A good reading if you want to learn more about misinterpretations of p-values. And it is open access, no subscription needed!
Our colleagues from the Swiss Institute of Bioinformatics organize a one-day course on reproducible research in statistical analysis. It will be held on January 26 2015, in the Biozentrum Annex, Klingelbergstrasse 61, Basel.
For more information, follow the link HERE
Remember to register to the LS2 meeting held in Zurich next January! We organize a poster prize.
Registration deadline has been extended to December 14 2014.
In 2015 we will launch a new initiative on our website to help you with your biostatistics: The Biotelligences Glosssary. We will regularly post definitions of statistical words and concepts (in plain English!) on a dedicated page you can access as needed.
An article in PharmaJournal 23/2014 presenting our poster recently awarded at the Swiss Pharma Science Day.
Download the PDF (French and German):
We had too many registrations for the capacity of the November lecture series. But no need to panic, we organise a second round in January 2015. Venue and content will be the same!
Visit the website of the Lemanic Neuroscience doctoral school for more information.
Congratulations to Sarah Mesrobian from the University of Lausanne (UNIL) who was awarded the first poster prize given by Biotelligences with her poster entitled "Working memory training and risky decision-making: An ERP study in ADHD and control participants".
In this poster, we liked the initiative taken to use non-parametric testing due to the non-normal distribution of the data, even though the large sample sizes (above 30) would have justified normal approximation (central limit theorem). Additionally, complete disclosure was made regarding the readout of Chi2 analyses including degrees of freedom and Chi2 values. Finally, details about study design and analysis were well displayed on the poster and Ms. Mesrobian could readily provide any further information regarding software used or alpha threshold chosen throughout the study.
This was far from the only poster to stand out based on good use of biostatistics and it was not easy to choose the winner. We would also like to congratulate the following authors for the quality of their design and quantitative analyses:
Scariati E. et. al, from the University of Geneva (title: Altered modular organization of the brain network at rest in 22q11DS, title changed to: Graph theoretical analysis of the functional network in 22q11DS: relationship to hallucination); Ciobanu A. et al. from the University of Geneva (title: The effect of action video-games on the processing of attended and unattended emotional stimuli; Schnider M. et al. from the Centre Hospitalier Universitaire Vaudois (title: Gene-environment interactions in psychosis: Evaluation of critical time windows for stress during the neurodevelopment in mice with genetic alteration in the glutathione synthesis).
The Biotelligences team.
Photos available here
Biotelligences poster has been awarded the Vifor Pharma special prize yesterday at the 2014 Swiss Pharma Science Day in Bern.
We thank the committee for this award and the organizers for the excellent meeting.
We look forward to the 2015 meeting.
Biomedical research is often image based; from the bright green cells generated by your transgene to the frenetic trace on your computer screen during an optogenetic experiment. Given this, we sought to generate visually meaningful word clouds, which are graphical representations of word frequency, from recent scientific publications. Our idea was to use these clouds to give visual information on the use of biostatistics by researchers.
Results paragraph: The obsession with P
We first used random text samples from Results sections and figure captions in 18 articles from 5 different journals (Nature, Neuron, Gastroenterology, Biological Psychiatry, British Journal of Pharmacology). The articles and journals were chosen randomly and a total of 2868 words were used. Words with very low occurring frequencies have automatically been excluded by the word generator.
This information clearly matches the results we collected from more than 300 articles in 25 periodicals indicating that: i) 98% of articles rely on p-values to conclude; ii) 84% use parametric tests (such as t-tests) despite detected violations of parametric assumptions; and iii) that 72% use SEM as an error bar, despite their well documented limitations (SEM do not inform the reader about variability but about precision of mean estimation, and the actual mean is not located within one SEM but within a 2-4 SEM-wide interval).
Statistics paragraph: Journal specificities
We are aware that not all statistical tests are cited in the Results sections, but rather in the Statistics paragraphs of Methods sections. Therefore, we also created the following word clouds, each corresponding to an individual journal, and made up of Statistics paragraphs of 4 articles (small/simple words not related to statistics were excluded).
Nature and Science:
Note the relative scarcity of words in the Science cloud, reflecting our difficulty to collect substantial statistical paragraphs. It is also remarkable that the names of statistical software are almost absent from the Science cloud. This points at a general lack of disclosure about statistical methods in Science, an observation supported by our quantitative study.
