[XFrequentist]

Agree that it's a poor headline and that the fructose/Alzheimer's hypothesis is quite speculative, but I don't think calling it a "junk study" or gesturing at evidence hierarchies is particularly helpful.

Theory (almost) always precedes evidence, and coming up with a novel, biologically-plausible explanation for a common ailment is absolutely a valid, useful scientific contribution.

Your general point, that drawing firm conclusions would be radically premature, is spot on. I just stiffen up a bit when I encounter "RCT or GTFO" type arguments; where in the world do you think the ideas for which RCT to run come from?

[colincooke]

While I agree that these theory based papers are useful, and are often the precursor to experiments, I believe the general understanding of "study has found X" in pop culture is that there is "hard evidence" of the finding being tauted. Theories are risky to place too much credence in without being steeped in the field yourself (is this a theory that most people in the field agree with, or is the one suggesting it an outlier?).

As usual science communication is never done as well as we could all hope, but I personally like this "hierarchy of evidence" approach in understanding if something is ready to be consumed by the general public, rather than requiring further discussion with the scientific community.

[jstummbillig]

Agreed. At this point I am also not willing to just let science off the hook and blame it all on the press: If our smartest people can not find ways to differentiate between ideating and good results in a way that a sensationalist press can't simply ignore, then just maybe they are not trying all that hard.

[dv_dt]

I wish sci comms practice would have a standard set of terms for stages of development/belief. Here the headline should be something like “theory proposed that …”

[birdman3131]

We have that term already. Hypothesis.

[a4isms]

Theory (almost) always precedes evidence, and coming up with a novel, biologically-plausible explanation for a common ailment is absolutely a valid, useful scientific contribution.

This is the case in physics and astronomy. People make predictions that are not only untested, but we have to invent equipment to test them.

One of the great early accomplishments in science was when astronomers, observing our Sun, noted an unknown yellow spectral line signature. In 1868, Norman Lockyer predicted that it must be created by a hitherto unknown element, which he named “Helium” after the Greek Titan of the Sun, Helios.

In 1895, two Swedish chemists detected helium in ore samples here on Earth, and in the great tradition of the scientific method, we had a theory, a prediction, and a confirmation of the theory by test.

http://braythwayt.com/2017/12/29/crown.html

[naasking]

I'm not sure it's always the case that theory precedes evidence, even in physics. We still have loads of evidence that theory does not explain, dark matter, dark energy and superconductivity being the most well known. Theory is often devised to explain something weird or unexpected that experimentalists observe.

[a4isms]

Upon reflection, you’re making an excellent point!

Einstein predicted things we’re still confirming empirically, but then again, he started by asking himself how the speed of light could be constant for every observer, which was the result of an experiment that failed to confirm a different theory.

[naasking]

Science is an ouroboros of experiment and theory!

[mtlmtlmtlmtl]

There's nothing wrong with research like this per se. I think the way it's publicised is the main issue. Broadly I think there are two types of papers: those who are only relevant to other researchers, and those who are interesting to the public as a whole. This is very much in the former category.

[kuhewa]

Theory is good. Not sure calling it a study in a headline is a good idea though.