Arnold Kling  

Changing People's Minds

Robin Hanson Argues from Autho... My Speech in Denver on Monday...

What causes people to change their minds? On this blog, I have argued that people do not change their minds on ideological issues. I have argued that econometric regression results typically do not change people's minds. Why is that?

1. Changing your mind could mean acknowledging a loss of status. If you are an academic, and you achieved status because of your articulation of theory X, then you will not give up on theory X, even if years later much of the profession has switched to theory Y. As Max Planck put it, "Science progresses one funeral at a time."

2. Changing your mind could mean loss of group identity. I suppose I could use the unkindest cut example. But let me stick with economic beliefs. If your think of yourself as belonging to a group of conservative Republicans, you probably are going to resist any analysis that promotes Keynesian economics.

On political issues, the cost of dogmatic beliefs is small. First of all, your beliefs have almost no chance of affecting policies, much less affecting your personal well-being. Hence Bryan's rational irrationality for voters.

Second, in the realm of public policy, there is always evidence that supports either side. As my father would have put it, it's the First Iron Law of Social Science: sometimes it's this way, and sometimes it's that way.

There are dozens of factors involved in the financial crisis. By focusing on the factors that involve private sector misconduct and market failures, you can build a case for enhanced government power. By focusing on factors that involve policy errors and bad incentives created by regulation, you can build the opposite case.

On political issues, I think that it is harder to change the mind of someone who is highly educated than someone who is not. The highly-educated person is more likely to have his sense of status and identity tied up in his political beliefs. He is more likely to have a made a larger investment in finding facts and theories that confirm his beliefs. See Jeffrey Friedman on "ideological elites."

I am still staggered by Mark Thoma's remarks about potential Democratic support for vouchers. Yes, there are pro-voucher Democrats. And there are pro-choice Republicans. But if you are prepared to believe that the Democratic Party is anything other than anti-voucher or that the Republican Party is anything other than anti-abortion, I have to think that you are suffering from some sort of defense mechanism for avoiding cognitive dissonance. That is, you have the problem that "I like my party. I believe in policy X. My party opposes policy X." You need a strategy to resolve the dissonance, and one strategy is to pretend that your party is not really hostile to party X.

In politics, love is blind, in the sense that you tend to filter out negative information about your preferred candidates and parties. The more highly educated you are, the more powerful are your filters.

In theory, higher education teaches "critical thinking," the scientific method, and the need for an open mind. In practice, this is not so simple.

What about instances in which people's minds have changed? Here are some:

1. Economists have changed their minds on the main sources of economic growth. They used to focus entirely on capital accumulation. Now they look at the intangible factors that Nick Schulz and I describe in our forthcoming book. Some of this change of mind even can be attributed to an econometric result--Robert Solow's finding of the "residual" in an equation that used capital to explain economic result. I would say, in retrospect, that this sort of equation by itself would not have persuaded anyone strongly committed to the view that capital accumulation explains growth. This may be a case where no one's status was really threatened by the finding, and eventually other economists developed status by confirming and Solow's the finding. Is this an example of academics changing their minds and, if so, what accounts for it?

2. Policy makers changed their minds about treating inflation as a problem of controlling the "wage-price" spiral. It is now viewed as a monetary phenomenon. Evidently, the notion that wage and price controls damaged the economy has sunk in. That is quite remarkable, when you consider all of the policies that I think have damaged the economy where the fact has not sunk in--housing policy, for example. Wage and price controls seem to me to be an exception. What accounts for it?

COMMENTS (33 to date)
Daniel Kuehn writes:

On the econometric findings - I think you also have to keep in mind that there really is no consensus even in the empirical work, so why would you expect one paper (like the paper you posted earlier) change anyone's mind? It shouldn't be dismissed, obviously, but I think people consider a wide set of econometric results and they are cognizant of the really tough identification questions associated with estimating a multiplier.

So it's not that "econometric findings don't change people's minds" - they do. The point is one econometric finding won't. And that's probably a good thing!

I like your First Iron Law of Social Sciences - that's a very good thing to keep in mind.

Mr. Econotarian writes:

I was actually on the ballot as a Democratic party elector at one point. The nauseating effect of the Democratic primary and its manipulative politics was very disillusioning.

