Did this get cut off early? I see section 1 but no others, and it seems to end rather abruptly for you.

Expand full comment

I'm inclined to blame / credit the media more than anything. They tend to celebrate liberal victories, bring awareness to their positive aspects, and decry conservative victories shedding light on their negative aspects. This may be naive, but it seems like a simple explanation thay fits the data?

Expand full comment

"Relative to the current status quo, are you pro-choice or pro-life" is a question that would produce the observed pattern, and it's plausible that many or most people would interpret the question this way, consciously or unconsciously.

Expand full comment

I think it's the media is liberal, and they do $(a few different things) to convince everyone the conservatives are evil, and everyone agrees for a while. I suppose if I were serious about proving/disproving my hunch (it is, alas, only a hunch) I would go back to the 1990's, or some other time when one could argue the media had a conservative bias, and find similar events during that time and see if there was a conservative-direction backlash.

Maybe some things during Bill Clinton's reign?

Expand full comment

Theory 1: Individual liberal policies poll much better than liberal politicians as a whole. Events that crystalize politics around specific tangible policies may end up favoring liberals.

Theory 2: Maybe the "shy Tory" effect is actually stronger for liberals, or at least liberal policies, than we realize. Perhaps there's a significant cohort of people don't want to be one of those bleeding heart liberals or blue-haired feminists attitudinally but when it comes down to actual women being forced to carry unwanted babies they line up on the liberal side.

Expand full comment

I think the sample size is too small to draw any meaningful conclusions. If this was a systematic pattern with decades of history I would be much more inclined to try to draw some conclusions, but otherwise I think each backlash is specific to the policy in question.

It may also be part selection bias: the last congressional session passed hundreds of bills, most of them didn't have backlash, so the baseline is to expect no shift in public opinion over the implementation of an idea (maybe unless the implementation is uniquely bad, like how brexit turned into a giant mess that left nobody happy).

Expand full comment

Number of undecideds and level of salience is probably a factor. Pre Dobbs a lot of Americans who never really thought about abortion wouldn't report strong feelings on it, but when Dobbs passed it rapidly increased in salience, and so reported pro choice.

In general a big salient event probably pushes undecideds towards the most popular side, since they are likely to break down the same way as the rest of the population when exposed to more information. In the gay marriage case majority support was already for gay marriage, so undecideds mostly tipped that way.

Expand full comment

I think Richard Hanania's observation that liberals are more tuned-in to politics and care more about politics than conservatives do can provide some explanation.

Expand full comment

Political victories which do not meet the approval of the hegemonic class.

Expand full comment
Jan 19, 2023·edited Jan 19, 2023

I'd theorize this happens when people think they can hold an opinion that will never come into effect. It probably felt nice for some people to vote for Trump in a rage against the machine, but when he came into power, they had to live with his actions, not just the message their ballot signaled. Similar story for abortion: I was talking to one of my closest friends a few days before the draft leaked; he said he wished abortion were illegal but "knew" the Court would never budge, and so it remained an abstract, lukewarm conviction. After the decision, the ensuing chaos, and a billion testimonies, he's now pro-choice. Maybe the decision encouraged people to proselytize outside their normal bubbles and punctured echo chambers; perhaps more people who knew the arguments and weighed them had their scales tipped. Not sure of any other examples to test this hypothesis against, but I'd love to hear some. (Edited for clarity and to remove Markdown syntax—sorry, first post <3)

Expand full comment

It's because when people are given conservative policies they realize those policies are terrible (abortion bans result in horrible criminalization regimes) whereas when they are given liberal policies they realize those policies are fine (gay marriage does not in fact undermine civilization and does not affect anyone's life negatively).

Expand full comment

I'd say the pattern is just that if people like how the change worked out, there's no backlash, whereas if it causes obvious trouble, there's a backlash. Dobbs->10yo victim->backlash. Gay marriage & Obamacare->happy gay couples & more folks insured->no backlash.

You could argue that this is about media bias, but I'd challenge that: if there were obvious problems after the gay marriage or Obamacare changes, CNN and Fox would both have been all over them.

You could say the public is biased in favor of policies that cause no obvious trouble (most people don't care much about curbing oil use until climate change starts affecting them personally)... well, of course they are.

Expand full comment

When the court makes a decision that catches up to where the public has been moving for years (e.g., Obergefell) it doesn't cause a backlash, but when the court makes a big decision that changes the status quo on an issue where the public has been stagnant for years (e.g., Dobbs) it causes a backlash.

You claim that the Obamacare chart shows no change, but to me it really looks like there's a higher pre-2008 plateau and then a lower post-2008 level that gradually starts moving upward again around 2014 (when the law starts coming into effect and Republicans start campaigning on taking it away).

Expand full comment
Jan 19, 2023·edited Jan 19, 2023

2022 May 2-22: pro-choice = 55% pro-life = 39%

2021 May 3-18: pro-choice = 49% pro-life = 47%

Legal under any, Legal under most, Legal only in a few, Illegal in all, No opinion

2022 May 2-22 any=35%, most=18%, few=32%, never=13%, no opion=2%

2021 May 3-18 any=32%, most=13%, few=33%, never=19%, no opion=3%

Expand full comment
Jan 19, 2023·edited Jan 19, 2023

The trend of support for a policy seems like a mix of two different trends- how much people support a particular outcome, and how much they think the policy supports that outcome.

In the case of gay marriage, there's growing support for the outcome of gay people getting married, and the policy of legalizing it clearly supported that outcome both before and after Obergefell, so the trend just represents the cultural shift.

In the case of abortion, however, the median person probably supports something like what you see in most European countries- protection for early term abortions and consistent outlawing of late term abortions. During the Roe v. Wade era, "pro-life" policy moved things closer to that outcome, while after its repeal, "pro-choice" policy does the same thing, so the trend looks like a sudden reversal even when the outcomes people want haven't shifted much. It's not that people always want a moderate outcome and will react thermostatically to policy changes- it's just the new policies may or may not overshoot the desired outcome, and that desire also shifts over time independently of policy.

I'll bet a lot of the Trump backlash was a similar sort of thing.

Expand full comment

For some of these examples, I think a relevant factor is "beliefs as attire". People sometimes claim to believe "nobody should ever murder an unborn child" or "international trade is bad for the economy" when they mean to express "boo feminism" or "boo China", but when these become actual policy options they're forced to reconsider whether they take these beliefs literally or only seriously. In contrast, "people should be able to marry same-sex partners if they want to" is plausibly straightforward enough to be mostly believed or rejected on the merits, and nobody is shocked when this policy is implemented and works pretty much as expected.

Expand full comment

The choice of examples of liberal "wins" not creating political backlash against liberalism is insufficiently precise and elides some pretty important individual features, particularly those of the Obergefell decision, that adequately explain why these instances don't disprove (or even counsel against) the thermostatic effect theory.

Let's be more specific, with an actual example. Indeed, if backlash only happened when conservatives got political wins, one would expect public support for the death penalty to have stayed the same or come down during the 1970's moratorium period. But this is not what happened. Gallup historical polling data (available here: https://news.gallup.com/poll/1606/death-penalty.aspx) shows a major and immediate increase in support for the death penalty following the landmark liberal Furman decision by the Supreme Court (which banned every existing death penalty regime), from below 50% to over 65% in less than 3 years. This is because, just as with the Dobbs decision, the Court weighed in on a contested political topic and pulled hard towards one direction, satisfying one group but creating comparatively greater anger and resentment on the other side.

This latter part happened to some extent in 2015 with Obergefell as well. However, it was immediately moderated (and eventually outweighed) by the fact that the decision was seen as the culmination of a decades-long movement to make homosexuality acceptable in "normie" culture. As such, lots of previously closeted gay people felt safe to come out to formerly homophobic (or potentially homophobic) family members, friends, or acquaintances.

I am convinced that the massive increase in support for LGB rights (and the GOP's rapid removal of anti-gay rhetoric from its platforms and public discussion in the 2nd half of the 2010s) was primarily because legal, political, and cultural changes caused more gay people to come out, and therefore more heterosexual people realized that others they knew quite well and got along with great were actually homosexual. It's a meme to say that "I have one black friend, how can I be racist?", but if SO MANY people all around you whom you had previously thought of as fundamentally good actually turn out to be gay, that makes it super difficult to keep believing that gay people are evil (or thay being gay is inherently bad). Homosexuality simply became normalized through exposure.

This is what makes it a bad case study for political wins and the thermostatic effect, because the real "win" wasn't making gay marriage legal (if this was the only thing that has happened, you WOULD have seen public opinion shift towards the right), but rather creating an environment that allowed centrists (and even socially slightly-right-of-center people) to realize that those who are homosexual or bisexual are just... normal people, and that you shouldn't get in the way of normal people living their lives and marrying each other.

Expand full comment

As a liberal example, I would expect the recent rise of BLM to result in a strong downturn in support for racial-justice causes. Anyone have any data on that?

Expand full comment

I think it's largely tied in to how high a probability you assign to the new law/policy either affecting 'you', directly, or (to a lesser extent) producing what you think are fundamentally unfair outcomes. The abortion case is a great example of this: Most people can imagine a scenario where a young woman they know (daughter, grand daughter, niece, etc) might really need to get an abortion, or they can at least imagine a case where a young woman to whom they are generally sympathetic would want/need one, so you see a backlash.

You don't see a backlash with issues like gay marriage because for all the chest-beating about 'the decline of Christian values, blah blah decadence blah blah', it isn't adversely affecting anyone you know or whom you would be sympathetic. So two gay guys get married, and move in down the block: That's....not really changing anything? You might be a little grossed out if you see them holding hands when they're out for a walk, but on the other hand you probably know first-hand some gay people (even if you live in a small town in the middle of nowhere), and likely have some sympathy for them even if you don't agree with their lifestyle.

Looked at this way, you can see some possible laws/policies originating from the left that could (or have) create(d) a significant backlash. These include:

1) Banning meat consumption.

2) Mandating 'CRT' classes in schools.

3) Allowing transgender boys to play girls' sports.

4) Using tax dollars for reparations (this is a big one. If you want to see us toss race relations 80 years into the past, give everyone who identifies as black 'x million' and watch what happens.)

5) Vaccine mandates/passports

6) Open borders

7) Total repeal of drug laws

By contrast, here are some policies/laws you can see originating from the left that would not (or have not) have/had this effect:

1) 'Common sense' gun control laws (no sales to felons, waiting periods, etc).

2) Massive infrastructure spending

3) Higher taxes on rich people

4) Stronger anti-trust legislation

5) Increases in social security payments

Expand full comment

I'm not entirely sure about the claim of "no effect" for the Affordable Care Act ("Obamacare"), which existed in some form since at least 2009 and was something then-Senator Obama campaigned on in 2008, and it does seem like things trend downward for Republicans/Independents on that graph. (It doesn't clearly start in 2008, but that's kind of a "Does reality drive straight lines...?" conundrum.)

I would also question whether the abortion ruling can truly be said to have no personality component - there was plenty of coverage related to President Trump, and related to the series of events that led to the current composition of the Supreme Court, e.g. https://www.cnbc.com/2022/06/24/roe-v-wade-decision-trump-takes-credit-for-supreme-court-abortion-ruling.html

Expand full comment

The following would be consistent with your data: many people don't think much about these issues and have opinions acquired from parents or neighbors or party. Then the issue becomes salient (per David Shor) so they think about it and most decide they agree with the liberals. To test this, are there cars where an issue became highly salient and the law did *not* change but opinions did?

Expand full comment

I think the sample size it to small to draw meaningful conclusions, but pontificating in the absence of evidence is what the internet is best at, so...

