‘Unfortunately, we have to inform you that after thorough review [YOUR FAVORITE FUNDING ORGANISATION] must reject your application’. Most of us know this sentence all to well, as most rejection letters of our grant applications contain it in a similar form. From a purely statistical point of view, we receive such letters quite frequently. In German biomedicine, the funding rates are between 5 and 25 %, depending on funder and program. Upon receiving a rejection we often feel personally offended. After all, we have put down our best ideas, often had already included some preliminary results and proposed experiments we had already conducted, even beautified the document with a lot of prose, and flattered the most important potential reviewers with strategically placed quotations, etc. And then the rejection! So we had to start over from the beginning, rewrite everything, submit it again, perhaps to another funding agency. This is how we spend a substantial fraction of our days at the office, if we don’t review applications of our colleagues. On average, scientists spend 40% of their time writing or reviewing applications. Continue reading
Triangulation! The Egyptians used it to build their pyramids. The Greeks developed a branch of mathematics out of it. Until the 19th century whole countries were charted in this way. Far into the 20th century ships have determined their position with it. To determine your position by triangulation you only need a set square and a protractor, which the surveyors call a theodolite, as well as the coordinates of two visible landmarks. It’s that simple!
Could it be that triangulation is also an important methodological approach in biology? A cure even for the replication crisis? Munafo and Smith recently postulated this in a commentary in Nature. Sociologists call it triangulation when they use two or more different methods to investigate one particular research question. If the results converge at one point, i.e. lead to the same result, this increases validity and credibility. Don’t we do this routinely in the experimental life sciences? Does the knock-out mouse have the same phenotype as one in which the signalling pathway was pharmacologically blocked? Do transcript and protein expression correlate with the phenotype?
Thus, basic biomedical research is familiar with ‘targeting’ a goal with different methods grounded in already established knowledge (the landmarks of the surveyor!). Are the results converging? Bingo, we have located the biological mechanism! Therefore it leaves many of us cold, if spoilsports with gradschool statistics argue that most studies in biomedicine must be false positive despite significant p-value. Because we don’t just rely on ONE result. Instead we triangulate by means of different approaches! In order to validate results, this might even be superior to replication. If something is simply repeated, it is not unlikely that a systematic error will be repeated too. This would make the result reproducible, but still not correct.
Were the skeptics wrong when calling out a crisis in biomedical research? Are we already doing the right thing? Continue reading
An article entitled “Growth in a Time of Debt” was published in 2010 by the highly respected Harvard economists Carmen Reinhart and Kenneth Rogoff. It dealt with the relationship between national economic growth and national debt. They reported on their discovery of an astonishing, globally observable correlation: As national debt rises, the economic growth of a nation initally also rises. If, however, the national debt exceeds 90 %, this ratio is reversed quite abruptly. Growth turns into contraction, and economic output then declines as debt rises further. The discovery of a “90 % debt threshold” hit like a bomb. Some suspect that the article was the basis for the European austerity policy after the 2008 financial crisis. What is certain, however, is that the paper was enthusiastically used by Western politicians to justify their restrictive fiscal policy. In 2013, Thomas Herndon, a student, reanalyzed the data of the Reinhart-Rogoff paper as part of a semester assignment. After some back and forth, the authors had given him the original Excel spreadsheet. And lo and behold, in a few minutes he found a number of serious errors in it! After correction, the debt threshold disappeared, and the data now appeared to prove the opposite, a steady, positive correlation between government debt and growth across the entire range! What do we learn from this? Apart from the fact that the fundamental error of Reinhart and Rogoff is of course the confusion of correlation with causation: Excel is not suitable for the analysis of complex scientific data. Even more importantly, scientists make mistakes, which can have serious consequences. Continue reading
Recently in a train station book shop I stood gaping in astonishment in front of a thematically highly specialized book display. It was the bowels-brain table. The books piled up on it promised enlightenment about how the bowel and in particular its contents influence us – yes – how, they verily steer our emotions. A selection of book titles: “Shit-Wise – How a Healthy Intestinal Flora Keeps us fit”; “Bowels heal brain heal body”; “Happiness begins in the bowels”, or “The second brain – How the bowels influence our mood, our decisions and our feeling of wellbeing”. Newspapers, magazines and the internet can also tell us this. The wrong bowel bacteria make us depressive – but the right ones make us happy … which is why yogurt helps against depression. Continue reading
U.S. economist Robin Hanson posed this question in the title of an article published in 1995. In it he suggested replacing the classic review process with a market-based alternative. Instead of peer review, bets could decide which projects will be supported or which scientific questions prioritized. In these so-called “prediction” markets, individuals stake “bets” on a particular result or outcome. The more people trade on the marketplace, the more precise will be the prediction of outcome, based as it is on the aggregate information of the participants. The prediction market thus serves the intellectual swarms. We know that from sport bets and election prognoses. But in science? Sounds totally crazy, but it isn’t. Just now it is making its entry into various branches of science. How does it function, and what does it have going for it? Continue reading
With a half-page article written about him and his study, an Israeli radiologist unknown until then made it into the New York Times (NYT 2009). Dr. Yehonatan Turner presented computer-tomographic scans (CTs) to radiologists and asked them to make a diagnosis. The catch: Along with the CT a current portrait photograph of the patient was presented to the physicians. Remember, radiologists very often do not see their patients, they make their diagnosis in a dark room staring at a screen. Dr. Turner in his study used a smart cross-over design: He first showed the CT together with a portrait photograph of the patient to one group of radiologists. Three months later the same group had to make a diagnosis using the same CT, but without the photo. Another group of radiologists were first given only the CT and then, three months later the CT with photo. A further control group examined only the CTs, as in routine practice. The hypothesis: When a radiologist is exposed to the individual patient, and not only to an anatomical finding on a scan, she will be more conscious of her own responsibility, hence findings will be more thorough and diagnosis more accurate. And in fact, this is what he found. The radiologists reported that they had more empathy with the patient, and that they “felt like doctors”. And they spotted more irregularities and pathological findings when they had the CT and photo in front of them than when they were only looking at the CT (Turner and Hadas-Halpern 2008).
