So the statistical term for what you're saying is that the paper didn't use bonferroni correction. Which is relevant (but unrelated to the p < .05 arbitrary boundary).
> That's why investigation is done until there are many more 9's
Nobody said the investigation is done. What is being said is that this alone is a red flag that warrants a lot of investigation (and the investigation shouldn't have to be research by colleges whenever a graduate student gets around to it). This is something the EPA should fund an attempt to reproduce.
It's the EPA's job to prevent dangerous chemicals from being sold in public goods. It's not college's job to need to conduct exhausting studies with absolute proof with many 9s that something is unsafe and then hand it off to the EPA to act. Once any reasonable doubt is raised the EPA needs to take ownership.
This investigation is the bare beginning. It may turn out to be completely wrong. It does not warrant 'a lot of investigation'. There are so many spurious results out there, it would be a fools errand to dig into every curious result.
And no I did not mean some correction. I meant, that publishing is a filter for any seemingly correlated result. It will inevitably include simple coincidences, because of the sample sizes involved.
To put it simply, all coincidental results end up being published, because they look so remarkable. It's a reason to be very skeptical of small samples.
>>> I meant, that publishing is a filter for any seemingly correlated result.
Great, we all know what publication bias is. Sure there's a non-zero chance that this is a statistical anomaly.
But my dude, what exactly is going on in your head? Numerous flame retardants have already been found dangerous and toxic [1]. Are you seriously concerned people are going to take their health too seriously and throw away their couches because of this study?
I think the exponentially more likely, and exponentially more dangerous concern is these flame-retardant chemicals aren't taken seriously enough... particularly since it's debatable whether they were ever necessary in the first place.
What's going on is, the 'emotional outrage of the minute' eco-topics that keep coming up. I have limited investment to spend in outrage. I'm not spending it on every preliminary report that might show some correlation.
For instance, tiny cats (6lbs? 10?) that sleep on furniture for 8-10 hours a day show elevated flame retardant in an absorbent tag.
Do the cats respond in the same way as humans to these chemicals? (No)
Is the acceptable dose for a 10lb cat different from a 100lb person? (Yes)
Was anything else measured in the experiment (No?)
Did the experiment seek out any other explanations? (No)
This reeks of a cherry-picked goal-directed experiment to create outrage. It seems the opposite of science.
> That's why investigation is done until there are many more 9's
Nobody said the investigation is done. What is being said is that this alone is a red flag that warrants a lot of investigation (and the investigation shouldn't have to be research by colleges whenever a graduate student gets around to it). This is something the EPA should fund an attempt to reproduce.
It's the EPA's job to prevent dangerous chemicals from being sold in public goods. It's not college's job to need to conduct exhausting studies with absolute proof with many 9s that something is unsafe and then hand it off to the EPA to act. Once any reasonable doubt is raised the EPA needs to take ownership.