It's possible that views are biased away from perceived extreme positions. We might be seeing opinions trying to regress toward the mean. So inasmuch as popular conception was that abortion was available and practiced with zero restrictions, many people hoped for something *slightly* more restrictive. With the Dobbs ruling, what we got, at least in some states, was a real extreme in the other direction, and so many people hope for something *less* restrictive.

Obgerfell didn't generally change the status quo much for most people. It didn't create a new extreme, it protected the norm in a number of states without forcing anything new in others.

On trade, many people might have felt that China had some sort of upper hand before Trump's edicts took effect, and wanted more consideration, but then felt that his edicts went much too far, and want something closer to the way things were before.

There was enough backlash to the ACA to help Trump get elected, so that's not nothing, but overall people were still able to get health insurance, and in fact more people were, and costs didn't skyrocket for most people, so there wasn't the same "avoid extremes" effect.

As a working theory, it fits those examples, but more examples might break it.

Expand full comment

I'd be interested to see an analysis of the Trump tax cuts. I'd expect that to be a low-perceived-impact conservative victory which wouldn't have a backlash, but I don't actually know.

Expand full comment

> Or maybe it’s because liberalism is “on the right side of history”, ie the direction things naturally tend, so nobody finds liberal victories surprising, whereas everyone finds conservative victories surprising and feels like they need to react against them.

This intuitively more plausible to me, but not in the context of whether these policies are ‘right’ or not, but rather the trend of public opinion.

Oh the graph showing the support for gay marriage, we can see that it has a clear upward trend—i.e., overall support for it goes up overtime.

Maybe these victories tend to accelerate public opinion in a certain direction if they’re particularly controversial.

The average person might read the Dobbs decision, and already slowly becoming more pro-choice, just becomes more so.

This could be because of the media, which does happen to be more left-wing, as Scott points out. These controversial decisions make political commentators rage, which then gets transmitted to NYT, CNN, FOX, etc., and then to other media outlets.

Perhaps the rate of change in these opinions is proportional to the overall exposure people have to them from the news, whereby a news cycle dominated by it increases the total amount, thereby changing opinions.

It would be really interesting for someone to model this, maybe by plotting keyword frequencies in the NYT for example versus the change in opinions for controversial issues.

Expand full comment

>My impression of the post-Dobbs debate on abortion was that it centered on an incredibly sympathetic 10-year-old rape victim from Ohio who really needed an abortion but was having trouble getting one

I have always wondered why anti-abortionists aren't doing anything about this. It's *incredibly* easy to paint abortion in bad light, bring out just one of those using it as a "A type of birth control that involves surgery", all those promiscuous $CENSORED, the sheer amount of debauchery and sluttiness their life consists of, that without abortion they would have thought 20 times before engaging in.

And then there is all those who are like "I was doomed by my unexpected pregnancy to be a basic ass stay-at-home mother but then I killed my future child and now I'm a successful wage slave at $BIG_CORP, and managing other wage slaves even. Please don't take away our right to kill the unborn uwu" in the newspapers and on social media. They talk about this rather proudly too, so it's not like there is any extra cajoling or deceit needed to bring it out of them.

In contrast to this, a single docuementary about the sheer magic of a mini-person rising like a phoenix from a single fertilized cell is enough to maybe shift **at least** 15% to 20% of the least-fervant pro-abortion base. It's an Epic, really, a compressed retelling of the story of Evolution of multi-cellular and primate life, on the spatial scale of a single organism and the temporal scale of an evolutionary nanosecond. Bonus points for showing the cute babies that would-have-been-aborted fetuses grew into, and interviews with the fortunate mother who changed her mind.

If somebody told me about the abortion debate for the first time in a vacuum, I would have expected that it's the *pro-abortion* side that desperately needs good faith and a devil's advocate. That reality is the exact opposite is so surprising to me. From a purely Machiavellian perspective, ignoring all utility calculations about who's "really" right, the anti-abortion side is severely under-using some of the rhetorical and propagandist assets at their hand, chief among which is an instinctive, millions-years-ancient, literally life-or-death love for babies and their survival ingrained in nearly all humans, including violently misanthropic ones like me.

Expand full comment
Jan 19, 2023·edited Jan 19, 2023

The Obamacare graph pretty clearly does show a swing against government healthcare coinciding with Obama pushing for Obamacare. It’s also a bit misleading, because it disaggregates responses into three partisan categories, and so ignores people moving between the categories. The aggregated graph shows the effect more clearly (second graph): https://news.gallup.com/poll/4708/healthcare-system.aspx

ie the least pro-Government healthcare Democrats becoming Republicans (a) is a thing that happened, and (b) would probably increase how pro-Government healthcare both groups are.

But even ignoring that… looking at a 20 point fall among independents and saying “no effect” is definitely wrong.

Expand full comment

Or maybe because the left is better at coordinating social meanness and ostracism against conservative political ideas and those who support them than vice versa.

Expand full comment

While I do believe media influence affects degree of the swing, I think there is merit to the thermostat concept. I just think people have a skewed idea of what the thermostat is measuring.

Complex trade policy? Mot people don't have a thermostat for that. They don't know what it is, how it works, or care to. But the media can hammer it hard and swap responses. Also keep in mind if your methodology for the poll is poor, it'll mask the don't care people. So with something like that you may have the number one answer be to hang up the phone and not waist their time. Normally this would be everyone without an opinion on trade policy so it works out. However if the media is hammering the issue, you may have a skewed population self selecting into the results rather than out because orange man bad per media.

Then with abortion, while the fact is most people are someplace between all the abortion and none of the abortion for an actual opinion, you are likely measuring which team you want to be associated with. Republicans are not uniformly driven on religious issues. A lot of republicans who are republicans on different issues were fine with it is an issue because they assumed it was a car the dog was never going to catch. The dog caught the car. Positions are being recalibrated. Some of the pro lifers are switching to an even more pro life stance and want it banned everywhere. Plenty of republicans would prefer not to hitch their horse to that wagon.

Healthcare? Perhaps it measured if people thought they were getting better health care from a fiscal point of view. People who saw the ACA as a huge failure for people who wanted single payer health care probably saw it as not a big improvement. Independents probably did the thing that independents do and decided that in the run up to enactment it was a political clown show they disapprove of. Republicans probably disapproved of it for a bunch of reasons. But then it got passed and lots of people got bought off with the subsidies and artificially low rates. Then premiums went up. Then the subsidies came to an end, so the brief uptick probably came to an end. Then post 2015 they start skewing with how much you hate Trump.

Gay marriage? The approval across the board is generally trending upwards and has been for years. I suspect the slope on the republican side will trend upward more sharply moving forward because for them you are likely measuring, in many cases, the calibration on their sense of moral peril. The "gay agenda" has now morphed into the "trans-agenda" and a lot of people concerned with that see the LGB portion of that as potential allies given the current direction of policy. There will also be a faction of Rs that just don't want to lose, don't care if gay people are married, and want it covered by defensible law rather than be beaten with a club that says "gay marriage and interracial marriage are next" post Dobbs.

Expand full comment

The health care one seems somewhat thermostatic if you base it on when the opposition starts scaremongering and not when legislation actually passes. Democrats, Independents and Republicans each become less approving of Government Healthcare starting in 2008 when Obamacare is being passed and Death Panels were all over talk radio. Then D'S & I's become more approving of it when Obamacare is threatened with repeal in 2016.

Expand full comment

It looks to me like there was a ~5% dip after Obamacare was passed, for a couple of years.

Expand full comment

Consider affirmative action in CA. Didn't that lash back pretty hard after the leg tried to institute it?

Expand full comment

It seems worth noting that "backlash" in the form of responses to public opinion polls do not necessarily correspond to actual substantive change.

In the case of public trade, for example, Biden has kept Trump's tariffs.

Expand full comment

I think some of it depends on really where the cultural norm sits for the politically disinterested. If something steps outside of their comfort window of norms, just enough, then they become engaged, usually in a backlashy kind of way. I know we've seen super high turnout post 2016, so maybe that supports this idea. Or maybe I'm just making crap up, dunno. I'm skeptical of the "liberal media wins" narrative because it seems too easy, but I live in the deep south so everytime I see a new channel on, its Fox or Newsmax declaring a woke crisis, so I probably get exposed to a whole lot more conservative doom porn than the avg ACX reader

Expand full comment

I hope it doesn't apply to Bruen, and I expect it won't. Primarily because I think most people don't even know it happened, let alone what it was.

Expand full comment

Could the change in abortion support reflect the number of people who mistakenly think the Supreme Court made abortion illegal?

Expand full comment

I think that the overturning of Roe v Wade* really changed what it meant to be pro-life. Whereas it used to be “less supportive of abortion than Roe v Wade” it became “less supportive of abortion than the laws in my state” and those state laws changed very fast.

Something “against gay marriage” doesn’t have as much middle ground I think, so changes on it doesn’t change the meaning of “against gay marriage”

*technically it was Casey, not Roe v Wade that was overturned, but I’m choosing to be slightly inaccurate to stay with the standard way of denoting the abortion cases

Expand full comment

It may be that people want some medium level of sexual liberation/libertinism, rather than some amount of gay marriage. The question shouldn't be whether opinion turned against gay marriage after Obergefell, but whether attitudes toward LGBT+ normalization efforts noticeably cooled. I'm not aware of a rigorous body of evidence suggesting that they did, but it seems plausible.

Expand full comment

Right side of history ftw

Expand full comment

I agree this isn't enough data, but some ideas that occur to me:

Idea #1: Maybe you just happened to pick a rare thermostatic issue for the conservative side? The idea that lots of people want a medium amount of abortion seems intuitively plausible to me--there's a wide gap between a maximally-sympathetic abortion and a maximally-unsympathetic abortion, so it's easy for me to imagine someone wanting to draw a line somewhere between them. I have a much harder time imagining that people want a medium amount of gay marriage; examples of gay marriage strike me as having a much narrower range of sympathetic-ness.

Idea #2: I once read about a study (sorry, I don't have a link) that asked politicians to predict public opinion. They found that liberal politicians believed the public was more liberal and conservative politicians believed the public was more conservative--so far, so boring--but that when they polled the public directly, the public was even more liberal than the liberal politicians thought they were. (The study suggested that maybe corporate lobbying skews all politicians' perceptions of public sentiment in the conservative direction.)

If true, that could perhaps be a reason that the public tends to be angrier about conservative moves than liberal moves--the liberal moves end up having more support than the politicians expected, and the conservative moves less than the politicians expected.

Expand full comment

I think the difference here is that the question is binary (for or against abortion), but the state fights have really been a matter of degrees (At how many weeks pregnant should abortion be illegal? Should there be exceptions for rape and incest?). Thus it's possible the issue has been reframed for those people in the middle - if you thought abortion should be illegal after the first trimester, but your state just voted on a law to ban it at 10 weeks, you might have initially believed you were against abortion, but now that you're opposed to the proposed law, you would describe yourself as being pro choice.

Expand full comment

Looking for a "backlash" effect is fun, but not good science.

If we were to slow the reasoning down, here's what it would look like:

(1) there's a metaphor we sometimes use in conversation: "backlash". The metaphor says that people are sort of like rubber bands, and when you pull them too hard in one direction, they tend to "snap back" (or something). We apply this metaphor in certain situations, as a vivid way to tell a story, or help make sense, of something that happened.

Then there are two options:

(2A) We believe that "backlash" is not just a figure of speech, but that it actually refers to a stable explanatory object, that science can investigate. It's a process that takes place in different situations, has causal powers, etc.

(2B) We're not going to be scientific about it — we're not trying to fit this event into a larger account of how humans reason about politics. Instead, we'd like to use the metaphor "backlash" in this particular discussion, and we want to put a number on that metaphor.

