Science - June 6, 2019

The proposition: High impact ≠ high quality

Tessa Louwerens

While he was working on his PhD, Jip Ramakers regularly discussed scientific articles with colleagues in a ‘journal club’. He discovered that even leading journals sometimes publish mediocre articles and he penned a provocative proposition about this: ‘The higher the impact factor of a journal the more suspicious one should be of its content.’

PhD candidates are required to append a few propositions to their thesis. In this feature they explain their most provocative proposition. This time, Jip Ramakers, who graduated cum laude on 8 May for his study on the influence of climate change on evolutionary processes among wild populations of great tits.

The proposition:

The higher the impact factor of a journal the more suspicious one should be of its content.

‘In a journal club, the members take it in turns to choose an article for discussion each week. No one wants to read 30 pages so people often choose shorter, eye-catching articles from high-impact journals such as Nature and Science. As a new PhD student, you are easily intimidated and it is not hard for an article to impress you. But as the discussion went on, we often realized that the data did not fully support the hypothesis or provide the basis for firm conclusions. Particularly in the major journals, news value sometimes takes precedence over  quality. Their popularity creates the impression that the articles are of higher quality, but that is not necessarily so.

Scientists are only human and we all want to be published in leading journals. Me too. One chapter of my thesis was published in Nature Ecology and Evolution. That was because it was a large-scale data study and because we used open science, which is hot right now. I wouldn’t mind betting that it wouldn’t have got into that journal otherwise, because we only found a very tiny effect. It is difficult to get those kinds of minimal results published, but we managed to sell the article anyway by focussing on specific, fashionable aspects. I don’t have a problem with that, as I know it was good quality research. And I also believe that negative results should be published, otherwise you get a bias in the literature.’