Art Carden  

Who Would Win in an Ideological Turing Test: Young Progressives, or Young Conservatives?

The Tragedy of Economath... Lake Wobegon on the Job...

Yesterday's post on what I've read recently provoked two interesting comments on the title of Michael Adams' Letters to a Young Progressive, one from ThomasH, the other from none other than David Friedman (!!).

They made me wonder: which group would be most likely to pass an ideological Turing test: young conservatives, or young progressives? I have a strong prior that young conservatives would beat young progressives, but such a test would still raise a bunch of interesting questions. How would the results differ in the rural South as opposed to the urban Northeast, for example? Would there be large campus-to-campus differences? Would we see different results at my institution--relatively conservative, Southern Baptist Samford University--versus a more mainstream institution? My prior here is that progressive students at relatively conservative institutions like Samford would do better than progressive students on a more progressive campus simply because they probably spend more time around people with conservative worldviews.

If you're a grad student and you're reading this because you don't want to face the terror of the blinking cursor or because you're not sure what to write about, there's a paper here. Get to work, and let me know what you find post-haste!

Comments and Sharing

COMMENTS (24 to date)
Daniel Kuehn writes:

I think distributions matter too. There are smart conservatives and dumb conservatives and smart progressives and dumb progressives. I think any progressive could probably pass a dumb conservative's Turing test at a higher rate than a conservative could pass a dumb progressive's Turing test. Among low-substance conservative and progressive commentary I find the conservative perspective on progressives to be far more outlandish and unhinged than the reverse.

But if we're thinking about a Turing test judged by a smart conservative or a smart progressive I'd probably agree with you in most cases, because of the exposure issue.

I do wonder sometimes, though. There are some awfully smart people out there with very odd views of what progressives think.

This kind of gets to another point... I think progressives (and to bring this into a more party/institutional discussion - Democrats) are probably a more varied bunch than conservatives. That obviously makes the task tougher. Conservatives are varied, but I feel like they come in much more distinct flavors than progressives, where the types of progressives bleed into each other more easily.

Daniel Kuehn writes:

That comment is 100% impressionistic, of course.

RPLong writes:

I think I can answer this question simply by reviewing my Facebook feed...

Daniel Kuehn writes:

rplong -
+10 on the one hand.

On the other hand, I think most people think this and draw very different conclusions :)

These questions are extremely tough to parse. I also don't always know what a Turing test is supposed to prove sometimes. It seems to me it says as much that one side is easier to mimic (or easier to fool!) than the other. I never know if it really implies that somebody has a deeper appreciation of the arguments.

And people will often switch between good arguments and bad heuristics of arguments themselves. How do we judge that?

Andrea writes:

When Jonathan Haidt asked self-described progressives and conservatives to answer the moral intuitions survey as if they were an average person of the opposing worldview, the conservatives' answers (as progressives) more closely resembled true liberal responses than the other way around.

Theories: conservatives might be more familiar with progressive positions because of media saturation, liberals might be less familiar with progressive positions because they don't actually know many conservatives, or conservatives might have a better understanding of the duality of all 5-6 moral dimensions because they utilize more of them (whereas progressives only employ 2-3.

RPLong writes:

Great points, DK.

For that matter, I've heard some arguments that I hope to never understand.

Daniel Kuehn writes:

Three other problems with ideological Turing tests in an old post I've got here:

I don't think they're pointless, I just always wonder about them.

I think the selection problem I describe is one of the biggest issues.

ThomasH writes:

Maybe I am overgeneralizing from the immigration debate (where conservatives are divided and liberals pretty unified) and climate change debates(where liberals are divided and conservatives pretty unified), but I get the impression that policy disagreements among conservatives concern electoral advantage, whereas liberals disagree about substance.

johnson85 writes:

In addition to the fact that it's difficult for a conservative to live in a bubble, I think the Turing test is easier for conservatives because it does not take any nuance to understand the average liberal's position and also they are not as susceptible to assigning sinister motives.