There's not really a good reason to believe (2A). We don't have (for example) a good model of cognition that would predict backlash as consequence, making it a repeatable phenomenon in the individual. (I'm not saying it's impossible to construct — just that one hasn't been suggested here or elsewhere, as far as I know; it would be a graduate-level homework to build a Bayesian model, for example.) Furthermore, without a good model, we can't say in which circumstances it should or should not appear, and so we're already in danger of p-hacking.

Meanwhile, (2B) is scientism. You can put a number on the metaphor, and even a p-value, but it doesn't actually mean anything. You're not talking about scientific objects (for example), or appealing to a scientific explanation. You're using statistics to boost your metaphor above someone else's, but really all you're saying is "here is a pattern, that is statistically significant, and I am going to tell the following story about the pattern — it's 'backlash'."

Expand full comment

One problem is that the graphs you're looking at only include "pro" and "anti". How many of those answers come from people who are really thinking "I've never really thought about it"? If there are large numbers of people answering essentially arbitrarily, what appears to be a controversial issue could actually just be a non-issue. Free trade seems like the obvious example -- it's hard to imagine average voters paying the slightest bit of attention to trade policy until there's a highly public change to it. In this case I would buy that the scale and direction of the change is due to Trump's unpopularity. Even some people who voted for him saw him as untrustworthy, so any major policy associated with him is going to have an uphill battle when it comes to public opinion. You could also look at the pro-life vs pro-choice graph: there's a big spike toward pro-choice at the start of the graph as well. Was abortion really more popular in the early 90s than it was in the 2000s, or was it just more in the public eye due to Casey reaffirming Roe in 1992?

The liberal vs conservative split is probably something of a coincidence. The conventional wisdom is that you can expand people's rights and benefits without backlash, but you can't take them away. Liberal policies tend to expand rights, conservative policies tend to take them away. People either don't notice or are actively happy about their rights being expanded, and they either don't notice or are actively angry about their rights being taken away. If a liberal policy were passed that hurt people, it would probably be unpopular.

Expand full comment

Why not assume this is entirely due to centrists? Let's say there's a government policy which can be on a scale from 1 to 10. People describe themselves as pro-increase and pro-decrease. If I want the number to be at a 5, and it's currently at a 6, I'll tentatively describe myself as pro-decrease. As soon as the number moves to 4 then I'll flip to pro-increase.

In other words, maybe there's no backlash at all, and people in the center just went from wanting slightly less abortion to slightly more abortion when the amount of abortion decreased significantly.

Expand full comment

Re Obamacare, (a) it does seem to have declined in the 2008-2010 period (which makes sense, it was a longer discussion than the dobbs decision and didn't come as a shock the way Trump's election win did), and (b) does seem to have suffered backlash by other metrics (Democrats lost the house in 2010 in a historic backlash, probably in large part due to thermostatic effects from obamacare). So I'm not sure it's strong counter-evidence.

Expand full comment

One obvious distinction is that the overturning of Roe vs Wade came from the courts, not from politicians. Politicians are less likely to do unpopular things than courts are.

Obviously Obergefell was a court decision too, but I think my model is that courts do a mix of popular and unpopular things whereas politicians very heavily towards things they think are popular, so Dobbs, Obergefell and Obamacare are all "business as usual" that fit the trend I'm positing, whereas the swing in favour of international trade is an outlier that challenges it.

Possibly international trade was previously an obscure issue most people didn't have strong opinions on, and it suddenly becoming associated with an unpopular politician made it a proxy for "do you like Trump"? In this model the trend against it is nothing to do with whether Trump won a victory or not, the causal thing was him becoming associated with the cause.

Expand full comment

Am I the only one who fails to see any obvious change in some of those three plots?

Expand full comment

I've often wondered whether certain beliefs function as predictions and other ones as markers of markers of tribal affiliation. If so, you'd expect "predictive" beliefs to be responsive to new evidence and "tribal beliefs" to not be responsive to presentation.

So if Trump champions protectionism in trade it might drive people's opinion on trade policy based on how much they like Trump. But if people predict that restricting abortion will cause people to be more careful about pregnancy, than stories of people who acted responsibly and still need an abortion would be likely to change their mind.

This is obviously somewhat post-hoc - it doesn't allow predictions of which positions will suffer from backlash, but it does suggest that political positions that are more rapidly reflected in outcomes are more likely to generate support/backlash depending on the outcome. That seem a bit like what is observed when it comes to things like gas prices - if Fox news says "Biden's anti-oil policies are pushing up the price of gas" then Biden's approval rating falls while gas is high.

Expand full comment

See your penultimate sentence: I think you can explain your observations with selective data and a bit of luck.

For the abortion issue, if you look at pew data ( https://www.pewresearch.org/religion/fact-sheet/public-opinion-on-abortion/ ) what you see is a long term trend of increasing pro-choice with a change from 21-22 that’s comparable with previous years.

If you look even at other Gallup data from the same page ( https://news.gallup.com/poll/1576/abortion.aspx ) the trend looks much less reactionary. Note for example that the figure in the post was for “asked after question on legality of abortion” whereas if you look at the table with “NOT asked after question on legality” we see a much smaller shift in line with fluctuations of previous years. One thing that likely changed after the ruling is people’s awareness of the law even in unaffected states.

I notice for the Trump issues you used figures from a number of different sources, which raises some flags for me about similar selectivity as the abortion data. Even if it’s a matter of honest search convenience, it makes the results a bit sketchy.

As you point out in the penultimate sentence, a handful of datapoints is selective even if the trends are consistent across polls, and your n~5 is really only 2 examples of the phenomenon. We can tell a story about reactionary responses to Trump without drawing general conclusions that the public is reactionary.

Expand full comment

I cannot express the breadth or depth of my indifference to the reaction against Dobbs. To be clear, I'm an atheist, but very anti-abortion; if that's confusing to you, it shouldn't be. Atheists don't believe in god, they believe in humanity (or at least I do).

Abortion kills humans. Whatever you think of the debate about when life begins or when that life is human or any of those, I don't think they matter. A human fetus or embryo will, in almost all cases, mature into a human and be born. This includes the fact that miscarriages and natural abortions are reasonably common (if you try to lawyer me on this point, I will ignore you. You will not convince me). It is not tenable to claim that, by intervening at any particular point in this process where we have decided no morality attaches, we also free ourselves of the moral burden of preventing what comes next, which is to say, the birth of a human.

We live in a society that can barely bring itself to put to death multiple murderers, rape-murderers, and the like. And yet a significant fraction are okay with annihilating the most helpless and innocent people? It's close to the definition of insanity, as far as I'm concerned.

Roe v. Wade is one of the most abominable decisions of any court. It offends my morality (as just described), and it offends my professionalism as an attorney. No one, not a single commentator, attorney, or jurist, seriously defends the reasoning of Roe v. Wade, because it has none. It was made up from whole cloth. The concept of the trimester was literally invented by Blackmun.

Just on this basis alone, Dobbs is worthwhile and correct. Egregiously wrong decisions should not stand if there is any opportunity to overturn them. The second-worst decision ever, Korematsu v. United States, remains good law and will remain good law unless and until someone tries interning Americans again. This is in the nature of our appellate system.

Even if Dobbs leads to more abortion rather than less, which I sincerely doubt, it will have been worthwhile. Now, the people and their representatives will decide. Not every right needs to be constitutional. If the people's decision is contrary to my morality, I have means of fighting back. By contrast, there is no gainsaying the Supreme Court.

Last and certainly least, the conservative legal movement was founded to oppose Roe. If the Court had blinked this time, after so much work and so many disappointments like Casey, it would have discredited the entire movement. I don't agree with every or probably even most conservative legal thinkers, but I do think the movement has done real good for real people, and is a valuable force that has corrected the lawlessness of the post-war legal profession and courts.

Expand full comment

This is one obvious theory:

A. None of the policy victories shifted anyone's opinion.

B. The shift in reported opinion reflects people's stated values moving into line with (or toward) their actual values.

It's pretty easy to explain why a policy victory would cause people's stated values to move from a misrepresentation in the direction of honesty - if you stand by whatever you were saying before, the policy you purport to like might gather momentum and score even more victories. If you didn't like the first one, that would be bad.

It's also easy to explain why people might systematically misrepresent their own values; in fact, that's been a recent focus of activity because the phenomenon appears to be so common.

However, this theory gives you nothing in terms of your ability to predict which victories will cause backlashes.

I don't intend to claim that this theory is correct as to the two opinion shifts you discuss, but it doesn't have trouble explaining (1) why they happened and (2) why other policy shifts don't cause similar opinion shifts.

Expand full comment

I think your analysis runs pretty far ahead of your data. For example, you say there's a clear backlash against Dobbs when you see a 6% shift in opinion (2nd graph), but you say nothing happened after Obergefell despite what looks like a 2-5% dip in support for gay marriage in 2015 (4th graph), depending on party. (Part of why it's hard to compare these graphs is that the y axis for the 2nd graph is magnified by ~2 compared to the 4th.) I'd be reluctant to draw sweeping conclusions from observing a 6% versus a 2-5% change in a noisy small sample. You might also be bemused by the fact that the Dobbs data stops while the issue is still very recent (2022), while the Obergefell data go on for another 7 years. Perhaps the Dobbs change is as ephemeral as the Obergefell data -- or maybe it represents a lasting shift. We'll have to wait to see.

I also think the link between the ACA and "government responsibility for healthcare" is pretty dubious. The ACA very carefully and loudly did *not* include a government-run option, because polls at the time suggested it would sink it politically (as well be completely unaffordable). The omission was widely decried at the time by people who like government-run healthcare, so it would be difficult for anyone paying any attention at the time to think the passage of the ACA represented the triumph of government-run healthcare. It was, if anything, on the contrary, good evidence that Americans wanted no such thing, even if they agreed healthcare was unreasonably expensive -- which means the thermostatic argument is weak.

Expand full comment

Preference falsification in the direction of a perceived trend?

Expand full comment

I think this is really very simple.

The conservative victories you cite - e.g. an unqualified likely criminal being elected President, abortion ban - made the world worse almost without qualifications. There's no upside except a feeling of moral superiority for a small group of hardliners.

The liberal victories you cite - e.g. gay marriage legalization, improved access to healthcare - made the world better, almost without qualification. Yes, Obamacare kind of sucks but it is clearly better than what came before. Nobody is hurt by gay marriage and it improves some people's lives immensely.

So, look for this - does the policy come with almost all downside and little to no no upside? People probably won't like it. Is it almost all upside with little to no downside? People will like it.

I really think it's that simple.

Expand full comment

Not speaking to trade here, but to Roe - I always felt that the overturning of Roe was the worst thing that could happen to the conservative movement. It was a brilliantly uniting quest against something that it's easy to form a strongly negative view on and see yourself as fighting for the voiceless, lending the movement a sense of moral righteousness and even a divine sense of duty. And there's the underdog facet of it too. The moral warriors fighting a huge, corrupt, system of death. I get that and though I definitely lean pro-choice, I'm sympathetic to the other side's points.

But once it's achieved, where is the unity? Where is the quest? It's done, the heroic party is disbanded, and there's no easy replacement for their motivation, though the debate about trans kids is moving into that same spotlight. It's not as potent, though. I feel like overturning Roe was a devastating mistake - the proverbial dog that caught the car and then had no idea what to do with it, or with itself. The chase was the point. I'm not surprised that as soon as it was achieved, opinions changed quickly. The whole thing was an ideological macguffin - only there to drive the story of a quest for glory and righteousness. But in epic quest stories, you usually end up finding out that the macguffin wasn't what you were *really* seeking all along. And I think that was very much the case here with Roe.

Expand full comment

Could it be that abortion affects people more directly than some of those other issues? It doesn't really impact straight people of hay marriage is legal or illegal, but not being able to get an abortion when you want or need one does affect most of the population. Women especially, obviously, but also men to some degree. In large parts of the country, abortion is seen as a very evil thing, so there is an inclination to opose it, but when one is faced with the possibility that it might actually be banned, maybe then people switch to thinking in terms of how that affects them and change their minds?

