If you read polls saying that the public supports a carbon tax, should you believe them? According to today's guest — journalist and blogger Matthew Yglesias — it's complicated, but probably not.

Links to learn more, summary and full transcript.

Interpreting opinion polls about specific policies can be a challenge, and it's easy to trick yourself into believing what you want to believe. Matthew invented a term for a particular type of self-delusion called the 'pundit's fallacy': "the belief that what a politician needs to do to improve his or her political standing is do what the pundit wants substantively."

If we want to advocate not just for ideas that would be good if implemented, but ideas that have a real shot at getting implemented, we should do our best to understand public opinion as it really is.

The least trustworthy polls are published by think tanks and advocacy campaigns that would love to make their preferred policy seem popular. These surveys can be designed to nudge respondents toward the desired result — for example, by tinkering with question wording and order or shifting how participants are sampled. And if a poll produces the 'wrong answer', there's no need to publish it at all, so the 'publication bias' with these sorts of surveys is large.

Matthew says polling run by firms or researchers without any particular desired outcome can be taken more seriously. But the results that we ought to give by far the most weight are those from professional political campaigns trying to win votes and get their candidate elected because they have both the expertise to do polling properly, and a very strong incentive to understand what the public really thinks.

The problem is, campaigns run these expensive surveys because they think that having exclusive access to reliable information will give them a competitive advantage. As a result, they often don’t publish the findings, and instead use them to shape what their candidate says and does.

Journalists like Matthew can call up their contacts and get a summary from people they trust. But being unable to publish the polling itself, they're unlikely to be able to persuade sceptics.

When assessing what ideas are winners, one thing Matthew would like everyone to keep in mind is that politics is competitive, and politicians aren't (all) stupid. If advocating for your pet idea were a great way to win elections, someone would try it and win, and others would copy.

One other thing to check that's more reliable than polling is real-world experience. For example, voters may say they like a carbon tax on the phone — but the very liberal Washington State roundly rejected one in ballot initiatives in 2016 and 2018.

Of course you may want to advocate for what you think is best, even if it wouldn't pass a popular vote in the face of organised opposition. The public's ideas can shift, sometimes dramatically and unexpectedly. But at least you'll be going into the debate with your eyes wide open.

In this extensive conversation, host Rob Wiblin and Matthew also cover:

• How should a humanitarian think about US military interventions overseas?
• From an 'effective altruist' perspective, was the US wrong to withdraw from Afghanistan?
• Has NATO ultimately screwed over Ukrainians by misrepresenting the extent of its commitment to their independence?
• What philosopher does Matthew think is underrated?
• How big a risk is ubiquitous surveillance?
• What does Matthew think about wild animal suffering, anti-ageing research, and autonomous weapons?
• And much more

Chapters:

Rob’s intro (00:00:00)
The interview begins (00:02:05)
Autonomous weapons (00:04:42)
India and the US (00:07:25)
Evidence-backed interventions for reducing the harm done by racial prejudices (00:08:38)
Factory farming (00:10:44)
Wild animal suffering (00:12:41)
Vaccine development (00:15:20)
Anti-ageing research (00:16:27)
Should the US develop a semiconductor industry? (00:19:13)
What we should do about various existential risks (00:21:58)
What governments should do to stop the next pandemic (00:24:00)
Comets and supervolcanoes (00:31:30)
Nuclear weapons (00:34:25)
Advances in AI (00:35:46)
Surveillance systems (00:38:45)
How Matt thinks about public opinion research (00:43:22)
Issues with trusting public opinion polls (00:51:18)
The influence of prior beliefs (01:05:53)
Loss aversion (01:12:19)
Matt's take on military adventurism (01:18:54)
How military intervention looks as a humanitarian intervention (01:29:12)
Where Matt does favour military intervention (01:38:27)
Why smart people disagree (01:44:24)
The case for NATO taking an active stance in Ukraine (01:57:34)
One Billion Americans (02:08:02)
Matt’s views on the effective altruism community (02:11:46)
Matt’s views on the longtermist community (02:19:48)
Matt’s struggle to become more of a rationalist (02:22:42)
Megaprojects (02:26:20)
The impact of Matt’s work (02:32:28)
Matt’s philosophical views (02:47:58)
The value of formal education (02:56:59)
Worst thing Matt’s ever advocated for (03:02:25)
Rob’s outro (03:03:22)

Producer: Keiran Harris
Audio mastering: Ben Cordell
Transcriptions: Katy Moore

Revisiting A Conversation w/ Tony Robbins