Spin found in clinical trial abstracts in top psychology and psychiatry journals

https://sciencebeta.com/spin-psychiatry-journals/

by Indirector • 7 years ago

66 21 7 years ago

notadoc• 7 years ago

They're called soft sciences for a reason, they're more about interpretation and opinion than hard irrefutable evidence. So this is not surprising, it's probably quite a bit higher than 50% too, and in all of the soft science fields where agendas or political motives may be present; sociology, psychology, political science, some archaeology, etc

wallace_f2• 7 years ago

If you can't follow the scientific method most of the time, should their work be considered science at all?

utopian3• 7 years ago

This is my concern. If you can only partially trust the science journals because the science is faulty, what “science” can you trust? ... unless you attempt replicate every piece of research yourself?

mattkrause4• 7 years ago

Are you actually arguing that randomized controlled trials aren't following the scientific method?

mattkrause4• 7 years ago

Rubbish.

The papers analyzed here are reporting the outcomes of randomized controlled trials. Political motivations, opinions, and interpretations have nothing to do with any of these papers' results.

A fair number of HN commentators have this bizarre caricature of psychology as someone lying on the couch, chattering about feelings while a tweed-jacketed man scribbles away. However, it hasn't been 1920 for a long time and the field has grown up and moved on. It's an actual science, with experiments and quantitative analysis\* and all that. The problems it tackles are really hard, in some ways evenmoreso than physics: you can generate an endless stream of identical electrons, but every human is different and has their own motivations.

\* Yes, yes, replication crisis, fMRI salmon of doubt. It's serious but psychology should get some credit for a) identifying it in the first place and b) fixing it. These problems certainly aren't confined to psych--it's mostly that psych experiments are easier to replicate.

candiodari2• 7 years ago

Have you ever talked to people in the social sciences ? There very much is political motivations, opinions and interpretations. And don't forget careers. Many people have built decade-long careers on being the go-to person (or even department, it can be a dozen careers) for specific treatments. Look up what EMDR is, for a particularly ludicrous example, and read the career of Francine Shapiro. Note that she had decades in which she had a great position to collect support and evidence for her therapy, as a professor, director of a treatment center and practicing clinician ... and somehow she never did (mind if I make the assumption that she did in fact collect plenty of data, it just said what all psychiatric studies say about all treatments)

Psychoanalysis or "someone lying on the couch, chattering about feelings while a tweed-jacketed man scribbles away" was done until deep in the 80s, where it was canceled because of the famous observation that it was exactly as effective as giving patients a free membership card to a sport club (not even checking if they used it at all: so some patients wouldn't have).

But truth is that it was not just done because psychoanalysis wasn't effective. It in fact psychoanalysis is definitely the most effective treatment in psychiatry. It was done so attention from expensive psychiatrists could be removed entirely from patients, as a cost-saving measure. Now all psychiatrists do is draw out treatment plans, to be executed by people with much less training than they have. Now a patient in mental health treatment is lucky to see a psychiatrist once a month.

All these treatments found are known to be far less effective than psychoanalysis. Psychiatry as a whole is simply a fraud. There isn't a single known treatment against depression, other than fixing the problem that causes the depression (which is external to the patient, and never has anything to do with psychiatry)

For the same story with added drama, see https://en.wikipedia.org/wiki/Rosenhan_experiment

gervase• 7 years ago

I think this is a natural outcome of the peer review process; you're not targeting research papers at statistical models, you're targeting them at human reviewers. This kind of behavior is equally prevalent in many fields, including computer science.

I recently reviewed a paper from a research team in China, on a scalable and ubiquitous ad hoc sensing platform. The presented "spin" was that it could potentially be useful for monitoring power distribution networks, thereby reducing blackouts, which was certainly true. However, it could also have been used for an inescapable, totalitarian panopticon (this part was implicit and unstated).

As a reviewer with explicit instructions to evaluate the ethical implications of the work, is it my responsibility (A) to interpret the stated "spin", (B) extract and evaluate strictly the technical contributions of the work, or (C) all of the above, and the implicit "anti-spin" interpretation that was provided courtesy of my Western worldview?

And here's a twist - six months previously, a similar situation occurred with my own work, in which a technology that I envisioned (spun) being used for disaster relief caught the eye of an alphabet agency for low-cost, secure data collection in third world theaters. This was deeply shocking to me at the time, because I hadn't considered this potential use case when I was developing the system. I had spun its impact so effectively in my own mind, I didn't even see the negative alternative.

It's my observation that human brains will inject their own spin if it doesn't already exist, using the social and cultural norms to which that brain is accustomed. I don't hold a researcher's desire to sell their work as effectively as possible against them, especially if the person to whom they're selling most effectively is themselves.

gentran• 7 years ago

I'm a novice in my IT career. If you're able and willing, would you mind explaining to me how software (spun as it may be) for disaster relief, could be used for data collection outside the scope of the objectives you had in the name of disaster relief? If this question misses the mark you're welcome to ignore it. Thanks!

AstralStorm• 7 years ago

It's probably image analysis from overflight, if I were to guess, or an ML-augmented database.

Noumenon72• 7 years ago

This article might be spin itself, given how it captures attention with the accusation but hides what was actually tested with the circumlocation "a previously published definition of spin". For all I know, they just counted every abstract that contained the word "may" and called it spin.

quickthrower2• 7 years ago

Also: "Spin" found in over 50% of top quantum mechanics journal abstracts.

pacbard• 7 years ago

The issue is probably two-pronged: 1. There is pressure from the department (and funders, as the article mentions) to produce publications. The need to publish leads to researchers to spin null results as a way to "cash in" the time and effort put into a research project (I'm assuming good intentions here on the part of the researchers). 2. Articles get published if the authors can argue that their work makes a new and unique contribution to the field. This argument usually rests less on the article's results and more on the discussion of these results. I am not surprised that authors spin the results in their discussion sections.

As a side note, most readers only read the introduction and discussion section of a paper. Those two sections usually have the most spin as they are written to "hook" readers in the paper (and to an extent, to be accessible to a non-technical person sometimes outside of the field). Lit review, methods, and results sections are usually more dry as they report on the technical aspects of a research project.

your-nanny3• 7 years ago

huh? we read the methods and results first, then the spin (exception, the abstracts)

Fomite• 7 years ago

Several major journals, including nature, put the methods at the end.

your-nanny3• 7 years ago

Ah, but if they had not found spin we would never have heard of their study...

SubiculumCode• 7 years ago

Abstracts are typically between 150 and 250 words, yet studies, especially clinical trials, are frequently complex affairs. Fitting nuance into an abstract is just often not possible, and frankly, the clearest most succinct language usually oversells the results.

Spin found in clinical trial abstracts in top psychology and psychiatry journals

Related Stories