Expand full comment
Jan 20, 2023·edited Jan 20, 2023

My bet is that it has to do with the media being firmly controlled by the left, and able to whip people into a frenzy, scare them, and do things like find the scary stories and publicize them. We know that there's gotta be no shortage of edge case one off bad effects from any policy - it's like the Chinese robber fallacy. The question is whether these things are brought to the public, and if people grow to think they are central examples, and meaningful examples.

Expand full comment

Abortion might be unique among these examples. I notice that in the health care and homosexuality polls, there isn't an option for "opposed to" the legislation, the way that there is for abortion. Anti-homosexuality certainly exists, but it isn't organized as a movement the same way that pro-life and pro-choice both are.

I'd guess that this affects polling. Pro-life scored a major victory, pro-choice suffered a major loss, how does that affect your opinion? As opposed to: Homosexuality scored a major victory, how does that affect your opinion?

Expand full comment

I suspect this has to do with the 'grounding problem', a quirk of polling-methodology, rather than meaningful change in opinion.

When pollsters ask vague questions like 'are you pro-choice' or 'is international trade good', the participant grounds that question in terms of how it relates to the specific policy debate around that topic in the present. If the nature of the policy debate changes, then the answers change, even if the base opinions stay the same.

Let's relate this to the 5 examples you used.

1. Abortion - When democrats were winning, being pro-choice debate was more about 'how much abortion is too much' rather than 'should be ban abortions all together'. Afterall, Roe-v-wade had supreme court precedent. As an abortion-moderate, your "abortions are okay sometimes" opinion now lands you in the pro-choice category, while it would have put you in the pro-life category before.

2. Global Trade - The US bringing back manufacturing & out-right banning Chinese goods wasn't in anyone's wildest dream pre-2016. With the goalposts shifted, restrictions on global trade came to imply isolationism rather than incentivizing local industry. Opinions stayed the same, but now global trade means collaborative goo-faith middle-grounds instead of shipping industries out to China.

3. The meaning of Gay marriage never changed. So you see a continued steady increase in the base rate without much drastic change in direction or velocity.

4. Obamacare was seen as a compromise bill for facilitating govt. subsidized healthcare. Democrats didn't get what they wanted and Republicans got a handicapped and inefficient program. Republicans are now increasingly answering that "we don't want obamacare" and democrats are increasingly answering "we atleast want to keep obamacare, but we would have liked something better". Obamacare gave the 'govt. subsidized healthcare' question grounding, but the outcome itself was not particularly appealing to either democrats and republicans. So while consensus opinion on govt. subsidized healthcare continues to evade us, giving it the concrete form of obamacare, also allowed people to have strong views on it. (this the polarization)

> n = ~5

Just say n=4. I reread this 3 times looking for the 5th example.

Across those 4 examples, I think my hypothesis holds. Opinions don't change, but a change in grounding changes how polls are answered.

Expand full comment

Is there a counterexample to your thesis in your second plot? What caused the big jump in support for the "pro-life" position between 1996 and 1998?

This was the Clinton era of controversy over things like partial birth abortions, and while I don't remember every single detail of how the debate played out I think it was a general era of victory for the pro-abortion side.

Overall I think it's an interesting thesis but needs more than two examples each way before it starts to look convincing.

Expand full comment

I feel like in the trump case, trump had the ability to change what it meant for someone to be "in favour of international trade". Before, if you asked someone their opinion on such things, there answer would be understood in terms of the present world order, which was strongly in favour of intl. trade. So to say they were against it would imply they were in favour of something more median. Trump was far off to the other side, so people felt the need to qualify that they weren't quite that extreme. In the abortion case, I imagine a slightly different effect: people felt like they ought to be against abortion, but still wanted to be able to get an abortion if they needed it. They wanted moral points from opposing it, but didn't actually want the status quo to change. This was fine, because the issue was decided by the supreme court, and thus totally out of anybodies hands - until it wasn't. The democrat cases listed are of a different nature. Gay marriage is an extremely simple policy proposition, that nobody can really be confused about. Either the gays can marry, or they can't. Health is different in that the democrats were forced to take an extremely moderate position. They might have wanted a full on nhs style government run hospital system, but in order to work with republicans they needed something barely different from the status quo. Anyone against public health care before won't be *more* against it in reaction, nor will anyone in favour decide this is too extreme for them. The general character, in other words, is that people react like this too *extreme* changes that go farther than they like, and not to moderate changes that are more in line with what they expected.

Expand full comment

It's weird. This all makes intuitive sense to me, but I'm not sure I can name it. Some of it is certainly the media effect you mention, but I think there's more to it than that.

Expand full comment

I am not surprised there are different reactions, as the two primary examples (1) pro-life vs. pro-choice and (2) gay marriage are different in kind. Of course, they're similar in that you can't be a little bit pregnant, and you can't be a little bit married. However, there's a healthy debate about when abortion should be legal. There's no such disagreement on a continuous scale for gay marriage (unless we want to roll the poly debate into this). It's two adults, yes or no.

Speaking of adults, adults can be studied, understood, and empathized with. Americans could see gay people before and after the decision and understand if their assumptions about their sincerity and dignity bore out. On abortion, the impact of significantly reduced access to abortion, especially in cases where it threatens the mother's life, is now a fact. Even if you hated the idea of abortion, this is real, significant, legible human harm to reckon with in the wake of the Supreme Court decision. On the other side of the same issue, human life at the scale of a few cells is much more abstract, though its fragility and helplessness raise the emotional stakes of a decision.

Expand full comment

Maybe the backlash represents a lot of people discovering, though experience, the effects of policies after they've been implemented, and changing their opinions based on that information. Like, if you don't pay a lot of attention to the media or don't trust it, you might not be updating until things actually change. Or maybe it just takes people like that long periods to update their priors.

Expand full comment

I think what you're looking at is an effect of narrowing the issue at hand.

In an open context, "pro-choice" can mean a lot of things. In the context immediately following the recent SCOTUS ruling, the context was more confined.

I'd guess if the pre-ruling poll question was worded so as to match the ruling itself closely, you'd fjnd a significantly smaller change in association with the ruling.

Expand full comment

I mean, the data set is so small it could literally be coincidence? (I don’t think it is, but theoretically it could be.)

Expand full comment

Too much overthinking here. True, (program x) cant be empirically quantified by one formula, but that does not mean phenom is "totally random." Examples complex but: Obergfell primarily recognized by smart gays and lesbians in 2015, but not general population; same logic applies to ObamaCare. Few voters fully understood it (that's hard btw) in 2010. Remember Holmes: "The life of the law has not been logic; it has been experience."

Expand full comment

I think the problem with the Obamacare example is that it just wasn't a large change. The poll question is "Government should be responsible for health care". There are not very many people who would say that government was not responsible for their health care before Obamacare and government is responsible for it afterwards. It was just a minor tweak to the system, compared to, say, a single-payer system or a government-run medical system.

In the context of abortion, a change along the policy spectrum of similar magnitude to the Obamacare change would have been a change of 2-3 weeks in the latest date for getting an abortion under most circumstances. I doubt such a tweak would have produced a backlash.

Expand full comment

I think it's a combination of speed and salience.

For gay marriage, and most other things on both sides of the spectrum, the change is gradual, widespread, and follows a predictable pattern. But when something sudden happens (eg, San Francisco doing same sex marriages at the decision of the city clerk) then there is a backlash (proposition 8, in this case). This is because a bunch of people revisit thier latent opinions in the face of a sudden and newsworthy event, and that intensifies soft biases against the thing, without affecting soft biases in favor.

Expand full comment

I don't think your two examples work: On the one hand, there were many widely publicized problems with Obamacare, from it being far more expensive than promised to Obama's lie that you could keep your insurance if you liked your insurance to the fact the computer kept breaking down for weeks after its introduction so that people spent hours trying unsuccessfully to register.

As for the young rape victim (certainly a tragedy) she had to take a 2 hour car ride from Ohio to Indiana to get the treatment she wanted. As you know, people with terminal illnesses often have to travel across the nation or even internationally if they can afford it to get the latest available therapies, so while certainly inconvenient and unfair and unfortunate, the girl was not forced into a back alley or forced to carry to term, but only required to drive 2 hours. Granted, the MSM played it up dramatically, but I'm sure there are many people in prison who wish their actions would have been legal had they only been able to drive two hours.

So, to summarize, I don't think everyone saw Obamacare as no big deal and the young rape victim's plight as horrific. So the explanatory force you see is, IMHO, less than you believe.

Expand full comment

I think there's a genie out the bottle effect. Conservativism thrives best when that which it conserves is existing. If you have to restrict something in order to create something to conserve again, it comes off as more authoritarian than conservative.

Expand full comment

The big problem with such reasoning is there haven't been that many real conservative victories, and a couple of the biggies were signed into law by Jimmy Carter (deregulating prices for trucking and passenger air).

When I was young, concealed carry and homeschooling were outside the Overton Window. I don't see a huge reaction against either. Most of the concern I hear about from the center is "assault weapons."

One could say that BLM is a reaction against conservative policing policies going back to the Nixon days, but boy, was that reaction a long time coming.

How about private ownership of gold bullion? Did that produce a backlash? (I'm too young to remember.)

Expand full comment

What the heck happened in 1996 - 1998?

Expand full comment

The really interesting part of this is the intersection between "the right side of history" and "liberals control more of the media", or perhaps in the space between these concepts.

Things seem to be trending in a direction, and also everything that anyone thinks (including about trends) is filtered through a variety of media-influenced biases. The older post about "do republicans care about babies or do they want to control women?" is an example -- we don't live in a reality of actual stuff that we sometimes have opinions about. Rather our only view on the stuff by which we're surrounded is an uncountable number of lenses (including suppositions about out-group thoughts and motivations). Nobody cares only about borders, abortions, trade, or any of these things -- we care about thoughts about what we're supposed to care about to belong to a team, what 'good' people care about or think, what will most spite the hypothesized-but-probably-imaginary bad people (anything from 'ha f**k you patriarchs!" to "haha look at these wh***s cry now that they can't abort babies!"), and too many more to count.

The people who 'control' the media (by being the media, not by plotting in smoke-filled rooms) are chock full of the same lenses. The lenses (in them, and subsequently in the rest of us -- not that we don't have other ways of getting the lenses, whether it's positive or negative experiences at bible camp or that time the girlfriend had a pregnancy scare or how sick we are of hearing a drunken racist uncle or shrill wine-aunt talk about this issue) create the illusion of trend, and create the trend, and create the space in which we look for trends, and create the ideology of caring about which side of the trend we're on, and whether we're scared or depressed or angry about our position vis-a-vis the trend, etc.

Expand full comment

Liberals control the media? Really? Rupert Murdock's newspapers. Jeff Bezos and Elon Musk are pretty Libertarian. The editorial boards of the NY Times, Washington Post, and Wall Street Journal are not known for their liberal polemics—and even their liberal columnists aren't very. Talk radio is completely right-wing. MSNBC is the only news station that leans liberal (and not so much anymore). Anyway, from my perspective most of the MSM seems pretty right-wing. Of course, I'm a damn commie. ;-)

Expand full comment

The conservatives appear poised to get a victory in the form of an affirmative action repeal. Based on the comments on this NYT article, it seems like liberals are largely neutral or supportive: https://www.nytimes.com/2023/01/15/us/affirmative-action-admissions-scotus.html#commentsContainer Will be an interesting test case to see how successful any backlash is.

If affirmative action gets repealed, the next generation of elites will disproportionately be Asians who know or suspect they are the beneficiaries of affirmative action repeal. I'm tentatively optimistic about that future.

