Responding to the discussion start around 21:00 on the Hispanic Red Shift in 2020. It's very interesting that apparently the shift was primarily among Hispanics that already identified as Conservative but had previously voted for Democrats. But I have some holes to poke in David's theory that Socialism and Defund The Police caused the shift:
First, AOC's own district is half Hispanic, and she only underperformed Biden in her district by 1.4 percent, after she spent months shouting from the rooftops that defeating Trump was The Most Important Thing Ever. If mere association with the likes of AOC and socialism and defund the police and so on, despite Biden's repeated denials, caused such a big drop, why didn't we see a much larger drop when it came to AOC herself?
Second, David has said on Twitter that "Hispanic turnout in the [2020 Democratic primary] went down by a lot (which probably was a big warning sign that persuasion was already happening)" https://twitter.com/davidshor/status/1343685995208839176 Obviously, this timeline does not at all fit with Defund The Police being a major cause. Why wasn't this mentioned during the interview?
Third, as to the correlations that they found in surveys with views on crime and socialism. Again around 22:30, David said that what happened in 2020 is that Hispanics who identified as Conservative became more likely to vote Republican. Do these correlations persist after you control for self-identified political ideology? If by 2020 more self-identified Conservative Hispanics were identifying with the GOP and consuming right wing media and taking their political cues from that, isn't this exactly what we would expect to see anyway?
I also note that it seems like it wasn't just Hispanics, nor was it just working class or less educated immigrants. Every Chinatown that I've looked up on NYT's precinct-level 2020 election map shifted red, sometimes by double digits. https://www.nytimes.com/interactive/2021/upshot/2020-election-map.html So did Chicago's Little India. It really seems like immigrant groups in general started voting Republican more this election, and I'd be really interested in seeing if the same pattern as in Hispanics is visible with them, IE self-identified Conservatives became more likely to vote Republican.
What's my theory? In 2004, George W. Bush also won an unusually high percentage of the Hispanic vote. Back in March 2016, David did some modeling that predicted that Rubio would do 10% better than Trump among Hispanics (is that 10% vote share or 10% margin?): https://twitter.com/davidshor/status/1012340380178755584 What do Bush 2004, Rubio 2016, and Trump 2020 have in common that's so important? My guess is that Hispanic voters could trust them not to carry out mass deportations and general anti-immigrant campaigns.
That may sound bizarre when talking about Trump, but if you look past his rhetoric and examine the facts on the ground, deportations were lower during Trump's first term than they were during Obama's first term. https://thehill.com/latino/470900-deportations-lower-under-trump-than-obama-report The stuff that made the news mostly involved Central American migrants, and less that 10% of American Hispanics are of Central American descent. And Trump's wall can be bypassed with a ladder, so the most relevant practical impact of building the wall was that Hispanic border communities got some increased economic activity. https://www.youtube.com/watch?v=ToFk0kwrN9k
Furthermore, during his 2020 campaign Trump largely stopped talking about immigration. When the subject came up during the second Presidential debate, Trump responded by first changing the subject, and then when the moderator brought him back on topic, deflected ("who built the cages, Joe?"). See the transcript: https://www.debates.org/voter-education/debate-transcripts/october-22-2020-debate-transcript/
So my theory is that from 2008-2016, a large number of right-leaning Hispanics and Asians voted for Democrats only out of fear that if Republicans got into power they would deport their relatives, round people like them up for "looking like illegals," as happened in Arizona under Sheriff Joe, and so on. Then when even Bad Orange Man failed to bring this about, they decided that Democrats had been crying wolf about this and voted Republican. The fact that, as David put it, the correlation between self-identified political ideology and voting behavior increased among Hispanics in 2020 would seem to be consistent with this.
Hispanics, more than any other group, supported bringing in the feds to stop the looting and rioting during the summer’s protests. They associate DTP and BLM with democrats, as a result Latinos in Texas/Florida defected to the right. The left continues to believe immigration is important to Hispanic, despite polls showing jobs and healthcare as their top issues.
Its easy to see why Latinos are defecting. Most Latinos (Cuban, Mexico, Puerto Rican) in the USA are mostly of European ancestry (based on genetic tests) and are becoming “white”. They have high marriage rates with whites. Pew released research that 1 in 5 Hispanic children has a white parent. Since they tend not to have college degrees they are starting to vote Republican, like most non college educated whites. Same happened with Italians. The WSJ interviewed Hispanics in Texas and many didn’t not consider themselves POCs. Latinos are more Henry Cuellar and Marco Rubio than AOC.
Prop 16 in CA died, because of the Hispanic vote. Point is, none of its shocking. Aoc’s district has a lot of white progressives.
