Thanks for your comment! And no worries about not polishing, I will do the same, so it will also be a bit long :)
I agree with your concern and it is something Iāve also thought about before (in other contexts as well). However, I see two reasons for why working in high-population countries should indeed be favoured:
At Animal Advocacy Africa weāre currently working on recommendations and implementation guides for advocates that aim to mitigate the rise of industrial animal agriculture in Africa. Based on our research, policy work is the top recommendation and I do think the expected value of this is higher in high-population countries. The reason is that it is hard to know where policy work is more likely to be successful (which you also mentioned). As long as we donāt have an indication that it is significantly less likely to be successful in higher-population countries, it seems fair to focus on the factor that we know will be important: the expected impact, if successful.
For work besides the area of policy/āregulations (e.g. working with farmers or certain public outreach interventions, which are our recommendations #2 and #3), I agree that scale considerations can be overblown. If we cannot cover the whole population anyway, there is no limit that should really matter. However, I think scalability and potential flow-through effects are important to consider here. If we can get a successful model to work for some part of a large country, there is the potential to scale this much further or to have it scale automatically across the country (e.g. word of mouth).
In short, there is a lot of upside to working in such large countries and as long as I donāt have evidence that working in smaller countries is much more tractable I would keep focusing on the large ones. However, if there is clear evidence that working in a specific country is likely to be significantly more tractable, we should give this consideration a lot of weight. Unfortunately our rough model is not well-suited for such nuances, so it should definitely be combined with contextual knowledge/āfactors.
That said, I think it is a good point that the weight might be too high and these weights are mostly based on our intuitions anyway. So itās great that you are challenging this. I think it would probably be fruitful to do some kind of MC simulation on how the scores change if we vary the weights of different parameters. Maybe Iāll find time for this somewhere down the road.
Thank you for your thoughtful reply. Some thoughts.
As long as we donāt have an indication that it is significantly less likely to be successful in higher-population countries, it seems fair to focus on the factor that we know will be important: the expected impact, if successful.
Lobbying smaller bodies of government is definitely easier. Whoever decides on policies in small countries has fewer bids of attention and is targeted by fewer lobbyists. You might need a lot of connections and effort to make your voice heard to a decision-maker in a big body of government. In a small body of government, you might be able to set up a meeting by writing an email without any prior connection. Thereās definitely a trade-off of scale vs tractability here. And to me, itās not obvious at all which choice would be more cost-effecitve. Iām not talking from experience here, itās just my common sense intuitions.
If we can get a successful model to work for some part of a large country, there is the potential to scale this much further or to have it scale automatically across the country (e.g. word of mouth).
I agree that country borders impact word of mouth but Iām not sure how much. Especially in Africa since Iāve heard that African borders were drawn kind of randomly and I donāt know how important they are culturally. For example, if I look at Africa language map like this, I see that bigger countries have many languages. Language barriers might limit the meme spread within the country. And it also seems that languages often cross national boundaries, Meme spread through internet content, TV, and radio might often transcend national boundaries, I imagine. But I donāt know how much, I know little about Africa.
Itās just food for thought, I think your view is reasonable and you probably have already thought about these things. You could just reduce the weight of the variable a little bit if I convinced you a little bit :)
Yes, this has certainly updated my view on prioritisation between big and small countries. So thanks for sharing your thoughts!
I think itās a good idea to reduce the weight of scale, though probably not as much as you might. Aashish and I might update this as soon as we got around to talking about it and are aligned.
In any case, we encourage people to just take the model, make a copy, and change parameters themselves, if it seems useful for their purposes.
Nice ^_^ One final thought. I mentioned that scale depends on multiple parameters:
Current human population
Expected growth in the human population
Current animal production per capita
Expected change in the production per capita
You account for 2,3, and 4 with a separate variable āexpected growth in animal productionā which would be something like āprojected number of farmed animals in 2050 divided by the current number of farmed animalsā. And then also have a variable āCurrent human populationā. I think it makes sense to split because these two variables matter for different reasons, and someone may put weight on one but not the other.