Expand full comment

Would be good to compare with what is happening in other countries in the world. It is not a given that the media leans left,liberal. Here in Australia for example all the commercial media leans conservative (The Guardian online newspaper is the only exception). The public though tends to trust the ABC, which the conservatives claim leans Left but I think most of the public trusts.

Expand full comment

I profoundly disagree that there was no backlash to Obamacare -- I might even voice an opinion like 'the backlash to Obamacare is the central political event in the United States in the 21st century': https://news.gallup.com/poll/4708/healthcare-system.aspx

Consider the second chart on that graph. 69% of Americans thought there should be universal healthcare in 2007. This crashed to a 50/50 tie when Obamacare was passed, and then further crashed to 42% when the rollout was considered a failure, in 2014. It's since recovered to 56% over the last several years, but this is still lower than any number *ever* recorded before 2009. Obamacare damaged the universal healthcare movement in this country in a way that it has still not really recovered from.

(But I think your broader point stands: it is true that there was no backlash to Obergefell, legalizing gay marriage, and a large backlash to Dobbs, which permitted states to legalize abortion. I wonder if there's some aspect of 'sometimes political victories reveal that beliefs were really belief in belief' here? For instance, while 50% of the population might have thought they were pro-life, the possibility of abortion actually getting restricted revealed that some were not; you could observe this by seeing that something like ~80% of Down's syndrome babies are aborted. 70% might have thought they were in favor of universal healthcare, but they had not considered the reality of greater government intervention in the healthcare sector. But something like rising support for gay marriage was real, not a 'belief in belief'. Of course these are often hard to tease out).

Expand full comment

Not understanding why you chose these particular issues and not others. For example, at least some liberals seemed to win a big victory with the post-George Floyd killing upsurge in support for defunding the police. Then, when that policy was at least partially implemented in several major metropolitan areas, support for defunding, and for "Black Lives Matter" seemed to collapse pretty precipitously. In other words, a large backlash effect. Isn't this pretty germane to your point?

The only major piece of legislation passed during the Trump administration was a big tax cut, and I don't perceive there was much of a backlash effect on that. (But I haven't seen statistics) Again, isn't that germane?

If we're gonna test how backlashes break down along politically partisan lines, shouldn't we look at a better survey of issues? Otherwise, it just seems to be speculation. It is true that politics are the most partisan I can remember them, and that there's backlash fairly often, but I can't see any pattern to it, other than people reacting to the implementation of the specific policy.

Oh, the Dobbs decision means, women can't get early abortions or even abortions in cases of rape and incest can't get them in many states, and even can't get healthcare that might be deemed an abortion by extremist local authorities? Oops. Oh, defunding the police doesn't actually either get rid of discrimination by police or make people safer? Oops. Makes some sense to me, anyway.

I think both these things have in common, they are strongly ideological measures applied to very practical problems, not usually a formula for success in even a limited democracy such as the USA.

Expand full comment

It may be apropos of nothing but I found the Ohioan rape victim story to be perhaps the darkest joke told around the whole overturning of Roe. The speed with which that story was dropped the instant particulars began to leak out into the wider world was something to behold, and the uniformity of its embargo was fun to observe as well. I personally bet it would never be brought up again, seeing you mention it here I can only attribute to your distance from the Overton window.

Expand full comment

Isn't the obvious theory that what predicts backlash is the extent to which it was a surprise to voters on the losing side? Backlash here means a sudden shift as a result of the victory. When I'm expecting a political deciscion to go against me either I'm already activated by this prospect so I won't contribute to a post-victory shift or I'm unlikely to be activated by the victory actually happening.

Yes, high-information voters knew that Roe was likely to be overturned but tons of normal everyday ppl had just accepted legal abortion as a fact of life and tuned out yet more people squawking about the republican threat to pro-choice. When the court overturned the case that population suddenly realized there was a real threat to legal abortion and started/shifted their votes.

Similarly re: Trump. A lot of liberals basically assumed he couldn't possibly win. Even to the point of cheering him on. So did quite a few conservatives. When he one it created an "ohh shit" moment where ppl suddenly thought: I gotta actually get out and vote to make sure Trump doesn't get his way.

Expand full comment

Hecklers veto, plus right side of history.

Loosening morals or cultural laws (aside from rules against murder, theft etc) feels good, and the people who protest stop caring as much once they see nothing bad happened. Eve and the snake.

Tightening laws on the other hand affects people clearly and directly

Expand full comment

I've thought about this a lot, and I think there are two main factors.

First, it takes a while for a political change to actually take effect legally or take hold in society, and the backlash only happens then. So no real backlash to the appointments of Kavanaugh and Barrett, even though they were widely expected to result in the overturning or weakening of Roe. Only when it actually happened was there a backlash, since there are real effects of it that people can see (and some of them people inevitably won't like, as with any change), and it's no longer just an abstract idea.

Second, backlash only happens when the opposite party makes a big deal of it, which they only do if it's a winning strategy. The republicans ran firmly against gay marriage in 2012, and lost. After that they moved away from the issue and stopped talking about it, for obvious electoral reasons, and so there was little basis for a coordinated backlash. If I remember right several conservative commentators switched their positions on the issue during Obama's second term.

On the other hand, the democrats won in 2020 on a clearly pro-choice platform, and that galvanised their opposition to Dobbs, meaning they and their media kept talking about it, producing a backlash. I suspect that if Romney had won and gay marriage had still been legalised there'd have been a strong conservative backlash.

Here are two ways to test this theory, for anyone who has the relevant statistics. See if there was a backlash to PC/wokeness and cancel culture during Trump's term when those things (begun during Obama's second) largely took hold in society. (And after Trump ran successfully against them). And see if there was little or no backlash to American global military dominance during Bush's second term, after the democrats ran against it and lost.

Expand full comment

I don't it's totally fair to say "the abortion ruling was a straightforward policy change with no extra personality component."

Certainly it wasn't as provocative as Trump (few things are), but the Dobbs decision was mired in scandal. There was violation of Stare Decisis (I don't endorse this description, but it was part of public perception) and the leak. More importantly, there were years of questioning the legitimacy of the court leading up to this, related to the manner in which the conservative supermajority was installed. That story, an integral part of Dobbs, is intimately tied to Trump himself, seeing as he appointed the conservative majority and all three of his appointees were "Trump-scale" spectacles.

Look at Supreme Court job approval, it went from +20 to -20 in 2 years, and half that change came before Dobbs. I don't think it's trivial to rule out the hypothesis that Abortion became more popular in part because the pro-life movement is now associated with an unpopular court.

Expand full comment

In the EU, there is freedom of movement in that anyone from any of the EU member-states can move for work to another member state.

When Eastern countries like Poland joined in 2004, there was, however, a transitional regime where immigration from the newer members to the older ones could be limited for a time: member states were allowed to limit immigration inflows if they chose.

The UK was one of the few member-states that kept its borders open to all EU citizens. Backlash against immigration exploded and now the UK is not a member of the EU.

Expand full comment

I think you get the backlash when the original position was membership signalling and/or trolling the other side, and seemed unrealistic.

So when there was no realistic prospect of an abortion ban, there was no downside to supporting it. Once once comes in, suddenly there are practical ramifications that affect you - even if male, it's going to affect your sex life - and the women in your life.

Same goes for Trump. It must have been great fun supporting him to "own the libs" etc. Less so to see him in actual power.

And somewhere in there is a learned instinct to shift the Overton Window by supporting extreme positions you don't really want to see actualised.

Expand full comment

Could be external influence. I.e. international community going "Really, America, you serious??!"

Expand full comment

My first thought on this is that the left actually has an effective machine for influencing peoples' outward opinions, and they use big events to coordinate on particular issues. A widespread change requires hyperfocusing on one thing, so that everyone opposed hears the push coming from all around simultaneously.

The right has no equivalent machine that can push politics on the public; no one shifts their attitude due to "public pressure" from the right, on anything. There's never a flood of right-wing arguments and slogans that everyone hears at once, complete with support from every major corporation, most media entities, public protests, pushy individuals, the internet, etc., to generate an illusion of a public consensus.

Expand full comment

I suspect part of it is just normal navigation. When you’re far from your destination, you follow the large highways or get on the trains that go in your general direction. But when you get closer to your destination, you need to make some choices, and figure out more precisely where you want to go. You’ll probably spend some bandwidth to take onboard a higher resolution map (i.e. think things through more carefully).

This is reinforced when, in times of heightened debate, relevant info and good arguments are prolific. So you encounter more solid counter-arguments to your own position, often forcing you to modify and nuance your views.

And I don’t believe the trend just goes leftward (see anti-woke sentiment), but rather toward some perceived political center or locus (x, but not a fanatic), that balances the concerns people have taken onboard.

Of course, these centers and loci are partly determined by social and psychological factors (the Overton window and personal biases), but also by objective reality (economic and biological factors). Currently, the political center in the west is pretty socially and economically liberal.

You can see it in that many or most people will modify their views when they see they may have to own the uncertainty and salient, negative consequences of those views. (I’m not sure there are many salient, negative consequences of Obergefell for most people, which may be why we haven’t seen a big counter-reaction.)

It’s everywhere, including on Twitter, in business meetings, and among friends out drinking: Someone will argue vociferously for some strategy or option, building up how amazing it would be to build the wall, or buy crypto, or follow a certain marketing strategy, or go sing karaoke. But when they suddenly, often to their own surprise, get buy-in, they will immediately modify their language and try to curb everyone’s expectations.

Expand full comment

What's going on with disappearing comments? Are we being quietly censored or is there a bug in Substack swallowing comments?

This is the second time, to my knowledge, that a comment of mine has disappeared from a comment thread with no notice. The first time, there was a possibility that I'd composed the comment and forgotten to go through the formality of posting it. But that definitely hasn't happened here, because I have email about a response to my now-vanished comment. What gives?

Expand full comment

It feels worth pointing out that the pro-gay-marriage chart has two downward spikes; one in 2012 when Windsor was decided, and one in 2016 when Obergefell was. So by that standard we should see abortion return to normal trends later this year.

I feel like the years-long clusterfuck of Brexit could explain the prolonged upswing of international trade love. Likewise the Trump administration had stories of sterilizing immigrants and sending them away without their children, which would change minds on immigration policies.

Expand full comment

What tends to happen with these hot topics is that people who didn't really think that much about them end up evaluating/re-evaluating them once it becomes a hot topic. This may of course look more like introspection, or more like just adopting the ideas held by your friend group, depending on the person. The reason why you have these shocks is just that suddenly 10+ years' worth of changing minds happen over a short time interval of perhaps 1-2 months. It's just like the "stochastic clocks" in financial theory, if you have heard... sometimes there just is a very intense amount of market behavior one day.

The reason why the shocks happen in the liberal direction is unironically "cthulhu always swims left" all over again. But why worry? I think 4/5 of those are good myself and I expect the 80-20 rule to continue holding for the foreseeable future. To the extent that liberals hold bad ideas they tend to be bad ideas in the sense that they make them unpleasant to talk with, or they make their families cringe at the dinner table. Their ideas are not usually (too) bad in the sense that they're a terrible direction to develop national policy in.

Expand full comment

Someone who thought Roe was trash but is ok with a European style 16 week cutoff used to be pro-life, now with the shifted landscape they are pro-choice

Expand full comment

Have you considered the possibility that people who were pro-life before didn’t realise that their commitments were closer to pro-choice until the decision?

Expand full comment

But the international trade thing didn't get walked back with Biden. If anything, the Biden regime has been tougher on China than Trump was. The poll results reflect the hoi polloi's virtue signaling against the bad orange man, while the actual decision makers have a free hand since no one normally pays attention to foreign policy anyway.