What put me over the edge into libertarianism was when a police office got behind me, essentially pushed me over the speed limit, pulled me over, and then suggested that he search the car for drugs. I asserted my 4rth Amendment rights, and after a long heated discussion I went on my way.

I found the experience very upsetting, having apparently been targeted by the War on Drugs because of my age, hair length, and a geeky bumper sticker that may have been misinterpreted as a Grateful Dead sticker.

Shortly thereafter, I became a Libertarian Party convert, and actually decided to try illegal drugs for the first time. Then I got to look into the blue glow of a core of a research nuclear fission reactor, and my friends took me handgun shooting for the first time.

It took time for me to drop back from extreme libertarianism. The reasonable (not perfect, but not horrible) track record of the Fed since the 1980's brought me over to fiat currency, along with learning more about how the rigid peg to the gold standard was a major cause of the Great Depression, and FDR's elimination of the peg caused some immediate economic improvement.

Learning more about economics has made me more "econotarian". Liberty is fine when it works, but regressions have shown me that specific government institutional quality is linked with economic growth: security of property rights, currency stability, etc. Economic Freedom research has been very illuminating as well.

Redland Jack writes:

I think there are two main problems econometric models have with overturning theories:

1) Omitted-Variable Bias - Almost anything that is interesting requires a large number of independent variables. Frequently, things that seem important to the model are difficult to measure/proxy.

2) It's difficult to tell whether a model was 'true' or the result of data mining. It seems that to get published you need to show 'positive' results, which often entails running your model repeatedly until it gets the 'right' results.

I've long said that it takes a theory to beat a theory (though I'm sure I stole that from somebody).

spencer writes:

I use to think that libertarians had something to offer until I started reading the GMU blogs and realized that it was a pure fantasy that had no basis in reality.

DW writes:

My favourite story about changing people's minds is in the concise encyclopedia of econ about Coase:

"Stigler recalled:

We strongly objected to this heresy. Milton Friedman did most of the talking, as usual. He also did much of the thinking, as usual. In the course of two hours of argument the vote went from twenty against and one for Coase to twenty-one for Coase. What an exhilarating event! I lamented afterward that we had not had the clairvoyance to tape it.1"

Bo Zimmerman writes:

Spencer, except for the enlightenment era, where a brief experiment with freedom caused the real, existing, non-fantasy Western world to emerge into our current prosperous modern world, you are probably right -- freedom sucks, we should all have just remained serfs.

Doc Merlin writes:

Hrm, I used to be left leaning on fiscal issues and right leaning on defense... then I became a rather extreme libertarian.

Two things did this, one was a 6 hour long argument about morality on a long car trip and the other was when I began reading Ayn Rand, David D. Friedman, and Heinlein.

People do change their minds, its just rare that ideologues change their minds because they have too much personal stake in it. Also, the older someone is the less likely they are to change, I find. This may be why large scale social change happens after baby booms?

Prakhar Goel writes:


It seems to me like you are focusing on the wrong thing. Would it not be much more effecient to accept people's unwillingness to change as a fact of human existence and spend time on charting its effects and uses?

Greg Ransom writes:

Mises book _Socialism_ changed Hayek's mind about Fabian socialism. Wicksell changed Hayek's mind about Bohm-Bawerk's average period of production model, as did exchanges and conversations between a whole group of economists.

Hayek's _The Road to Serfdom_ and conversations with Stigler changed Friedman's ideology from non-ideological to classical liberal.

Hayek's _TRtoS_ changed Popper from a socialist to a market supporting social democrat.

Ideas do change minds.

Wage controls are associated with war, and no one likes the idea of his own wage being controlled. A huge majority has a wage. So, it was easy to reject wage controls when there was some data that they actually hurt the overall economy.

Price controls on gasoline produced waiting lines that eventually convinced people that something wasn't working. Those lines disappeared when the controls were lifted. The harm was obvious and annoying, and the removal of controls worked nicely. So, no more obvious price controls.

Prohibition of alcohol didn't work, is associated with funding criminals, and was denounced by it early supporters. A huge majority likes a drink. So, it was possible to change away from Prohibition, even though it required a constitutional amendment to start it and to end it.