Thanks for your comment! And no worries about not polishing, I will do the same, so it will also be a bit long :)
I agree with your concern and it is something Iāve also thought about before (in other contexts as well). However, I see two reasons for why working in high-population countries should indeed be favoured:
At Animal Advocacy Africa weāre currently working on recommendations and implementation guides for advocates that aim to mitigate the rise of industrial animal agriculture in Africa. Based on our research, policy work is the top recommendation and I do think the expected value of this is higher in high-population countries. The reason is that it is hard to know where policy work is more likely to be successful (which you also mentioned). As long as we donāt have an indication that it is significantly less likely to be successful in higher-population countries, it seems fair to focus on the factor that we know will be important: the expected impact, if successful.
For work besides the area of policy/āregulations (e.g. working with farmers or certain public outreach interventions, which are our recommendations #2 and #3), I agree that scale considerations can be overblown. If we cannot cover the whole population anyway, there is no limit that should really matter. However, I think scalability and potential flow-through effects are important to consider here. If we can get a successful model to work for some part of a large country, there is the potential to scale this much further or to have it scale automatically across the country (e.g. word of mouth).
In short, there is a lot of upside to working in such large countries and as long as I donāt have evidence that working in smaller countries is much more tractable I would keep focusing on the large ones. However, if there is clear evidence that working in a specific country is likely to be significantly more tractable, we should give this consideration a lot of weight. Unfortunately our rough model is not well-suited for such nuances, so it should definitely be combined with contextual knowledge/āfactors.
That said, I think it is a good point that the weight might be too high and these weights are mostly based on our intuitions anyway. So itās great that you are challenging this. I think it would probably be fruitful to do some kind of MC simulation on how the scores change if we vary the weights of different parameters. Maybe Iāll find time for this somewhere down the road.
Thank you for your thoughtful reply. Some thoughts.
Lobbying smaller bodies of government is definitely easier. Whoever decides on policies in small countries has fewer bids of attention and is targeted by fewer lobbyists. You might need a lot of connections and effort to make your voice heard to a decision-maker in a big body of government. In a small body of government, you might be able to set up a meeting by writing an email without any prior connection. Thereās definitely a trade-off of scale vs tractability here. And to me, itās not obvious at all which choice would be more cost-effecitve. Iām not talking from experience here, itās just my common sense intuitions.
I agree that country borders impact word of mouth but Iām not sure how much. Especially in Africa since Iāve heard that African borders were drawn kind of randomly and I donāt know how important they are culturally. For example, if I look at Africa language map like this, I see that bigger countries have many languages. Language barriers might limit the meme spread within the country. And it also seems that languages often cross national boundaries, Meme spread through internet content, TV, and radio might often transcend national boundaries, I imagine. But I donāt know how much, I know little about Africa.
Itās just food for thought, I think your view is reasonable and you probably have already thought about these things. You could just reduce the weight of the variable a little bit if I convinced you a little bit :)
Yes, this has certainly updated my view on prioritisation between big and small countries. So thanks for sharing your thoughts!
I think itās a good idea to reduce the weight of scale, though probably not as much as you might. Aashish and I might update this as soon as we got around to talking about it and are aligned.
In any case, we encourage people to just take the model, make a copy, and change parameters themselves, if it seems useful for their purposes.
Nice ^_^ One final thought. I mentioned that scale depends on multiple parameters:
Current human population
Expected growth in the human population
Current animal production per capita
Expected change in the production per capita
You account for 2,3, and 4 with a separate variable āexpected growth in animal productionā which would be something like āprojected number of farmed animals in 2050 divided by the current number of farmed animalsā. And then also have a variable āCurrent human populationā. I think it makes sense to split because these two variables matter for different reasons, and someone may put weight on one but not the other.