Expand full comment

I think that at any given time there are way, way to many issues to think about consciously and to be constantly updating our beliefs on each and every one of them, even if we are professional opinion-havers. So, for most of them, we default to the dominant ideological position (the "common sense") until the issue comes into our focus for whatever reason and only then we are kinda forced to think more thoroughly about it (and maybe change our minds).

I eat meat without considering its ethical consequences until I gaze into the eyes of a living cow that is about to be killed and seems to know it and fear it or I see some vegan activists on tv or read a blog post or whatever.

Expand full comment

I think your social bubble is seriously influencing what policies you see as having "a backlash"; I'd say the strongest backlash effect I've seen was against the ACA, which resulted in a historic defeat for the Republicans, which many were claiming would be the end of the Republican party, into a historic victory in the next election cycle.

There are two significant issues. First, separating people out by affiliation erases the backlash effect when it causes people to switch affiliation - which I think was a much more powerful part of the effect with respect to the ACA compared to other issues (The ACA coincided with an increase in conservative affiliation by 4, and a decrease in liberal affiliation by 4). Second, the polls there aren't asking about the actual policy, but a different policy entirely - the ACA isn't government-run healthcare, and the mandate, which was a central point of the backlash, was in fact a massive gift to privately-held insurance companies. I expect you'd see a radically different picture if you asked about support for mandatory health insurance.

There are a few key ingredients in a backlash.

The biggest is novelty. The public doesn't have the energy for a backlash about the same issue every year; I think this is the most important part of Obergefell, which clearly has a backlash among all parties in the graph you posted. It's just that the majority of the available backlash energy was exhausted in 2013, when ten states legalized gay marriage. People still got annoyed in 2015, but it lacked staying power, because most people were already tired of talking about it.

The next is that the policy is boolean, but the subject of the policy is not; polling and voting tends to be directional in these cases, rather than about desired policy outcomes. Abortion is perfect here; the majority of US citizens want some abortion restrictions, but do not want it to be completely illegal. In the pre-Dobbs political context, if you want some abortion restrictions (and this issue is important to you), you are in a sense on the same side as people who want abortion to be illegal. In the post-Dobbs political context, if you want some abortion restrictions, you're in opposition to the people who want abortion to be illegal. Directional polling is hugely significant.

And, finally, identity. Your identity group has to care about the issue - and keep caring about the issue. To some extent this ties in with novelty - if I look around, and the people I'm trying to fit in with don't seem to care that much, I'm not going to try to appear to care too much. If many people are exhausted of the issue, this makes my identity group appear not to care that much about the issue, so I'm not going to make too much of an effort either.

Expand full comment

My understanding is that we went from one of the most permissive abortion regimes in many states to one of the least permissive in many states (when compared to international norms).

I'd credit this one to some people holding the position of "relative to third-trimester abortions I'm pro-life" and "relative to a complete ban on abortion I'm pro-choice."

Expand full comment

This is pretty obvious. The media

Expand full comment

My initial reaction is to endorse the "media effect" thing, but that's because I always forget that most of the U.S. follows deeply unhinged conservative media instead of somewhat unhinged liberal media.

Instead of looking inward I'm just gonna revise my hypothesis to save it even though the facts have proven it wrong - most people in the U.S. don't consume any news media at all and instead form opinions based on opinions based on opinions based on...based on opinion makers - prestige media like WaPo and NYT.

Expand full comment
Jan 20, 2023·edited Jan 20, 2023

These graphs show only PRO and CONTRA, they don't show fraction of those who don't have an opinion. And the truth is - in the end - everyone on every question is either-or, rarely there can be a political question that truly doesn't matter, but a lot of people just can't / don't want to spend resources on analyzing which side they are slightly leaning from the middle.

So, my pet theory - those who were pro mostly stayed pro, those who were contra mostly stayed contra and the changes are due to people in the middle more pushed to make a decision when their opponents win. And why it happens sometimes and not always is due to the fraction of "hidden pro" / "hidden contra" in the middle - on some questions it's far from 50/50...

Expand full comment
Jan 20, 2023·edited Jan 20, 2023

It's clear in the health care graph that there was a huge drop in support for government healthcare as Obamacare was being widely discussed in 2008-2009. Looks like a 20 point drop among independents! There wasn't backlash after it passed because the backlash had already happened. So the idea that there's no backlash against Democratic priorities seems definitely wrong. As for why the backlash happened before the law passed, this probably reflects the different dynamics between long public legislative processes vs. supreme court decisions.

Expand full comment

Surprisingly you are missing the second most obvious (aside from Liberal dominance in the media): maybe the polls are also influenced heavily by Liberal power and not lying just like the media doesn't lie.

It's pretty easy to create a poll to say almost anything you want. And if you don't get the results you want with one poll, just do a few more. Then report out the poll that give the right answer.


Expand full comment

There are confounding variables on these examples. Your Republican examples are about taking things away (trade access, abortion access) while your Democratic examples are about getting or protecting something (gay rights, single payer healthcare).

It may just be that in the examples you've selected, the Republicans picked less popular fights than the Democrats. Taking things away is not popular.

Suppose the following: the public doesn't want abortion access taken away. However, they're not paying attention to day-to-day political signaling, possibly because they're tired and don't have time. Then, when SCOTUS does overturn it, nobody's priorities or stance actually changed. All that changed was the public's ranking of where the fires are. That would be sufficient to explain the effect we see.

Expand full comment
Jan 20, 2023·edited Jan 20, 2023

I think we have a deep cultural bias in the west and particularly America in favor of change and progress. We love innovation. Our attitude towards the past ranges from a contemptuous dismissal to hatred and resentment.

All in all its worked out pretty great, its given us scientific and technological powers beyond imagining. Its made us the richest, freest, most powerful civilization ever. Its abolished slavery and aristocracy most other antediluvian forms of arbitrary hierarchy.

But that same impulse for change has led us into all kinds of mistakes and crimes and disasters along the way. Nationalism and hyper-nationalism and fascism were progressive in their day. Eugenics was progressive in its day. Lobotomy was progressive in its day. The Indian residential schools were progressive in their day. Its only with hindsight that we re-code everything failed and monstrous as right wing or reactionary.

Conservatives want to conserve what's working now, or was working 10 minutes ago, and they always have. There's nobody in our politics that thinks the revolutions of 1688, or 1776, or 1865 were bad and want to go back. There's hardly anyone left who seriously proposes undoing the New Deal -- or at least not the parts of it we're still doing. If today's progressives even knew what a NRA blue eagle was they'd say it was right wing and fascist-adjacent.

Progressivism is a random walk, moving us in every direction. Sometimes it works, and that's great. Sometimes it doesn't work so great. When the wrong directions are finally recognized, they're either forgotten or re-coded as reactionary and right wing in our historical imagination.

Expand full comment

The asymmetry is that conservatives are willing to follow the rules and accept their opponents' victories as legitimate, but liberals are not. The clearest example of this is that conservates worked within the system and established rules to get Roe overturned, but liberals are already beginning to treat the Supreme Court as illegitimate. And of course this is starting to change for conservatives as well, since you can't have one side flout the rules _forever_ and still claim that the rules actually hold.

Expand full comment

Idea: US political messaging is targeted at an ingroup, disregarding how it's perceived by an outgroup. Maybe the exact same hype and media that brings political "success" is what polarizes the opposition, so we should expect them to be coupled. No idea how to validate this, esp. with respect to your examples, but I think we'd somehow need to quantify the framing and intensity of messaging around a topic. Legislative/court victories should be minimally coupled since they probably don't depend much on messaging, while elections/ballot questions should be tightly coupled because of all the advertising? Supreme court leak confounds your example because it stirred up so much media coverage? Idk, curious if this resonates with anyone.

Expand full comment

Do you actually see an effect on abortion views? I'll spare you the Fenyman quotation.

Because what I see is pretty much a series of coin flips, or draws from a box with white and red beads.

You buried your lede: "Or maybe this is totally random and I shouldn’t try to conclude things from n = ~5 examples." Statistical thinking is what helps cure apophenia. If you are going to be a rationalist be serious when looking at data.

As to "Trump beliefs" - he's got beliefs?

Expand full comment

I think it's related to focus of attention. People tend to accept what those they respect (for whatever reason) tell them is proper, as long as there's no effect. When it becomes real, they start thinking how it might affect them and those they care about. Liberal causes, on the average, tend to positively affect more people than they (obviously) harm. Conservative the opposite. People here has to be understood as "those who do or could respond to surveys".

I've got beliefs about how the (obviously) maps onto reality, but they don't follow "liberal vs. conservative" lines.

Expand full comment

Is it possible that left-wing media gets angry about politics while right-wing media just counters whatever left-wing media is saying (thus allowing them to set the topic of debate)?

Expand full comment
Jan 20, 2023·edited Jan 20, 2023

The typical non-evangelical conservative voter didn't really care that much about abortion, but sympathized with fighting coastal elites (read: the Supreme Court) telling them what to do. Pro-life was an easy cause for conservative politicians to adopt without risking anything.

Once the Supreme Court withdrew feds from the abortion business, non-evangelical conservatives had nothing left to sympathize with. Now that red-state conservatives actually have the burden of implementing policy, the simplicity and upside of the pro-life cause has evaporated.

Expand full comment

It’s precisely that, which side owns the Cathedral institutions and can thus generate backlash among independents. In the USA and W. Europe that is of course, progressives.

Expand full comment

The "direction things naturally tend" argument makes most sense to me. It's as if there is sort of a "ratcheting" mechanism as the general public mood slides slowly in the direction that most commonly is called liberal.

When things push in that "forwards" direction (not intending to make a value judgement here), it's like a frog in boiling water, most people just tend to shrug and go along with it. Yes many people will cry about the degeneracy of modern culture, but the specific policy positions get more and more unfashionable over time and the mean conservative/reactionary stakes out more and more leftward territory as the decades go by. How big of a persuadable audience could someone gather in 2023 arguing against interracial marriage, for example? Even the prominent conservative voices today express a lot of opinions that would be very, very left not too long ago (notice how lots of conservative commentators had at least a few complimentary things to say about Bernie Sanders).

Whereas big swings that go against the "ratchet" give lots of people the sense that something has gone *very* wrong vis-a-vis the direction that things are "supposed" to be. People as a general rule can be OK with complacency but they HATE to backslide (a lot of the MAGA psychology is "we had more status in the past"). Which is why the center-left (and the median American really is center-left) that is mostly OK with the slow or even stalled-out advance of progressivism can get very seriously perturbed when Roe gets struck down.

The average American voter is desensitized to the idea that things will move moderately leftwards over the course of their lifespan. (how many times have you heard someone say "our grandkids will probably think it's outrageous that we ate as much meat as we do"). From that point on the distribution, "lots of progress" is about as equally far from the middle as "no progress", which means that "backwards progress" (or perceived backwards progress) is even farther away from expectations than lots of progress.

Expand full comment

I'm imagining a story where public opinion drifts around, and but people refrain from really identifying with a position until there is a catalyzing event, at which point people begin to firmly identity with the positions that they've been gravitating toward. Perhaps a month before the abortion ruling, there are a lot of people who have become gradually more pro-choice, but their shift doesn't necessarily mean a different position on a survey. Then the ruling happens, and they realize that they need to "pick a side" and so it looks like a rapid shift happens.

According to this story, the effect might not show up in situations where people have reason to identify with the position before the catalyzing event, or where the catalyzing event might not cause people to want to snap to an identity.