Beliefs and policies change when the problem is personal, annoying, closely related to the policy, and affects a large majority.

Sean writes:

Did the stagflation of the 70's play a part in reconsidering the dominant models because it did not comport with their world-view?

Ryan Vann writes:

I believe the answer to the question is rather simple. Regression results can be dismissed primarily because the error of omission critique is never "not valid." Ideological change requires convincing rhetoric, empirical data plays the supporting role.

MikeDC writes:

The message on wage and price controls has sunk in? Hence the country being slow motion frog marched to a single payer health care model?

fundamentalist writes:

"I have argued that people do not change their minds on ideological issues."

1. Young people are more likely to change their minds than older people, which explains why young people were so quick to trash a century of excellent classical economics and embrace the hodgepodge of nonsense snake oil that Keynes peddled.

2. Ideas are divided into opinions, attitudes and world views (also called paradigms). Opinions are easy to change. The examples you give are examples of opinions. Attitudes are very difficult to change. Attitudes are like choosing between Austrian and Keynesian economics. I'm constantly amazed at the fact that mainstream economists would rather take a hot poker in the eye than read Hayek. World views are like the differences between capitalism and socialism that involve fundamental issues like the nature of mankind and evil. Worldviews are next to impossible to change, though it's easier with young people.

3. Hazlitt has a great book available at "Thinking as Science" in which he details how prejudice works. Very enlightening.

4. In order to know the truth about anything, you have to be able to put yourself in a state of mind in which you don't care what the consequences of truth are; you just want the truth. Only then will you be able to overcome prejudices and find truth. Most people are not capable of reaching that state of mind.

bgc writes:

There have been vast, rapid and continuing changes of mind among the elites of Western nations since about 1964 - political correctness, for example, is an inversion of the common sense of humankind at all other times and places.

Maybe this provides the answer: people will change their minds if by doing so they believe they can avoid being publicly denounced as evil.

This rationale also covers the other rapid mass brainwashings in 20th century totalitarian societies.

jr writes:

"I have argued that people do not change their minds on ideological issues. I have argued that econometric regression results typically do not change people's minds."

This is in direct contrast to my experience. Ideologically, I was educated in Marxism. At the same time I developed a vague leaning towards capitalism. (I didn't even know all that it entails.) When I grew older, I became a liberal Democrat, if I could use American spectrum to describe my experience. At one point I was reading anarchist materials and seriously contemplating all that many left and right variants of anarchism. Then I found libertarian and I thought I was settled. But now, I consider myself a conservative of the Rush Limbaugh brand.

Empirically? Data always give me heart attacks. I never can dismiss data simply because it contradict my ideology. On the other hand, it should be common sense to not to change position by every paper you read. At least read two, I mean.

So your reasoning about the two propositions is, I am afraid, a mirage. At least to a significant degree 'cause I doubt that I am that much exceptional.

fundamentalist writes:

jr: "I doubt that I am that much exceptional."

Actually, I think you are exceptional. Look at the history of the US. We have been moving toward greater socialism since the late 1800's. There was no break in the trend until roughly Reagan's presidency. The growth in socialims slowed, but did not stop. Now it is going full speed again.

You mentioned Rush Limbaugh, but I have to disagree with you that Limbaugh is a conservative. I used to be a huge fan of his. But his support of Bush's socialist policies in "No child left behind", the Medicare drug presciption program, and stimulus package made me realize that Limbaugh and all of the other "conservative" talk show hosts are not conservatives. They're Republicans.

And Republicans have never been conservatives. They keep conservatives in the party by tossing them a fake bone once in a while and using conservative slang, but if you ignore the talk and look at their actions, they're only slightly less socialist than Democrats. Look at John McCain!

I'm hard core libertarian, but I voted for Obama because I knew what a fraud McCain is. He is a socialist as any Democrat. If McCain had won the election, he would have fooled people into thinking they had a conservative president while he advanced socialism. As a result, voters would leave the Democrats in charge of Congress in the 2010 elections.