Still working my way through this, but I want to first respond to, at 38:42, David saying that increasing pro-immigration sentiment was caused by the increasing college-educated share of the population/electorate.
I'm pretty sure that this is completely wrong. Pew has a graph of immigration attitudes over time broken down by generation (https://www.pewresearch.org/wp-content/uploads/2019/01/FT_19.01.29_ImmigrantsBurden_Generationaldifferences.png?resize=420,394), and the trendline for Generation Xers starts increasing around 2010 at the exact same time that it does for Millennials, and tracks the Millennial trendline almost exactly from that point on, even though Gen-Xers (defined in this context as people born from 1965-1980) were long past their college years by then. Boomers and Silents also show the same general post-2010 trend of increasing pro-immigrant sentiment, though they started from a lower base.
The pre-Trump trend (I agree that when Trump came onto the scene there was an additional pro-immigration backlash to his rhetoric and actions, but the trend was already visible before 2015) of increasing pro-immigration sentiment was, in my opinion, almost certainly primarily caused by the end of net illegal immigration with the Great Recession (https://www.pewresearch.org/fact-tank/2019/06/12/5-facts-about-illegal-immigration-in-the-u-s/). There's an inflection point around 2007-2010 in basically every immigration public opinion graph that I've looked at. Opinions changed because the situation changed.
Also, as for the general question of whether a college education makes people more pro-immigration, a longitudinal study in Switzerland tracked people from before they entered college until age 30, and found that people's opinions on immigration stayed basically the same between entering college and graduating, and that highly educated individuals actually became a bit more anti-immigration after they entered the labor market. Quoting the study's abstract: "This suggests that differences between educational groups are mostly due to selection effects, and not to the alleged liberalizing effect of education." https://academic.oup.com/esr/article-abstract/31/4/490/496810
There are similar studies using similar strategies in South Korea and the US that find similar results.
The mechanisms are a bit unclear. I *don’t* think it’s due to colleges indoctrinating people (the Euro study I linked to isn’t even about college! Though the Korean one was). My main guess is that as an age cohort gets better educated, there are both partial and general equilibrium effects on attitudes that are driven by changes in media consumption, partly because middle/high-brow cultural production is produced by very left wing people.
But intuitively, I am just very very skeptical of the selection hypothesis for two reasons:
1) Younger cohorts are way more cosmopolitan than older ones, here and everywhere else in the West, and a very significant fraction of age effects decline when you adjust for education. That wouldn’t by itself be enough, but:
2) Education divides in attitudes are roughly equal in magnitude among both younger and older people. Education levels are much higher among younger cohorts, so if it was just selection then we’d see much smaller college-noncollege gaps in support for younger voters and we do not.
This reminds me a bit of the debate about the impact of schooling. If you took panel data on cognitive performance, it would tell you that college students being smarter is entirely selection (there is actual literal selection on test scores) and that schools don’t make you smarter.
But that’s not true! We know from a variety of quasi-experimental and cross-sectional and experimental studies that school funding boosts cognitive performance, even if most variation in individual cognitive performance is not due to schools.
None of this even attempts to address the fact that ~2009-2014 saw a large increase in pro-immigrant sentiment among generations too old to go to college, so clearly other factors must be at work here. But leaving that aside...
Page 5 of the study that you linked shows that the latest reform studied took place in 1972. I have serious doubts that we can generalize that to the post-Internet era. Maybe marginal students were made more likely by the reform to move to a big city, which made a big difference in the information environment back then.
"My main guess is that as an age cohort gets better educated, there are both partial and general equilibrium effects on attitudes that are driven by changes in media consumption, partly because middle/high-brow cultural production is produced by very left wing people."
"2) Education divides in attitudes are roughly equal in magnitude among both younger and older people. Education levels are much higher among younger cohorts, so if it was just selection then we’d see much smaller college-noncollege gaps in support for younger voters and we do not."
In a vacuum, yes you would expect that, but a lot of other stuff has changed over the past few decades. Suppose, for instance, that a lot of the increase in social liberalism particularly among younger generations over the last few decades was driven by improvements in communications tech. One would expect that households that send their kids to college would tend to get cable TV, then internet, then broadband internet, before households that did not send their kids to college.
And indeed, that's exactly what we see. Click the "Income" and "Education" tabs above the "Who uses the internet" and "Who has home broadband" graphs here: https://www.pewresearch.org/internet/fact-sheet/internet-broadband/ You see pretty large income and education gaps during the period that the internet was being rolled out. In 2004, for instance, 86% of college graduates used the internet and 47% had home broadband, but among adults with only a high school diploma the numbers were 53% and 18%, respectively. As late as 2019, 93% of college graduates had home broadband compared to just 59% of adults with only a high school diploma.