Expand full comment

UCLA economist Brian Wheaton has a paper on this with n>5 I think. Here is the abstract: "Do laws affect the beliefs and attitudes held by the public? Using data from the ANES,

the GSS, and Gallup along with a difference-in-differences identification strategy, I find robust

evidence that virtually every major U.S. social policy law of the past half-century has induced

significant backlash. That is, the public moved in the opposite ideological direction of each law. I

show that my results are consistent with a model whereby individuals care about the ideological

beliefs of other members of their society – such as their children and peers – but only have

imperfect ability to influence said beliefs, which are also influenced by the law and the actions of

others. I empirically test the implications of said model; for instance, I show that the backlash is

successfully transmitted to children, more persistent in ideologically-homogeneous communities,

and exists on both sides of the political spectrum. Furthermore, I provide evidence against a

variety of alternative models that would also generate backlash." and the link: https://www.anderson.ucla.edu/sites/default/files/document/2022-09/Laws%20Beliefs.pdf

Expand full comment

I think there's a connection between this and prior discussions on "belief" vs. "belief in belief."

That is, there exists some number of people who take positions on political issues because they believe they believe them, or they want to be the kind of person who believes them (maybe call this "aspirational belief.") There's no immediate cost to holding these kinds of positions as long as they're in opposition to the status quo or strongly entrenched as the status quo.

For example, if I'm a devout Catholic, I should be strongly opposed to abortion in almost all cases. The strongest way to oppose something in a democratic society is to try to make it illegal. As long as abortion remains a protected right and I don't actually personally need one, I can safely believe that I believe it should be illegal. But when the *Dobbs* ruling comes down and I'm confronted with the real possibility that I or someone I care about might be affected by a total abortion ban, maybe I start to question whether the law is the right tool to express my moral conviction.

On that theory, the victories that are likely to result in backlash are the ones where some poll respondents are expressing something other than how they want the future to look. Someone who says they support overturning *Roe* may just be saying "abortion is bad;" someone who says they oppose free trade or immigration may just be saying "the modern global economy has left my community behind;" someone who says same-sex marriage should be illegal may just be saying "I'm a Christian."

The majority of respondents in either case may actually believe that the policy they're supporting is good. But some fraction of them just believe they believe that because that's what people with their values/emotions/identity are supposed to believe.

I think the reason you can only pull up conservative examples on the national scale is that the liberal issues with this dynamic tend to be local. It's easy to imagine ways it could play out on a national scale in a hypothetical universe where the far left had more power at the federal level: polled support for slavery reparations would drop even lower than it already is if there were a real chance that white people were going to be differentially-taxed to fund large cash payments for the descendants of slaves. But in real life, most of the left monkey's paws are local concerns like housing/homelessness, school policy...the stuff that dominates local elections in blue cities.

Expand full comment
Jan 20, 2023·edited Jan 20, 2023

I think the most interesting aspect of the Gallup poll is that the 2022 poll began on May 2nd, the same day the Supreme Court draft memo was released. It could be a coincidence. Gallup starts this poll in early May every year. The case was argued in early December 2021 and the decision made in late June 2021. The opinion had to be written before it could be leaked. So there may have only been a couple months of time in which it could be leaked, so April-May might have been a natural time to leak it.

Even if we grant the leaked memo would have had a similar effect if released a week before the poll, that still leaves about a 10% likelihood of it coinciding with the start of the poll. And it's obviously even less likely that it happens to be released on the exact day the poll starts by chance. Even if the leaker didn't know the exact start day for sure, Gallup is very consistent in starting in the first few days of May, so they'd have stood a good shot of releasing it on the actual first day of polling by chance.

This makes me entertain the possibility that the leaker deliberately timed the leak to coincide with the Gallup poll so that they'd pick up the salience of the issue in their polling. If that on balance led more people with earnest pro-choice views to take the survey, that could explain the shift in expressed support without actually reflecting a shift in real opinions.

Having worked in a call center for a year, my experience is that people are much more likely to answer the phone if they perceive participating in the call as being valuable to them. It would not surprise me that pro-choice people who just discovered Roe v. Wade was about to be overturned for real viewed the ability to express their views in a survey as being of unusually high value at that time, and chose to take those calls more often than they normally do.

My best hypothesis to explain why this happens after some political victories and not others is that expressing support for a popular political view is most useful when fighting against an unpopular position or person that has nevertheless achieved political victory. "You may have won formal power, but most people don't like you and we're going to make sure you know it and we know it." Under this hypothesis, people would tend to respond to surveys more when they've just suffered a loss despite holding a politically popular position.

According to Gallup, 80-90% of people support some level of abortion access and 70% support gay marriage. Support for government guaranteeing access to healthcare is 50-60%. Victories in favor of gay marriage and government access to healthcare were politically popular, so under my hypothesis, people don't feel that participating in a survey would be an especially useful. The majority who support gay marriage and government healthcare feel that their institutions are reflecting their belief, and those who are against feel that wielding public opinion is unlikely to be a useful way to fight back.

Trump lost the popular vote, and international trade was a majority favorable position before 2016. So this again fits my hypothesis in which the victory of an overall politically unpopular figure leads to a surge in willingness to take the time to make expressions of support for the popular position. This surge in expressiveness then gets picked up by polls.

Expand full comment
Jan 20, 2023·edited Jan 20, 2023

Re: how the media is involved --

After the abortion ruling there has been an obvious uptick in stories about ridiculous injustices due to abortion being illegal, probably in part because they're topical, in part because they get a lot of clicks, and in part because there are more actually relevant things happening as a result of the ruling (directly or indirectly -- like some people now can't easily get abortions, but also maybe some anti-abortion people are galvanized to be annoying about it after the ruling). So the average member of the public might be more generally aware of how bad abortion restriction is to people than they were before the ruling, and hence their opinion might swing in favor of abortion.

I mention this to say that it's important to keep in mind that this cause & effect is not simply "the media tends liberal -> more abortion horror stories -> more anti-abortion sentiment". The causal diagram also has a node for "abortion restriction is actually inhumane" (or if you don't like that phrasing, "abortion restriction causes major and easily preventable problems for people"), and this is also a cause of "more abortion horror stories". (It's also a cause of "the media tends liberal" but that's probably less important). This is an argument against the simplistic stance (expressed in another thread) that "well the media tends liberal so that's probably the main explanation for backlashes against conservative victories". Part of the backlash is _caused_ by the bad thing that liberals not-coincidentally tend not to like (abortion restriction being actually really bad for people).

Anyway this is also the reason that liberalism is on the right side of history: the stances weren't randomly assigned to conservative vs liberal camps; they've assigned based on whether they're, like, good stances, and the good stances are usually going to be on the right side of history. Or, in my biased opinion: the liberals took all the good stances on social issues so no wonder they come out ahead in the long term. (It's not at all clear to me that economic stances weren't basically randomly assigned though.)

Expand full comment

>Maybe because liberals control more of the media than conservatives; when conservatives win something, the media does a good job making everyone panic that conservative ideas are taking over, and convinces them to do a thermostatic reaction. But when liberals win something, most people either don’t hear about it or don’t hear anyone telling them to worry, so they don’t.

I like how Scott buries the actual answer in a few subclauses of one paragraph.

Expand full comment

I find it hard to believe that Americans are more pro trade than ever before. Maybe my perceptions are just wrong, or maybe people have a different idea about what it means to be pro trade, so you can’t compare the top line number.

Expand full comment
Jan 20, 2023·edited Jan 20, 2023

Here's a facebook video via the daily mail of dogs barking at each other until a gate is opened.


Is it possible we're running into some signaling that might be explained by game theory? That is, opposing abortion might be +1 more virtuous until enough people oppose it, abortion is banned, and then it switches to -3 abortion is actually illegal. People then readjust to try to relegalize abortion, but once it's legal/accepted opposing it becomes more virtuous again. If this model is correct then similar effects might be seen for liberal policies like more extreme energy policies.

It's not exactly the same, but Germans are now more pro nuclear power. https://www.energymonitor.ai/policy/weekly-data-shift-in-germanys-perception-of-nuclear-energy/

I think being extremely pro "green energy" or "anti-racist" probably benefits from the same signaling game theory. Maybe there are some examples of this sort of thing happening locally and getting pushback.

Expand full comment

> Or maybe it’s because liberalism is “on the right side of history”, ie the direction things naturally tend

As a slightly weaker/less-universal explanation, is it possible that there is some non-partisan conception of "quality" of policy, that is orthogonal to the partisan dimension, and which most people are capable of honestly assessing (sometimes in advance, but much more easily after the fact)? By which I mean, for a given partisan position, a "good" policy would actually give precisely the disired effects when the plan makes contact with reality, and the "bad" policy has unintended consequences and does harm, perhaps doesn't even improve the thing it was targeted at.

For example, Trump's tarrifs did substantial harm to US businesses, and those knock-on effects had to be countered with direct subsidies. A better policy for that position might have been to go out with some subsidies at the same time; even Republicans recognized that it was just a bad implementation.

Pro-life Republicans want less access to abortion, but my understanding of the polling is that most don't want a total ban, which is what Dobbs produced in some states. So enacting that policy resulted in much more serious consequences than most Republicans were actually in favor of (even if Republicans in a given Red state might have got broadly what they hoped for inside that state).

A problem with this analysis would be Scott's point that the "support for international trade" increased _before_ the tarrifs were implemented. I think this suggests that there is no monocausal explanation for this; we have to include press coverage as a major driver of sentiment in addition to ex post analysis of side-effects.

But even then, quality and content still matters; I think there's only so much that the press can convince people to support. Without any specific proposal to reduce trade, one might say "sure trade is OK I suppose, but I do have some concerns about {globalization and sweatshop labor | NAFTA taking our jobs | colonialism | wars in the middle east | insert ambi-partisan objection here}". On the other hand, with a specific proposal for less trade (say the Trumpist protectionist agenda), now we have a concrete position to argue against, and so if one doesn't support that specific agenda, one might move from "undecided on trade" to "enemy of my enemy is my friend, I support trade against this form of anti-trade".

So, three related possibilities: bad quality of policy evaluated ex post, press coverage of policy ex post, and a process for coalescing opposition in response to a specific policy.

Expand full comment

1) Defund the Police and BLM generally basically collapsed in popularity after the November 2020 election.

In fact I would say this is Round X in this typical dance on how the public views crime since the 60s.

2) COVID NPIs basically collapsed after the Omicron wave (I would argue the second Youngkin won in November 2021, but some time elapsed to save face). If public health tried to try shit again, I think it would get slapped down.

3) Although the Republican Party in general couldn't take advantage of the above, certain Republicans could (Ron DeSantis, etc).

Expand full comment

This seems more situational than pattern-based. What if we looked at progressive ideas like "low-income housing projects," "cut back your use of energy," and "defunding the police"? Sure, these are not policy victories, but they're still policy ideas that had some level of implementation at different points in time.

In other words, are you doing some selection bias based on framing it exclusively as policy victories vs. political salience of the core ideas?

Expand full comment

Another obvious (to me) hypothesis is that democrats are disadvantaged in Congress and in the courts, so they can only get wins on issues where they have popular support, whereas conservatives can get wins on issues with less support thus producing greater backlash.

Expand full comment

I think the most relevant analogy to Obamacare would be the TCJA. Was there a major backlash to TCJA (I'm not aware of one, but I haven't looked). If not, that would be a strike against "liberal propaganda controls everything" and evidence in favor of "laws that do popular things don't cause a backlash".

It's also worth noting that Obamacare got a lot more popular in 2017 once Republicans seriously started trying to kill it.

Expand full comment

The media bias theory makes the most sense. Most mainstream media sources are skewed left. This includes 3 of 4 broadcast networks, 2 of 3 cable news networks, the NYT, NPR and so on:


I’ve also noticed that the major news aggregators that supply the news to our browsers and phones (Google, Apple) favor these sources. That means the “default” news most people see will overwhelmingly skew left — unless you intentionally curate your news to include sources from the right. The one exception is Fox, but I think most people are acutely aware of Fox’s bias, so it’s different than some other sources that are perceived to be neutral and are not.