So why do I care if I see no difference in the parties? Because while they're both socialists, they care about power and being in power. One will oppose the other simply because it is not in power. For example, the Democrats fought vehemently against Bush's prescription drug plan. Had a Democrat president proposed it, they certainly would have supported it. Libertarians will never be numerous enough to elect a libertarian president. The best we can hope for is gridlock. If Congress and the presidency are held by opposite parties, gridlock will result.

Steve Roth writes:

And there's the (current) 800-pound gorilla of false beliefs--that rich countries with lower taxes grow faster than their higher-taxing counterparts.

The econometrists have demonstrated that it just ain't so. But most people--even on the left--still believe it.

Steve Roth writes:

Richard Dawkins tells a remarkable story on this subject. Check out the paragraph at the bottom of this page, beginning, "It does happen."

When I heard him recount this in person, it brought tears to my eyes.

Michael Rulle writes:

Excellent post.

Still feel vaguely dissatisfied with it however. It is probably the nature and complexity of the subject matter itself---human economic and political activity.

I disagree that more educated people are less likely to change their mind. People in general will be less likely to change their minds if it negatively impacts their self interest---and vice versa.

If one measures growth of GDP or GNI in US over last 100 years it seems to have little to do with which party is in power. This could be evidence that within a very broad spectrum of possible political and economic choices our disagreements exist within a very narrow zone.

You seem to use the concept of "ideological" interchangeably with concepts of "true or false". Admittedly, many things are true and false subject to a first premise such as a moral belief/ideology.

Your Thoma example is very compelling but probably at the extreme of unusual and subject to change even by him.

But what is "vaguely" bothering me the most?

1)I guess it seems bizarre that any economist can support massive new regulations and taxes as proposed in health care. The world's health care economic model seems bizarre and we are moving toward it. The model is "free for all" or "cheap for all" and "we will subsidize it for you and hide its real costs until we have to ration it from you". The health care system is already bizarre, with this being designed to make it more so. Yet economists are not uniform---publicly at least---in opposition.

2)Man made Global Warming. Putting aside the paucity of falsifiable experiments, it is virtually without doubt that all that is proposed in the these Waxman-Markey monstrosities will simply subsidize certain businesses, make energy more expensive, and do little to zero about CO2 emissions--or at least enough to impact global warming--whose theories themselves are almost entirely data-fit. Yet economists are not publicly uniformly against it

3)Stimulus spending, higher marginal tax rates, a hodge podge of randomly inserted vacuum sucks of private money to government---all currently being proposed---and all creating uncertainty. Not just on their passage---but even with the "certainty" of knowing they would be passed we retain the uncertainty of government's ability to randomly change incentives whenever it feels like it. Again, no uniform disagreement by economists.

It is possible these are simply subjective value differences and all agree on the economic impact of such policies. But that is not my impression.

Economics is much harder than physics--clearly. But certain principles seem obvious too. Yet many are being violated now--with economists seemingly equally split--or even leaning toward support of these policies.

It just feels like we are missing something.

Alan Crowe writes:

There is a conspicuous omission from the post which leaves me
struggling to determine the actual question.

Common practice in econometrics is to acknowledge that
correlation does not imply causation, but to then carry out
regressions, controlling for this and that, under the
mistaken impression that this helps. When the researcher is
confronted with his error he makes Bambi eyes at his critic
and pleads "Please don't make me read Pearl".
Or perhaps he bargains "I won't criticize your research if you don't
criticize mine. Don't worry, Clark Glymour won't find out about us."

The apparent question "Why econometric regression results
typically do not change people's minds." asks why invalid
arguments are unpersuasive. Duh? Because they are invalid?

I'm more aware of this in the medical field with all its talk of
risk factors and healthy living advice that unravels every time
it is subject to a randomised controlled trial, but there is
plenty to discuss in the economic area.

What do I believe about minimum wage laws? I think they destroy
. That is not my only opinion in this area. I think
that stupid politicians, such as John Prescott, don't
understand this, while clever ones, such as Tony Blair,
do. Obviously the clever one becomes Prime Minister and the
stupid one becomes his Deputy, so it is the clever one who
decides policy. Over-raising the minimum wage is a political
disaster. If an over-high minimum wage starts sticking out
as a destroyer of jobs, the government has no good
option. Cutting the minimum wage poisons the party's
feel-good brand, it makes them the hard-hearted don't care
party. Leaving it high, and letting the unemployed rot on
the dole poisons the party's feel-good brand, it makes them
the hard-hearted don't care party. That is not the kind of
choice that a politician wants to face.

