paper and is intended to encourage browsing, promote interdisciplinary scholarship, and help readers identify more quickly which papers are most relevant to their research interests. Authors must provide an image that clearly represents the work described in the paper. A key figure from the original paper, summarising the content, can also be submitted as a graphical abstract”(3).

Scientific Politic

Hopefully this excellent idea will spread throughout the world of scientific publishing, and perhaps even extend to conference posters. If the use of well-designed, clear, infographic-type displays of scientific information becomes widespread, it will greatly facilitate and increase the communication of scientific results, with the potential to hasten the pace and value of scientific discoveries. Given the range of environmental and technological threats the world currently faces, this is surely something we can all support.


1. Schmid C, Statistical graphics: Design principles and practices: p2, 1983
2. Tufte E, Visual display of quantitative information 914: p40, 1983
3. Visit: journal-authors/graphical-abstract

Read full article from PDF >>

Rate this article You must be a member of the site to make a vote.  
Average rating:

There are no comments in regards to this article.

Barry Drees received his PhD in Molecular Genetics at the University of California, San Francisco. Following his postdoctoral work as a fellow of the US National Institute of Health, he worked as a medical writer in the pharmaceutical industry at Hoechst/Aventis for 12 years, setting up a Phase 1 writing group and leading several regulatory submission teams. Barry is a frequent speaker on medical writing, statistics, and other scientific communication topics for various pharmaceutical associations. He is the former Editor-in-Chief of The Write Stuff, and was President of the European Medical Writers Association. He is currently a co-founder and Senior Partner of Trilogy Writing & Consulting, leading submission teams and providing international training.
Barry Drees
Print this page
Send to a friend
Privacy statement
News and Press Releases

Sphere Fluidics Closes a $4.8 million USD (£3.7 million) Funding Round

Cambridge, UK, 10 June 2019: Sphere Fluidics, a company commercializing single cell analysis systems underpinned by its patented picodroplet technology, announced today that it has closed an additional tranche of investment of $2.9 million (£2.2 million), securing a total of $4.8 million (£3.7 million) in investment. The funding will be used as working capital for the Company’s Cyto-Mine® Single Cell Analysis System, expansion of the UK facility and enable the doubling of personnel through the opening of a new US Sales Office and Demo Lab in California.
More info >>

White Papers

Planning the Execution of a Viral Clearance Study

Eurofins BioPharma Product Testing

Viral clearance studies are a necessary component of any regulatory submission for clinical trials or commercial product approval for all biopharmaceutical products. These studies are performed to evaluate the capability of the purification process to remove or inactivate viruses that could potentially contaminate the starting material. They are complex studies that require substantial financial and personnel resources, as well as specialized scientific expertise to perform. As such, viral clearance studies are often performed at a qualified contract testing laboratory rather than in-house. When multiple parties are involved in this process, clear communication and a comprehensive understanding of the approach and timeline is critical.
More info >>

Industry Events

The Universe of Pre-filled Syringes and Injection Devices

22-23 October 2019, The Swedish Exhibition & Congress Centre Gothia Towers Hotel

PDA’s Universe of Pre-filled Syringes and Injection Devices has become the must-attend meeting for everyone working in the field, and is now the world’s largest conference on this subject
More info >>



©2000-2011 Samedan Ltd.
Add to favourites

Print this page

Send to a friend
Privacy statement