4 ms·

Here’s the source of that statistic: https://dariuszzdziebk.wpenginepowered.com/wp-content/uploads/2015/09/Loss_et_al._2013-Impacts_Outdoor_Cats.pdf https://dar

by dahart 3y ago

Here’s the source of that statistic: https://dariuszzdziebk.wpenginepowered.com/wp-content/uploads/2015/09/Loss_et_al._2013-Impacts_Outdoor_Cats.pdf https://dariuszzdziebk.wpenginepowered.com/wp-content/upload... The number does sound high, and the paper acknowledges that, but I guess something published in Nature should be taken seriously. I’ve only had time to barely skim it, but it appears that they might be saying the majority of the issue comes from un-owned cats that roam free, not cats that live with (and are fed by) humans in a house. I think they took data on the number of kills per cat and then multiplied by the estimated number of cats. Edit: I also looked for rebuttals or critical review, and came across a defense of the paper by the same authors, which has references to the criticism: https://link.springer.com/article/10.1007/s10530-018-1796-y https://link.springer.com/article/10.1007/s10530-018-1796-y

kjkjadksj 3y ago
The big assumption with the data is that all cats are equivocal in hunting ability and access to resources, when we know this isn’t the case.
- dahart 3y ago
  That is one of the accusations to their methodology, yes, but their defense makes a good point that the accusations here have been levied by pro-cat organizations that are firmly against considering policies to reduce free-ranging and feral cats, and that these accusations have not been peer reviewed and are not correct. They address multiple claims about misusing extrapolation in their defense article. I’m not defending them because I don’t know enough about this and it would take a long time to fully evaluate the methodology and criticisms. But, I really doubt this is as obvious or clear-cut as you suggest. It seems like a bad idea to jump to the conclusion that they’ve made the simplest of statistical mistakes and got it past the reviewers at Nature Comms. This journal is one of the more difficult to get into. The authors appear to have tried to be careful, and moreover they are at least using actual data, where the rest of us arm-chair commenters are basing our ideas off of beliefs and emotions and wishes for certain outcomes, which are all notoriously inaccurate, right? Another thing we do know is that behaviors tend to lump into distributions (often Normal Distributions), and that given enough samples, we can make pretty strong claims about whole populations (as long as you’ve captured true representative samples). I did skim the paper enough to see that they put together probability distributions for the bird and other-mammal killing rates. They explicitly acknowledge a pretty wide range of feline hunting rates, no?

by dahart 3y ago

kjkjadksj 3y ago

The big assumption with the data is that all cats are equivocal in hunting ability and access to resources, when we know this isn’t the case.

dahart 3y ago
That is one of the accusations to their methodology, yes, but their defense makes a good point that the accusations here have been levied by pro-cat organizations that are firmly against considering policies to reduce free-ranging and feral cats, and that these accusations have not been peer reviewed and are not correct. They address multiple claims about misusing extrapolation in their defense article. I’m not defending them because I don’t know enough about this and it would take a long time to fully evaluate the methodology and criticisms. But, I really doubt this is as obvious or clear-cut as you suggest. It seems like a bad idea to jump to the conclusion that they’ve made the simplest of statistical mistakes and got it past the reviewers at Nature Comms. This journal is one of the more difficult to get into. The authors appear to have tried to be careful, and moreover they are at least using actual data, where the rest of us arm-chair commenters are basing our ideas off of beliefs and emotions and wishes for certain outcomes, which are all notoriously inaccurate, right? Another thing we do know is that behaviors tend to lump into distributions (often Normal Distributions), and that given enough samples, we can make pretty strong claims about whole populations (as long as you’ve captured true representative samples). I did skim the paper enough to see that they put together probability distributions for the bird and other-mammal killing rates. They explicitly acknowledge a pretty wide range of feline hunting rates, no?