Take minimum wage for example. A thoughtful liberal might weigh the hardship on workers forced out of the labor market because of a min wage and the short term and long term disemployment effect and decide that the benefit for the workers receiving a bump in wages outweighs the negatives. But most liberals don't get any further than "it feels good to say low wage workers should get paid more." That's pretty easy for conservatives to understand. On the flip side, liberals have to understand the economics to understand opposition to minimum wage, or else they just jump to "conservatives don't care about poor people."

Even when there is a simple position such as "conservatives don't support affirmative action because they think racial discrimination is wrong", or "low taxes are better because people deserve to keep as much as possible of what they earn", there is still the emotionally satisfying explanations of "conservatives are racist" and "conservatives are greedy" that will be to tempting for many liberals to resist.

Carl writes:


Interesting. I think Caplan touched on this a while ago. That while a progressive intellectual may have sophisticated reasons for supporting the minimum wage, to the average lay progressive the justification amounts to little more than "helping the poor". The means are simply presumed to satisy the ends.

As for the Turing test, I don't think there would be much difference between any two average Joes who identify with either label. When it comes to intellectuals though, it seems obvious that conservatives would pass themselves off more convincingly than progressives. "Right-wing intellectual" already sounds paradoxical enough seeing how inseparable intellectuals usually are from the state which supports them e.g left wing intellectuals spending their entire careers investigating new and exciting ways for government to spend everyone's money in the pursuit of social perfection.

Foobarista writes:

The original Turing Test was to see if you could come up with a sim that could fool a human user.

As a right-leaning libertarian, if I were to come up with simple rules to answer "Turing Test" questions as a faux-progressive sim, they would be something like

o Government exists to right social wrongs.
o The only way for meaningful social change to occur is through the organized efforts of government.
o Private charity is guaranteed to be insufficient to the task (whatever it is), as the only entity with enough power to truly make social problems disappear is the government.
o Law enforcement and military are secondary functions of government.
o Expanding government staffing is a good way to expand the middle class.
o Policy is good, more policy is better, trade-offs can be overcome with a proper application of "will".
o Public choice theory issues and bureaucratic priorities as impediments to good government aren't meaningful (except for isolated instances of corruption) because electoral democracy is a frictionless way to transmit the public will onto the bureaucracy.
o Electoral democracy only fails when the people are misled by corporate propagandists, whose actions in the public sphere should be ended by government policy.
o People who work in the government are working for the common good, as is anyone working in the nonprofit sector.
o People who are in the for-profit private sector are not, as proven by the simple fact that they report to owners who expect to profit from their actions. (Only exception: tech companies or companies run by "patriotic" - ie, left-leaning, billionaires)
o The government doesn't really exist as a distinct entity because it's "us".

BZ writes:

I read Mr. Kuehn's criticisms of the ITT. While they seemed to boil down to "we don't have good data yet", he made several interesting points, my favorite of which was something to the effect of: Is there really a Conservative and Progressive (_I_Will_Not_Call_Them_Liberals) philosophy to separate into two neat camps to begin with?

Speaking for myself now, I've long suspected that conservatives and progressives are really groups of people with very different priorities and worldviews, who seem united into 2 distinct groups because, for political reasons, the coalitions inside those groups have each others rhetorical backs.

That makes the question more like: Do true-believer economic progressives understand free market economics and vis-a-versa? Do social progressives understand christians and vis-a-versa? And also, how well *really* do economic progressives understand civil-libertarian progressives?

Max writes:

Andrea's post says (and links to) what I was gonna. Jonathan Haidt has already asked and answered this question. Conservatives are better at pretending to be liberals/progressives than liberals/progressives are at pretending to be conservatives. Further research may undermine or even reverse this finding, but for now the evidence is pretty clear.

Max writes:

I don't think johnson85 would do very well on an ideological Turing test.

Daniel Kuehn writes:

Max -
Precisely my thought. Foobarista too. You guys are starting to move my priors further away from Art.