Expand full comment

Seems like in the abortion rights and Trump free trade cases, new constraints were put on the populace; or, alternatively, freedoms were taken away. In the former, a right that previously was taken for granted was put into question; in the latter, trade faced new restrictions that, in principle, limited businesses’ options relative to the status quo and drove up costs for everyone.

In the Obamacare and gay marriage cases, constraints were taken away / freedoms were added. The lives of people with insurance and heterosexual unions weren’t really different after those events, nor did the potential for significant change really increase.

Generally, people don’t like when freedoms (or anything really) are taken away. Also, they don’t really care when new freedoms they won’t exercise are gained. I don’t think you need to invoke left/right dynamics to make sense of this, though I’m sure they might affect the intensity of the reactions.

Also, with respect to the small number of examples, maybe you could look at sentiments after different COVID restrictions, prohibition, women’s suffrage, civil rights, and/or Brexit if we’re including non-American events.

Expand full comment

There doesn't have to be the same reason for two similar reactions. The Trump backlash could be rooted in how individually shitty he is, while the post-Dobbs reaction could be about something else.

If you do want to find a common factor here's a possibility: events making some people focus on an issue more deeply than they have previously. Pollsters have terminology for this which I do not remember right now, but the basic idea is that a lot of the time a certain fraction of public opinion about issue X is not based on a lot of deep thought or strong feeling for/against X. But then something happens (e.g. the SCOTUS suddenly and sharply changes the policy reality of X) to make some people give X some real thought for the first time.

And the Trump factor could be synergistic with that. E.g. people having no strong opinions about immigration policy being startled and turned off by Trump's open racism towards immigrants (I have one of these in my own household), which makes them more likely to respond a certain way when asked by a pollster.

Expand full comment

What about Defund the Police. How did that go.

I agree with other commentators that it's basically about whether the "win" has immediate trouble associated with it that was ignored when pushing the belief, because the belief was really about something else (e.g. abortion is often about traditional family values and don't be promiscuous, defund the police was really about inadequate training and empathy in some high profile cases). It was kind of a posture. In fact I'd say voting for Trump itself was kind of like that for some people. Things are so bad I'm gonna vote for TRUMP! Oh wow, he won, what happens now.

You could speculate about how long-term goals play into this. For example, short-term pain might be more apparent than long-term gain, short-term pain might cause more backlashes (not that there is any long-term gain, but that it would suddenly take a backseat, for example if liberals just shut off all the oil overnight).

Expand full comment

I'm going to hypothesise three different sources of backlash.

1) The Trump: This one is all personality. A big figure becomes associated with a policy and people who don't like that person turn against the policy.

2) The Insulation Effect, AKA, The Abortion: This backlash requires people who don't like a policy, but feel safe to advocate it. Imagine if you will a Republican woman making a trade with a bible thumping preacher "I'll vote for your abortion policies if you vote for my tax cuts". Because the supreme court will strike down any anti abortion law this trade costs her little. But as soon as Rowe vs Wade is struck down she has to seriously consider whether those tax cuts are worth control over her reproductive system. So all the people who were pro-abortion as a political trade, as tribal signalling, basically anything except genuinely believing abortion is evil. They all have to reconsider.

3) The Enemy Victory, AKA, The Brexit: This one is straight forward. The backlash comes from people who were always against a policy moving from passive support to active vocal support once their policy is under threat. If you like being a member of the EU in 2000 you probably don't think about your membership much. Tomorrow the sun shall rise and you shall be an EU citizen. In 2016 campaigning could make the difference between membership or leaving, so you campaign hard.

P.S. I think there's some of 1 in the backlash to Rowe vs Wade. I'm sure clips of politicians talking about said 10 year old triggered people's disgust in the same way that Trump can.

Expand full comment

If a government takes *turns* whacking both sides of an issue which divides us, that's not justice. Government should only intercede when we've had enough dialogue and evidence to reach a super-majority on ethically-charged issues. It's safer to be conservative about the *application* of government in the ethical sphere.

Expand full comment

Implicit in the question is an assumption that voter opinion on specific issues can be modelled statistically *without taking into account what the issues are*. One can go a surprisingly long way, and yet there really is more to politics than simply rolling dice. Decisions have impact, and whether they make life better or worse is not entirely subjective.

After a decision is made, if life becomes noticeably worse, one may expect some backlash. Whether or not this will happen isn't something you can predict *purely* by looking at which tribe made the decision or what political colour it is. You have to also consider the actual effect.

Unfortunately, whose life a decision makes better and whose worse is much more difficult to quantify than how many people style themselves red vs blue, and so people keep pumping their intuition engines with the latter as a poor proxy for the former, beliving the output uncritically then being surprised when it fails to match reality.

Expand full comment

Prediction: The upcoming (and all but certain IMHO) demise of affirmative action/racial preferences in higher education will be broadly mildly-popular despite being narrowly extremely-unpopular.

Expand full comment

My guess would be that policy changes that are perceived to increase the typical voter's ability to do as they please are unlikely to provoke significant backlash, but new restrictions on perceived freedom will tend to create backlash.

People just don't like it very much when they're told a freedom they've come to expect is being taken away. Even someone who is nominally opposed to abortion on a moral level may balk once they realize the government is *actually* going to force them to carry a fetus to term.

But when a new freedom is introduced, for instance gay marriage, it doesn't really hit the same way. Even if you're opposed, your life isn't actually changed. Your theoretical opposition is still mostly theoretical, and will tend to erode over time as people do the thing you don't like and the world carries on just fine.

Expand full comment

Part of it is that Democrats tend to be more, well, democratic: they are more likely to pick fights and policies which already have widespread public support (it wasn't so long ago that Obama publicly disavowed gay marriage, for example). This is in part due to the voting system (Democrats are the urban party and Republicans the rural one, and the voting system gives rural voters outsized power, so Democrats need way more popular for equal electoral performance than Republicans do), and in part due to Republicans really having gone off the deep end recently (I think Hanania's recent take at https://richardhanania.substack.com/p/why-the-media-is-honest-and-good is the correct one on this: their extremely low standards for truthfulness in allied media are causing their core voter base to increasingly lose their grip on reality).

Another thing is that on some topics, one side is just right (or at the very least, significantly more convincing to the average person after careful consideration and examination of arguments). So while the topic is low-salience, people tend to think whatever their preferred media thinks; but once it becomes a major point of discussion, most people can't / won't avoid engaging with arguments from both sides, and truth might override ideology. This happened with abortion, IMO: until it was regulated by courts, the average person did not have much reason to care; once it ended up on ballots and became a point of intense national debate, people reexamined their views, and pro-life views are just harder to defend.

Trump is a bit of an exception, I think, because he tried to take the Republican party in a new direction, so for a while his proposals went against both the liberal press narrative and the conservative think-tank narrative. (And it surely didn't help that he was incredibly unpopular at the time + did not look very competent at being a president.) So that was a combination of the issues suddenly becoming more salient + his ideas not really having an intellectual "heartland".

Expand full comment

Scott, If I read your source on the abortion graph right the 2022 datapoint is from May and Dobbs was decided in June. So this may not be the effect you are looking for.

Expand full comment

For me the clue is in the names “conservative” and “progressive”. People vote for conservatives to maintain a status quo/tradition and for progressives to make change. I can understand why there would be backlash if conservatives made changes, and why there is so little surprise when the same is made by progressives.

Expand full comment

The media is certainly a big part of the equation, but aside from that: people have an aversion to losing their rights. They may not be very happy about their political enemies gaining rights, but they'll panic and get angry if you try to take something from them, particularly if they used to take it for granted.

On the other hand, if an issue doesn't concern them particularly and they don't have strong convictions about it (no religious beliefs, no partisan solidarity), they may have an interest in picking the victor's side (particularly if they're surrounded by people who have strong convictions about it).

Expand full comment

Less backlash to laws that introduce freedoms than that take them away? This would not seem shocking to me.

(Obamacare is something of a mix, though, so it might not apply.)

Expand full comment

Roe was popular and fits what most people want. IMO the reason for the change.

Expand full comment

If the mechanism that causes a backlash is seeing the results of a decision, then you are not going to get a backlash every time, because people are not going to hate what they see.every time. Similarly, if the mechanism is a public debate, that might change minds, but not alway.

Expand full comment

The huge tea party victory in the 2010 midterms seems like a pretty big backlash against the Obama presidency. For that matter, Trump himself seems a lot like a backlash against Obama. Those are pretty big conservative backlashes against liberal victories. Are you counting them?

Expand full comment

There is a tendency towards the belief that there are only two positions you can publicly take: fully support or fully oppose. In reality most people have nuanced positions and faced with a sudden loss of rights might flip from oppose to support. If this is true then looking for cases where liberals want to take away rights should provide a mirror case.

Expand full comment

I think that at least part of the answer requires an *honest* understanding of why the "losing" side is against whatever the issue is. Such honesty is, of course, not exactly common...

In the case of gay rights, I think the explanation is pretty simple. The issue of gayness (being honest now, as opposed to the usual story) I think was primarily viewed by its opponents as set of anti-bourgeois values and morals. And why wouldn't it have been so? In the late 60s/early 70s gayness presented as an all-party all-the-time lifestyle built on endless drugs, zero personal responsibility, and destruction of the family:


The infamous 1972 Gay Men’s Liberation demands looked to most people like a demand for legalized pedophilia and does the usual political yutting thing of throwing in multiple demands that have zero to do with gayness but an awful lot to do with making themselves unpopular with most of America.


Read the above document. THAT is what most people thought they were protecting America from in the 1970s, and who can blame them...

What happened in, I guess the 90s, is that a smart enough group of gay people managed to wrest control of the agenda away from these lunatics and worked hard to ensure that the ONLY issue on the table was gayness. Not gayness and how the military sucks. No gayness and how the family should be abolished. Not gayness and drug legalization. NO OTHER CRAP except gayness and laws related to that. And it turns out that, big fscking surprise, Americans did not have a problem with gayness per se, once it was stripped of the lunacy.

To the extent that other agendas like "racism" or "sexism" win without backlash, it would be by following the same agenda. But it appears that both of these are in too deep in terms of having defined a totalizing world view that is anti-bourgeois and anti most of what Americans support (including such basics as decency, honesty, truth, and rationality).

In a sense I think you have the story backwards. Both race and sex got most of what was reasonable in the late 60s and early 70s, but were not content to take the win and build on that; they created a backlash by refusing to take yes for an answer.

Gay rights stand out as being a rare case in history where common sense prevailed, where the winning group was content to accept its winnings, shut up, and stop fighting. If gay marriage had immediately been followed by other items on that 1972 agenda ("Americans remain as homophobic as ever, until they are willing to destroy that most homophobic and repressive institution of all, the family!!!") yes, there would have been, and would continue to be, massive backlash.

Expand full comment

Probably been said before, but have you considered that trump caused a lingering "rage against conservatives" effect? does the trend to not backlash against liberal victories but to do backlash against conservative ones exist before trump as well?

Expand full comment

The media also found and amplified said story…any victims of liberal policies are not sought out and suppressed when they come forward (detransitioners, COVID vaccine injured people being 2 examples)

Expand full comment

The abortion, medical care, gay marriage issues impact people's lives directly. If the Right supported real policy that would have a positive, realizable benefit on people's lives, you would not see this effect. Since the Right stands for the rights and privilege of a privileged or successful minority, any real policy they propose is not going to have this sort of property.

Trade is an abstract thing, so people have no direct feel for it. Free trade is good for rich people, who control media. Hence media will work to change opinion to oppose what they oppose like Trump's tariffs. To test this idea, look at opinion on how much tax rich people pay and how that is affected (or not affected) by conservative-supported tax cuts.

Expand full comment