This is very valuable input, thanks! In particular, I might have overestimated the cost of such polls. We’ll definitely look into this further and will strongly consider doing such polls.
As a result of this thread, we’re also in the process of obtaining data from a votematch tool. It’s not representative but contains demographic data, so we will be able to extrapolate.
Why should this matter? That ballots succeed without measuring their potential and optimizing against it doesn’t mean anything.
It does mean something, but I’m now 1) less confident in the belief that people don’t do such polling, and 2) think that we have more reason to update away from that prior (how well-informed is it, really?).
Direct conversations with politicians and charities seem similarly helpful for informing framings/communications at much lower cost.
How so? As a charity worker and former political staffer I think I’d have minimal information on a ballot initiative for something like this that I hadn’t done polling on. Why would their expertise be helpful? What information do they have that you don’t? Campaigners would only know about issues on which they’ve researched voting behavior.
I think polls are mostly helpful for strategic decisions such as how many resources to invest into the vote campaign. For refining our communication strategy, observing the media coverage and updating our strategy based on that will be more useful. Politicians also have intuitions for how people will react to various points, e.g. for the kinds of arguments opponents will bring up, or whether a particular party will support the initiative. They might be able to give us recommendations for how to make sure the parliament recommends accepting the initiative. This is especially true in the frequently-debated question of foreign aid. It’s not clear what a poll result like “45% plan to vote YES” means for the communication strategy.
I don’t think the last point is decisive though, we should simply do both.
This is very valuable input, thanks! In particular, I might have overestimated the cost of such polls. We’ll definitely look into this further and will strongly consider doing such polls.
As a result of this thread, we’re also in the process of obtaining data from a votematch tool. It’s not representative but contains demographic data, so we will be able to extrapolate.
It does mean something, but I’m now 1) less confident in the belief that people don’t do such polling, and 2) think that we have more reason to update away from that prior (how well-informed is it, really?).
I think polls are mostly helpful for strategic decisions such as how many resources to invest into the vote campaign. For refining our communication strategy, observing the media coverage and updating our strategy based on that will be more useful. Politicians also have intuitions for how people will react to various points, e.g. for the kinds of arguments opponents will bring up, or whether a particular party will support the initiative. They might be able to give us recommendations for how to make sure the parliament recommends accepting the initiative. This is especially true in the frequently-debated question of foreign aid. It’s not clear what a poll result like “45% plan to vote YES” means for the communication strategy.
I don’t think the last point is decisive though, we should simply do both.