So my prediction is that politicians will try to predict
when a tightening labour market is going to be raising wages
and will try to time increases in the minimum wage so that
they can claim credit. I also predict that they will be
cautious about this, just in case the labour market fails to
tighten and they end up in the lose-lose scenario. So I know
what I expect to see: a negative correlation between changes
to the minimum wage and unemployment, due to cautious

Some people think that minimum wage laws don't destroy jobs
and we should raise it higher. The usual idea for
"empirical" research in this area is to look at increases to
the minimum wage and see whether such increases are followed
by rises or falls in unemployment. The "logic" is that if
increases are followed by falls in unemployment then that
proves that increasing the minimum wage doesn't destroy
jobs. That reminds me of the argument that the Sun goes
around the Earth because that is how it looks to be. How
would it look if the Earth went around the Sun? Just the same.

I think that asking why people don't change their minds is
asking the wrong question. Empirical work in economics is
incredibly difficult. If you make drastic simplifying
assumptions, such as supposing that politicians are clueless
and bugger about with the minimum wage at random, you will
not persuade anyone. If you make realistic assumptions you
can easily end up with opposing theories leading to the same
predictions, so you cannot tell them apart. You should not
be expecting people to change their minds. The interesting
question is where do peoples ideological commitments come
from in the first place? People are not getting their
commitments from evidence, economics is too difficult for

hhoran writes:

The answer, as always, is power and status. Do you have a situation where people of divergent views and background share a strong interest in solving an important problem, or does enhancement of personal/group interest overwhelm the abstract pursuit of truth?

In academia, the biases of professors protecting their status, or folks from faction X fighting faction Y are well known. There is also the issue of whether the accepted wisdom supports entrenched economic interests. Economic criticism of rent, wage or price controls carried the day, but really rich/powerful interests had no interest in keeping those controls. But similar economic criticism of intensive zoning laws, strict command-and-control highway/parking spending and rules goes nowhere as they are each protected by powerful vested interests.

The overwhelming use of regressions isn't at the Solow level--it is much more pedestrian arguments in legal fights over economic disputes. Thus you have the unfortunate confluence of a context where bad analysis is welcome because nobody cares about abstract analytical standards and the ridiculous academic cultural bias favoring quasi-math methods over any other type of evidence. Your original question reflects that bias--you clearly presupposed that regressions, per se, were higher quality/more objective than alternate forms of evidence. If the objective is "winning" an argument, or protecting powerful interests, or enhancing personal/group status, regressions can be debased just as easily as any other form of evidence.

steve writes:

I think people change their minds all the time.
The basic act of studying a subject in detail results in multiple previous beliefs being changed. Of course, the rate of change slows down once one becomes deeply familiar with the material. While experts occasionally change their minds their just aren't that many occassions during an experts carreer that result in unequivacal proofs occurring. For example, whether or not a nuclear bomb could be built was genuinely and legitimatly questioned at one time. Lets say the proof was convincing.

steve writes:

When it comes to economics, proofs on the order of a nuclear bomb are hard to come by. While it may seem obvious to an economist that taking a certain action will improve the economy, there are always losers and winners. Naturally, the losers don't see those actions as convincing from their point of view.

I think to get a mass change of opinion on the order of price controls one needs the argument to to be clearly articulated to the population in genral and plainly true for a significant percentage of them.

Unfortunately, this is why I believe that socialism continues to reoccur with such frequency. It promises a better life for significant numbers of people. When one raids the seed corn so to speak, this promise can be kept for a period of time before it proves false and is abandoned once again.

Jacob writes:


Your post reminded me of a fellow I ran into the other day. He is a Marxist and strongly believes in the labor theory of value. He was surprisingly intelligent for espousing such a belief.