Note the importance gained by the Kolmogorov-Smirnov test, often used by electrophysiologists to analyse distributions of cell responses. Interestingly the word "one-tailed" is way larger than "two-tailed". Finally, it is worth mentioning that the arsenal of statistical software is larger in this cloud.
British Journal of Pharmacology:
In this cloud, the obsession for the "0.05" threshold and the domination of "SEM" over other error bars are visible again. However, the most visible difference concerns the popularity of ANOVA as compared to other tests, probably linked to multi-groups experimental designs as well as the occurrence of doses responses studies in pharmacology.
We were not aiming, in doing this, to carry out a scientific demonstration into the misuse of biostatistics. Our sample size of a few articles from a handful of journals would be too low to conduct proper descriptive statistics and no real quantitative investigation was performed. In addition, the clouds are only based on disclosed information making flaws by non-disclosure invisible. This later point is critical because a comprehensive and quantitative study that we are currently conducting on hundreds of articles shows that the absence of disclosure is a major flaw in biomedical publishing. Nevertheless, some intriguing features emerged from our clouds and many words highlighted here tend to be in line with the results we obtained in our quantitative studies. In particular the quest for "significant" "p-values" below the "0.05" threshold using parametric "Student's t-tests" or "ANOVA" is pervasive, even when parametric assumptions are not respected.
The Biotelligences team
All word clouds were generated on www.wordle.net
Download the PDF of Biostatistical Word Clouds:
We are delighted to announce the launch of our courses about good practices in statistics! It will take place in November at the Department of Fundamental Neuroscience in Lausanne
For more information visit the website of the school, course title: "Biostatistics for non-statisticians: good practices, misuse and pitfalls"
Detailed programme and registration details here (PDF).
We are happy to announce the first Biotelligences poster prize organised at the Lemanic Neuroscience Annual Meeting. A prize of CHF 500 will be awarded to a poster displaying experimental design and quantitative analysis of outstanding quality, including biostatistics.
Biotelligences Fortnight is no longer displayed on our News page, but you can find its dedicated section here!
Neuroscientists can meet us at the 9th FENS Forum (July 5-9) in Milan (link);
Pharmacologists can visit our poster at the 7th Swiss Pharma Science Day (August 20) in Bern (link);
Biochemists and molecular biologists are invited to attend our minisymposium (September 1) at the FEBS-EMBO meeting (August 30-September 4) in Paris (link).
We look forward to meeting you there!
Romain-Daniel Gosselin from Biotelligences has joined the Editorial Board of Advances in Regenerative Biology (ARB), a new international peer-reviewed open access journal in the areas of developmental biology, stem cell biology, and regenerative medicine. One of ARB highest priorities is to improve the transparency and reproducibility of the published data. To this end all papers are especially evaluated for their technical aspects, research design and analytical methods, in particular in biostatistics.
Visit the journal website here.
Download the promotional card:
We will soon be launching Biotelligences Fortnight, which aims to showcase a recently published article chosen by us, based on scientific impact (the importance to its field) and merit (especially the experimental design, statistical analysis and presentation). Biotelligences Fortnight will be released on the News page of our website and on our Facebook page every two weeks or so. Do not hesitate to send us suggestions from your recent reading.
It was the first poster from Biotelligences and we had a great success. That deserves a thumb up!
Many thanks to Dr. Marc Suter for taking the picture.
You can visit the symposium website to see the official pictures.
We look forward to meeting you in Paris!
Here is a Link to this event:
We will present a poster at the first Symposium organised by the Department of Fundamental Neuroscience, on May 9th at Lausanne.
Visit the website of the symposium.
We are a proud sponsor of the 2014 Orpheus meeting: The international conference on PhD education.
A special issue, with recent archive articles form the Nature Publishing Group, about reproducibility in science:
And a News Feature article in Nature about misuse and misconceptions about p-values.
We can only approve and suggest these readings.
Website of the event:
We will be present at the meeting and look forward to all these riveting talks and posters!
For all attendees: We are sure you will have an amazing scientific (and not scientific) time in Switzerland!
Here: a link to the meeting website.
In the PDF file below you will find the answers, or at least the explainations, to the questions.
Once again thank you for your time.