Hazel Meade writes:

This would be most interesting if it could be made into an online game.

I wonder if you could design a program to "detect" your political allegiance based on answers to questions and then have it ask you to "pretend" to be a progressive/conservative when answering the questions.

The questions would have to be carefully designed not to have obvious answers like political polling questions. I'm guessing that what you would want is some sort of back-and-forth challenge questions that try to get at the subject's intuitive reasoning.

Say on gun control, see if the respondent can explain the rationale for their position in the same way that the liberal/conservative side explains it.

Maybe you could have some sort of machine learning algorithm doing the categorization, which has been trained using actual straight progressive/conservative answers, and see if you can trick the machine into categorizing you the opposite way.

Foobarista writes:

One problem with the "ITT" would be the "test target". Would you be trying to simulate policy level intellectual positions and arguments for the other side, or the "tribal" arguments? (My little rules above were more "tribal" than "intellectual"...)

After all, most people who self-identify as liberals or conservatives didn't arrive at their particular persuasion after long, careful voyages of intellectual self-inspection, but out of a sort of tribal affiliation. (Libertarians seem to me to be less tribal and more intellectual-journey-oriented than either, although not always.) Is the ITT testing whether one could simply pass as a member of the "other tribe", or is it testing whether you deeply understand and can reproduce arguments and rationales of intellectuals of the other side? (The original Turing Test was a simple test: human or clever AI?)

G. White writes:

Daniel Kuehn writes, "There are smart conservatives and dumb conservatives and smart progressives and dumb progressives"

First thing to note is that there are no progressives, everything about their socialist religious doctrine is regressive to constant and consistent failure since it was first dreamed up by that man long ago who feared self responsibility.

So how smart is it for the regressives to be so devout to a failed doctrine that they want to implement it again, force it on people, and see it fail again.

Isn't there a street level definition of insanity in that, trying the same failed thing again and again, each time expecting different if it all can be changed by attaching another lying label on it?

G. White writes:

[Comment removed for rudeness. Email the to request restoring your comment privileges. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Daniel Kuehn writes:

I'll try to rephrase. Anyone who accuses one of these groups of being (1.) socialists, (2.)devotees of a religious doctrine, (3.) not smart, and (4.) liars will probably not do well on a Turing test (the original subject of this post), because none of these rather snide claims really has anything at all to do with modern conservatism or modern progressivism. It would stand to reason, then, that a person with these sorts of views of other people is not going to be able to imitate other people in a Turing test, because they seem to be operating off of an extremely rude (not to mention faulty) understanding of their fellow human beings.

Rob writes:

This question has actually been answered by a liberal social psychologist who studied it. See The Righteous Mind by Jonathan Haidt. His study proved that conservatives are significantly better at predicting what a liberal will believe than vice versa and was able to explain why this is so. Ironically, conservatives operate with a much broader moral foundation; consequently they can understand the arguments of a liberal and reject them, but a liberal simply can't understand the arguments of a conservative because they involve a broad moral perspective that liberals lack.

[urls and html code fixed--Econlib Ed.]

Max writes:

+1 to Daniel Kuehn's most recent comment.

Rob, that is an amusingly biased way of presenting Haidt's research, which has already been mentioned a couple of times earlier in this thread.

Ken P writes:

Here is Krugman showing that he can pass the Turing test:

"You have a Republican base that truly believes that guaranteed health insurance is the work of the devil."

That sounds like many of my otherwise intelligent friends who make statements like "Republicans are pure evil and don't care at all about the poor."

I hear similarly poor characterizations coming from conservatives.

Ken P writes:
I have a strong prior that young conservatives would beat young progressives

That is interesting. I believe the youth in either camp would outperform the elders. But more importantly, I think that the young conservatives are very different in views from the older ones and that the sterotype is diverging from the reality.

Arnold Kling's book the Three Languages of Politics is pretty interesting.

Comments for this entry have been closed
Return to top