You’re reporting statistics (mean, quantiles) of the ratio of cost-effectiveness estimates, and not the ratio of the mean cost-effectiveness estimates, right? I think the ratio of the means is more informative, since we normally compare expected values per $ spent.
Assuming QALYs and cQALYs reflect similar circumstances and are normalized similarly (I think that’s the intention), the most informative figures for me would just be the ratio of means assuming moral weight is 1, times a placeholder discount factor for moral weight. So something like “CCCW is 5,000m times as cost-effective as MIF, where m is the relative moral weight of chickens to humans times the relative probability of chicken consciousness to human consciousness.” (Replace 5,000 with whatever number it turns out to be, but I’m guessing it’s around that, based on your mean moral weight of 2.41.)
If you’re including meat eater problem effects (and only considering chickens/poultry, and not other animals, farmed or wild), then I’d guess the ratio of means would look like f(x)=ax/(b+cx), where x is the relative moral weight times consciousness probability variable, and you could report the function and its constants, and plot its graph.
You’re reporting statistics (mean, quantiles) of the ratio of cost-effectiveness estimates, and not the ratio of the mean cost-effectiveness estimates, right?
Yes, right.
I think the ratio of the means is more informative, since we normally compare expected values per $ spent.
That makes sense. I have updated the post such that I now analyse the ratio between the statistics of the cost-effectiveness distributions instead of the statistics of the ratio between the cost-effectiveness distributions.
Both the ratio between the mean cost-effectiveness of CCCW and MIF, and the mean ratio between the cost-effectiveness of CCCW and MIF are 12 k, so there were no changes in the conclusions.
Assuming QALYs and cQALYs reflect similar circumstances and are normalized similarly (Inthink that’sthe intention), the most informative figures for me would just be the ratio of means assuming moral weight is 1, times a placeholder discount factor for moral weight.
I think I had better use a concrete best guess for the expected moral weight, such that the expected conclusions are clearer. That being said, I agree that is quite informative, so I have added the following to the Discussion:
This ratio [between the mean cost-effectiveness of CCCW and MIF] is equal to the product between 4.78 k and the mean moral weight of chickens relative to humans if these are moral patients, which would have to be 2*10^-4 (= 1/(4.78 k)) for the ratio to be 1.
For this analysis, I have not accounted for the “meat eater problem effects” I analysed here.
You’re reporting statistics (mean, quantiles) of the ratio of cost-effectiveness estimates, and not the ratio of the mean cost-effectiveness estimates, right? I think the ratio of the means is more informative, since we normally compare expected values per $ spent.
Assuming QALYs and cQALYs reflect similar circumstances and are normalized similarly (I think that’s the intention), the most informative figures for me would just be the ratio of means assuming moral weight is 1, times a placeholder discount factor for moral weight. So something like “CCCW is 5,000m times as cost-effective as MIF, where m is the relative moral weight of chickens to humans times the relative probability of chicken consciousness to human consciousness.” (Replace 5,000 with whatever number it turns out to be, but I’m guessing it’s around that, based on your mean moral weight of 2.41.)
If you’re including meat eater problem effects (and only considering chickens/poultry, and not other animals, farmed or wild), then I’d guess the ratio of means would look like f(x)=ax/(b+cx), where x is the relative moral weight times consciousness probability variable, and you could report the function and its constants, and plot its graph.
Thanks for the feedback!
Yes, right.
That makes sense. I have updated the post such that I now analyse the ratio between the statistics of the cost-effectiveness distributions instead of the statistics of the ratio between the cost-effectiveness distributions.
Both the ratio between the mean cost-effectiveness of CCCW and MIF, and the mean ratio between the cost-effectiveness of CCCW and MIF are 12 k, so there were no changes in the conclusions.
I think I had better use a concrete best guess for the expected moral weight, such that the expected conclusions are clearer. That being said, I agree that is quite informative, so I have added the following to the Discussion:
For this analysis, I have not accounted for the “meat eater problem effects” I analysed here.