4 ms·

There is also another hypothesis that likely is messing with the data, more people are working at home and from what I can tell from friends, it is making women

There is also another hypothesis that likely is messing with the data, more people are working at home and from what I can tell from friends, it is making women buy a lot more scented candles (they’re literally selling out once stock is added), which combines with the time and convenience of being able to write reviews that would otherwise not be written. Or, as you described earlier, there would be no review or rating left when smelling candles in store. This seems to be a form of selection bias. I also find it rather shocking how nonchalant people who likely consider themselves educated and scientific, just jump to rather dangerous conclusions; dangerous in that it can contribute to hysterics that have been driving this pandemic response and has us in a catch 22 type situation.

gruez 6y ago
Isn't this a non-issue because they're measuring the proportion of reviews rather than the absolute amount? The selection bias effect that you mentioned still applies, but as long as the effect is consistent between months it shouldn't affect the conclusion.
- ashtonkem 6y ago
  You can’t just count on proportion, because the cohort of people buying candles online in March includes a bunch of people who used to buy candles exclusively in store. If the previously in-store people were more likely to be picky or have different needs, there is a good chance that they would leave different ratings online. It’s hard to separate those effects. It seems plausible to me that someone who insists on buying candles in store is likely to be more picky than someone who buys online, (or based on my time in Candle sections, is already nose-deaf and need very strong candles), and thus would be more likely to complain. The big question is what are they buying. If these reviews are coming from people buying the same brand/scent as before, then Covid related scent loss is the most plausible explanation (although I wonder why some one in such a situation would bother to leave a review....). But if these shoppers are buying new scents for fun or availability reasons, it’s hard to separate those who are losing their smell due to Covid from those who are just dissatisfied with a candle they purchased scent unsmelled.
  - LorenPechtel 6y ago
    Note that in the replies to that tweet there's a Google trends for "I can't smell". The graph looks just like the Covid case counts.
    - ashtonkem 6y ago
      I agree that the similarity between reviews and Covid cases is compelling; I’m just pointing out that there are other possible factors.
mindslight 6y ago
I agree with your comment, except what appears to be driving our response to the pandemic is a debt spiral and cognitive dissonance. I wish people would get hysterical enough to eg not travel for the holidays.

gruez 6y ago

Isn't this a non-issue because they're measuring the proportion of reviews rather than the absolute amount? The selection bias effect that you mentioned still applies, but as long as the effect is consistent between months it shouldn't affect the conclusion.

ashtonkem 6y ago
You can’t just count on proportion, because the cohort of people buying candles online in March includes a bunch of people who used to buy candles exclusively in store. If the previously in-store people were more likely to be picky or have different needs, there is a good chance that they would leave different ratings online. It’s hard to separate those effects. It seems plausible to me that someone who insists on buying candles in store is likely to be more picky than someone who buys online, (or based on my time in Candle sections, is already nose-deaf and need very strong candles), and thus would be more likely to complain. The big question is what are they buying. If these reviews are coming from people buying the same brand/scent as before, then Covid related scent loss is the most plausible explanation (although I wonder why some one in such a situation would bother to leave a review....). But if these shoppers are buying new scents for fun or availability reasons, it’s hard to separate those who are losing their smell due to Covid from those who are just dissatisfied with a candle they purchased scent unsmelled.
- LorenPechtel 6y ago
  Note that in the replies to that tweet there's a Google trends for "I can't smell". The graph looks just like the Covid case counts.
  - ashtonkem 6y ago
    I agree that the similarity between reviews and Covid cases is compelling; I’m just pointing out that there are other possible factors.

mindslight 6y ago

I agree with your comment, except what appears to be driving our response to the pandemic is a debt spiral and cognitive dissonance. I wish people would get hysterical enough to eg not travel for the holidays.