Issues with Modern Science


You may not realise how much published scientific research is wrong. If you know about the scientific method and statistics, you might expect the proportion of studies which are incorrect to be around 5%, because that’s the normal level of confidence many areas of science are held to. However, scientific practices in the modern world distort this and make it highly likely that a significantly higher percentage of scientific papers are wrong, let alone the issue of some scientists being immoral. It’s mainly due to problems with the market of scientific journals and researchers misusing statistics.

Journals

Scientific journals are where papers based on studies are published, which show collected data and the conclusions drawn from the data. However, the peer-review process is not cheap — it costs a lot of money to pay qualified scientists to check that a given study is valid and containing as few errors as possible. Hence, the journals require payment for a study to be peer-reviewed and published. This creates a competitive marketplace which contains a lot of bad actors.

Predatory journals don’t follow the rules and are primarily made to make money. Some hide their publication fees until late in the submission process and then prevent a paper being published elsewhere. Some hire peer-reviewers which are not adequately qualified nor vetted just so they can use the number of peer-reviewers they have to try and prove their legitimacy. Some don’t even peer-review at all, they just accept any paper submitted and will publish it if paid, while that journal claims to have a rigorous peer-review process. Scientists who want their paper published but don’t have much money may struggle to pay for submission in a widely respected journal. Instead, they seek journals with lower publishing fees, many of which are predatory but appear legitimate.

A minority of researchers don’t care about the validity of their research, they’re more focused on trying to sustain their career and build prestige through a long list of published papers. They may purposefully try to publish in journals with weak or no real peer-review, while that journal claims to have a proper peer-review system. These bad papers are then published and may be cited by other researchers or used as evidence in an argument. Anyone on the internet can then refer to the deceptive paper as evidence to further their point, but the study may not have stood up to peer-review or may not have been reproducible.

This lets anti-vaxxers and climate-change deniers believe that ‘science’ has evidence in favour of their case; they can just reference these false scientific papers. People looking at the argument will rarely check whether the citation leads to a valid paper, and most people won’t have an easy way of knowing. Even the ones who do will not want to go through the effort of checking every referenced source that they come across is valid. Checking whether a paper is valid can be quite challenging, the publisher of a study will not leave evident deceit in plain sight. And with the papers written in good faith, sometimes the collected data indicates that an effect is real, but it could be due to random chance rather than an observed effect actually being real, or an unconscious misuse of statistics.

This makes it much more challenging to trust science, as any cited source may not be valid, and it may be incredibly difficult to check. Most people won’t try to check at all and may believe a lot of misinformation that provides references to fraudulent or false scientific papers. Newspapers can often publish articles based on invalid studies, spreading misinformation — then it will have already done its damage. Wikipedia tends to be a reasonably reliable source when it comes to this, as it filters out most of the garbage to try and paint a best a picture as possible of the current research in any given area.

Statistical Issues

Another issue is that of random chance and misuse of statistical methods, such as the practice of p-hacking — which can be used maliciously or unconsciously. There is a lot to this, and it’s quite tricky to write about to an audience in which many readers will not have training in university-level statistics, so I’ll refer you to a video by Veritasium on the topic, which also covers other problems with publication in journals:



And here’s an article explaining the set-up study that ‘found’ that chocolate causes weight loss. They consciously used statistical p-hacking to find the result and published in a predatory journal that didn’t review the paper, which was then shared with newspapers who were unaware of what was going on and published articles on the study — possibly deceiving millions of readers of those newspapers.

So, should you trust scientific studies? In the context of a Wikipedia article or from a respected national/international organisation, they are more likely to have checked that a scientific paper is known to be from a reputable source before quoting it. When hearing about scientific studies from other places, you may not want to completely trust them 100% of the time. It’s scary that there can be so much misinformation out there, though it doesn’t mean that most science should be ignored.