Rachael Miller

This chapter does four things. First, it considers several questions about the proposed methodology. Second, it answers several objections to the methodology, many of which center on the results of implementing it. Third, it identifies several ways we could improve the methodology going forward, improving the empirical rigor of our approach. Fourth and finally, it takes stock of the project and provides our overall view of its significance. We emphasize that insofar as it’s appropriate to use our welfare range estimates for decision-making, they are placeholders for the main result, which is that the differences between species are smaller than some may have anticipated (or, more cautiously, that there is little empirical basis for postulating larger differences).

This chapter provides some tentative welfare range estimates, where a welfare range is understood as the difference between the most intense positively valenced experience and the most intense negatively valenced experience available to members of a species. These estimates are conditional on hedonism, which means that they don’t reflect the implications of uncertainty about the correct theory of welfare, and they are intended as a proof of concept, so they do not factor in every possible complexity even given hedonism. The chapter’s main aim is to explain a method we can use to generate estimates. Along the way, we provide an overview of the proxies we used for differences in the possible intensities of valenced states across species and a summary of the literature reviews about those proxies for several farmed species. Finally, we explain how those reviews can be converted into specific estimates.

Given that there are no direct interspecies measures of the intensity of valenced experiences, we outline a methodology for estimating welfare ranges that does not rely on such direct measures. This methodology has four steps: First, specify the determinants of welfare. Second, identify measurable proxies for variation in the ability to realize the determinants of welfare. Third, survey the empirical literature for evidence about these proxies. Fourth, aggregate the results of that literature review to generate welfare range estimates. We then review some of the limitations of this approach, which include the problem of missing data, the difficulty of navigating phylogenetic distance, and puzzles about scoring and aggregating the proxies.

This chapter does four things. First, it considers several questions about the proposed methodology. Second, it answers several objections to the methodology, many of which center on the results of implementing it. Third, it identifies several ways we could improve the methodology going forward, improving the empirical rigor of our approach. Fourth and finally, it takes stock of the project and provides our overall view of its significance. We emphasize that insofar as it’s appropriate to use our welfare range estimates for decision-making, they are placeholders for the main result, which is that the differences between species are smaller than some may have anticipated (or, more cautiously, that there is little empirical basis for postulating larger differences).