3 ms·

> Funding: This work was supported by Procter and Gamble Not bash the researchers or the quality of their research but I'm unsure if a negative result would ha

> Funding: This work was supported by Procter and Gamble Not bash the researchers or the quality of their research but I'm unsure if a negative result would have been acceptable to their continued funding. I'm sure P&G would like to use this research to sell their diffusers after all.

jklinger410 3y ago
This scientific study brought to you by Glade's new oil scent diffusers. Buy one at your local Wal-Mart or Target today! This is where conspiracy theories are born, and why they are so hard to fight.
- gumby 3y ago
  These disclosures are a relatively new phenomenon, and sure, they can feed conspiracy theories but they can also be used to fight them.
culi 3y ago
Well if you read the paper you'll see that the Introduction cites a lot of other research that is not supported by P&G. Not that you shouldn't be skeptical. Just pointing out that they didn't just pull the idea for this study out their ass. Also the statistical significance and magnitude of effect is quite mindblowing
godelski 3y ago
These are always hard to read and I want to make a mention because we are critical in this area (like medicine and food) but when it comes to stuff like ML no one brings it up (pretty much all the ML papers you see are supported or directly from large tech companies). Usually if the results are positive they're not exactly made up. Negative results just don't get reported. You should always maintain skepticism though and be critical. Those are still good qualities and I don't want to discourage that in any way. (Sorry, little ranty below) I say "not exactly" because every researcher does embellish their work a bit. I'm not going to directly criticize the researchers here (well I will) because this is more a product of the reviewing process (i.e. researchers) who are often quick to reject. The bigger issue is that reviewers see themselves as filters (i.e. adversarial to the reviewed work) and not as editors or consultants (i.e. on the same team as the reviewed work). I'm not sure if this is a psychological thing or incentive (both for sure) but it's probably something we should be discussing due to the feedback loop that it creates and is resulting in more a prolific environment where it is impossible to distinguish an ad from science. Which isn't good for science and especially for people trying to read science (i.e. why science communicators are often shit). I'm also a big believer that we should be publishing negative results. Fuck man, most research is failing. You find out how to do something by failing to do it a hundred/thousand different ways. Why aren't we encouraging this? I thought papers were just about us scientists communicating with one another...

jklinger410 3y ago

This scientific study brought to you by Glade's new oil scent diffusers. Buy one at your local Wal-Mart or Target today! This is where conspiracy theories are born, and why they are so hard to fight.

gumby 3y ago
These disclosures are a relatively new phenomenon, and sure, they can feed conspiracy theories but they can also be used to fight them.

culi 3y ago

Well if you read the paper you'll see that the Introduction cites a lot of other research that is not supported by P&G. Not that you shouldn't be skeptical. Just pointing out that they didn't just pull the idea for this study out their ass. Also the statistical significance and magnitude of effect is quite mindblowing

godelski 3y ago

These are always hard to read and I want to make a mention because we are critical in this area (like medicine and food) but when it comes to stuff like ML no one brings it up (pretty much all the ML papers you see are supported or directly from large tech companies). Usually if the results are positive they're not exactly made up. Negative results just don't get reported. You should always maintain skepticism though and be critical. Those are still good qualities and I don't want to discourage that in any way. (Sorry, little ranty below) I say "not exactly" because every researcher does embellish their work a bit. I'm not going to directly criticize the researchers here (well I will) because this is more a product of the reviewing process (i.e. researchers) who are often quick to reject. The bigger issue is that reviewers see themselves as filters (i.e. adversarial to the reviewed work) and not as editors or consultants (i.e. on the same team as the reviewed work). I'm not sure if this is a psychological thing or incentive (both for sure) but it's probably something we should be discussing due to the feedback loop that it creates and is resulting in more a prolific environment where it is impossible to distinguish an ad from science. Which isn't good for science and especially for people trying to read science (i.e. why science communicators are often shit). I'm also a big believer that we should be publishing negative results. Fuck man, most research is failing. You find out how to do something by failing to do it a hundred/thousand different ways. Why aren't we encouraging this? I thought papers were just about us scientists communicating with one another...