As a quick comment, I think people are just poor at convincing others. When trying to convince someone else of a fact, we tend to explain a concept how we understand it, not how the person we're explaining it to understands it. Those who know me well also know how to convince me of their position.


steve writes:

This is speculation on my part but I think marxists ultimately base their beliefs on what I think of as the difference between price and value. Water is essential to life, gold is not yet the price of gold is higher while the value of water is higher.

Ultimately I think a marxist bases their belief in economics on the idea that all human life has equal value. I think this is true. They then claim that it logically follows that the price of all human effort should be equal. Unfortunately, this conclusion is disasterous for the purposes of efficiently organizing production.

Adam writes:

I do think some people are open to evidence and argument. Sorting through evidence, assessing it's methodological status and making reasoned judgments takes work. As Daniel Kahneman noted in a 2004 AER article, relatively few people put in the work to make reasoned judgments. In place of reason, people tend to substitute learned, automated patterns of thought.

What are my bona fides in terms of changing my mind? Here are some few changes over the course of years:

1. Secular humanist to born-again Christian.
2. Socialist/statist to libertarian.
3. Utilitarian to ethical libertarian

and many other lesser decisions.

Tracy W writes:

I wish the people who write popular psychology books would devote more time to explaining when and why people do change their minds. I read quite a few of these books and they tend to be spend a lot of time listing cognitive biases, like confirmation bias, that make us resistant to changing our minds. But any look at the polling data or election results over a period of time shows that some people do change their minds from time to time about which politician to vote for, and I haven't run across a book yet that goes into detail as to why mind changes do sometimes happen even so.

George writes:

fundamentalist wrote:

If Congress and the presidency are held by opposite parties, gridlock will result.

And yet you voted for Obama. Exactly how many seats did you expect the Republicans to pick up in Congress? In a two-party democracy, you have to pick the lesser of two evils. Period.

steve wrote:

When it comes to economics, proofs on the order of a nuclear bomb are hard to come by.

The stimulus is Fat Man and Cap-and-Trade is Little Boy. (That makes the looming but not certain takeover of the medical system Thin Man.)
I hope there's something left of our economy to study.

The Cupboard Is Bare writes:

People who are most likely to change their minds are those whose primary concern is to know the truth.

People who are least likely to change their minds are those who have win/lose mentalities. The weaker their position becomes, the more likely they are to dig in.

steve writes:

George writes:

"The stimulus is Fat Man and Cap-and-Trade is Little Boy."

Good point. Unfortunately, I think the cause and effect will be spread out over a greater period of time and not evenly distributed making it necessary to tease out a number of other potential influences from the data. Thus, the cause and effect will not be so obvious to a lot of people.

Mr. Econotarian writes:

"And there's the (current) 800-pound gorilla of false beliefs--that rich countries with lower taxes grow faster than their higher-taxing counterparts."

However the data is clear that higher levels of Economic Freedom are associated with higher GDPs per capita;

Economic Freedom measures do include tax levels, but we should keep in mind there are many other things that make up economic freedom.

The Netherlands is my favorite example of a country with higher personal taxes than the U.S, but lower corporate tax rates, greater investment freedom, greater freedom of the finance industry, and less corruption.

Outside of the traditional economic freedom measures, the Netherlands has less drug prohibition, and less government control of schools than the U.S.

pk writes:

If two entities with opposing views have a dialog or series of dialogs, and neither is ever willing to be wrong, neither can advance beyond their starting position.

The key, I think, to winning these arguments is by willingness to be wrong, which *can* beget a willingness by other parties to be wrong.

Of course, there is willingness at the margins and willingness at the core. Changing a core belief can have a massive domino impact, and for some, that may just be too uncomfortable to bear. "What if I've been wrong about everything? What an idiot I have been!"

At that moment is the best of learning -- a brand new perspective gained by looking at a problem from an alternative point of view.

So I think changing minds means investing not so much in the underlying data/logic, but a bit more in influence and communication strategies, and there is a great deal of research in this area. The problem here is the slippery slope to manipulation rather than persuasion. The statist does not mind manipulation -- he thrives on it. The libertarian is repulsed by manipulation and control, and rejects its use. That puts the libertarian at a distinct disadvantage for winning political arguments.

Comments for this entry have been closed
Return to top