So how about showing researchers in basic and preclinical biomedicine photos of patients with the disease they are currently investigating in a model of the disease? Continue reading
It struck at the end of July. A ‘scandal’ in science shook the Republic. Research by the NDR (Norddeutscher Rundfunk), NDR (Westdeutscher Rundfunk) and the Süddeutsche Zeitung revealed that German scientists are involved in a “worldwide scandal”. More that 5000 scientists in German universities, institutes and federal authorities had, with public funds, published their work in on-line pseudoscientific publishing houses that do not comply with the basic rules and for assuring scientific quality. The public and not just a few scientists heard for the first time about “predatory publishing houses” and “predatory journals”.
Predatory publishing houses, whose presentation in phishing mails is quite professional, offer scientists Open Access (OA) publication of their scientific studies at a cost, whereby they imply that their papers will be peer reviewed. No peer review is carried out, and the articles are published on the web site of these “publishing houses”, which however are not listed in the usual search engines such as PubMed. Every scientist in Germany finds several such invitations per day in his or her e-mails. If you are a scientist and receive none, you should be worried about it. Continue reading
- Let’s get this out of the way: Reproducibility is a cornerstone of science: Bacon, Boyle, Popper, Rheinberger
- A ‘lexicon’ of reproducibility: Goodman et al.
- What do we mean by ‘reproducible’? Open Science collaboration, Psychology replication
- Reproducible – non reproducible – A false dichotomy: Sizeless science, almost as bad as ‘significant vs non-significant’
- The emptiness of failed replication? How informative is non-replication?
- Hidden moderators – Contextual sensitivity – Tacit knowledge
- “Standardization fallacy”: Low external validity, poor reproducibility
- The stigma of nonreplication (‘incompetence’)- The stigma of the replicator (‘boring science’).
- How likely is strict replication?
- Non-reproducibility must occur at the scientific frontier: Low base rate (prior probability), low hanging fruit already picked: Many false positives – non-reproducibility
- Confirmation – weeding out the false positives of exploration
- Reward the replicators and the replicated – fund replications. Do not stigmatize non-replication, or the replicators.
- Resolving the tension: The Siamese Twins of discovery & replication
- Conclusion: No scientific progress without nonreproducibility: Essential non-reproducibility vs . detrimental non-reproducibility
- Further reading
There is a lot of thinking going on today about how research can be made more efficient, more robust, and more reproducible. At the top of the list are measures for improving internal validity (for example randomizing and blinding, prespecified inclusion and exclusion criteria etc.), measures for increasing sample sizes and thus statistical power, putting an end to the fetishization of the p-value, and open access to original data (open science). Funders and journals are raising the bar for applicants and authors by demanding measures to safeguard the validity of the research submitted to them.
Students and young researchers have taken note, too. I teach, among other things, statistics, good scientific practice and experimental design and am impressed every time by the enthusiasm of the students and young post docs, and how they leap into the adventure of their scientific projects with the unbent will to “do it right”. They soak up suggestions for improving reproducibility and robustness of their research projects like a dry sponge soaks up water. Often however the discussion is in the end not satisfying, especially when we discuss students’ own experiments and approaches to research work. I often hear: “That’s all very good and fine, but it won’t get by with my group leader.” Group leaders would tell them: “That is the way we have always done that, and it got us published in Nature and Science”, “If we do it the way you suggest, it won’t get through the review process”, or “We then could only get it published in PLOS One (or Peer J, F1000 Research etc.) and then the paper will contaminate your CV”, etc.
I often wish that not only the students would be sitting in the seminar room, but also their supervisors with them! Continue reading
I failed to reproduce the results of my experiments! Some of us are haunted by this horror vision. The scientific academies, the journals and in the meantime the sponsors themselves are all calling for reproducibility, replicability and robustness of research. A movement for “reproducible science” has developed. Sponsorship programs for the replication of research papers are now in the works.In some branches of science, especially in psychology, but also in fields like cancer research, results are now being systematically replicated… or not, thus we are now in the throws of a “reproducibility crisis”.
Now Daniel Fanelli, a scientist who up to now could be expected to side with those who support the reproducible science movement, has raised a warning voice. In the prestigious Proceedings of the National Academy of Sciences he asked rhetorically: “Is science really facing a reproducibility crisis, and if so, do we need it?” So todayon the eve, perhaps, of a budding oppositional movement, I want to have a look at some of the objections to the “reproducible science” mantra. Is reproducibility of results really the fundament of scientific methods? Continue reading