We'd see similar effects if, for whatever reason, poltical polarization in general increased between demographics that are likely to go to college and demographics that are not.
I have my own reason for intuitively favoring these sorts of explanations: if college education was the primary driver of "cosmopolitan" liberal attitudes, then high school students should be comparatively very conservative. Political Tik Tok should resemble a junior version of Fox News or Newsmax. I see no evidence that this is the case, and the one study that I've found that examined the political beliefs of Generation Z, and included a fairly large sample of under-18s, found them to be quite similar to Millennials: https://www.pewresearch.org/social-trends/2019/01/17/generation-z-looks-a-lot-like-millennials-on-key-social-and-political-issues/ 61% of Millennials and 62% of Zoomers believe that increasing racial and ethnic diversity is good for society, for instance.
"This reminds me a bit of the debate about the impact of schooling. If you took panel data on cognitive performance, it would tell you that college students being smarter is entirely selection (there is actual literal selection on test scores) and that schools don’t make you smarter."
Do longitudinal studies really not find that the same individuals score higher on cognitive tests (after taking into account renorming for age and the Flynn Effect and so on) after they graduate than they did before they entered college?
- Do you think there is such a thing as a generational voting block (controlling for all other demographics)? If so, how has this influenced policy in the past and how will it influence policy in the future? For background, I was shocked to see Baby Boomer's made up nearly 40% of eligible voters during the 2000 election cycle: https://www.pewresearch.org/social-trends/2019/01/30/an-early-look-at-the-2020-electorate-2/.... I'm just trying to figure out if I'm reading too much into this data point.
- In regards to class, what is your understanding of the polling on tax cuts and the Iraq War? My understanding is support correlated positively with respect to amount of education and amount of income.
you know someone is interesting if you've heard him interviewed by R. Khan, Julia Galef, Politico, NYMag already and I still watched:) I was a big fan of "Breaking the Two Party Doom Loop." great book that I think is a key idea after seeing what has happened after RVC was implemented in other countries. Had to laugh at this one though: https://www.theatlantic.com/politics/archive/2020/12/democrats-election-ranked-choice-voting-new-york/617461/ You can't please everyone!
I truly enjoy and learn a lot from the discussion. I agree with David on pushing for economic issues that are common concerns to many undecided voters. I am an immigrant and it took me decades to become center left democrat. My friend who is a highly educated immigrant (Phd) did not vote in the 2020 election b/c his family is upset about the leftest leaning agenda and defund policy slogan in the Democratic Party. They are from China and the idea of socialism, eve misperception, is a big turn off for them. Luckily, we are from Massachusetts so the votes from his family did not count. Massachusetts is a solid blue state. It took months before I can talk to them b/c they skipped the most important election in recent history.
Noah, excellent interview with Shor (and I've read quite a few Shor interviews by this time!) He makes a lot of important points in this one that deserve wide circulation. I wonder if you have thought about posting a transcript of the conversation--that might help bring the very enlightening discussion to more people. Btw, great job with the substack in general; I consistently find that your newsletters provide useful analysis and data about stuff I'm interested in. Keep up the good work!
I think he really hits on something when he starts talking more broadly (on context of the Hispanic shift) about the working class/educated divide. Remember: it used to be that the working class was not a good fit at all for Republicans. They were seen as “fat cats” who didn’t give a shit. When you thought about the Republicans you knew in real life, perhaps as a working class person you’d think of the big dog bosses at your company who were too good to talk to you or ever really acknowledge you.
Now.... my how things have changed, and I think Trump helped accelerate the cultural shift that was already happening with the working class. Look, here’s the deal: I know lots of working class people, and if you want them to look at you like you have two heads tell them how great of a President you think Hilary Clinton would have been. You’ll get laughed at probably 9 times out of 10. And since Hispanics are much more likely to be working class vs whites in this country, and the churches they attend also don’t have the traditional Dem backing that blacks get, Hispanics are getting it: they increasingly have more in common with working class GOP voters than the college educated Dems. This is going to be tough to be reverse I think, but you never in politics. Things can change quickly with the right candidates/message.
Highly educated immigrant groups also moved right in 2020 Look up any Chinatown on NYT's precinct-level election 2020 map: https://www.nytimes.com/interactive/2021/upshot/2020-election-map.html One precinct in San Francisco's Chinatown shifted its vote margin right by 22 points. The margin in Chicago's "Little India" shifted right by 11 points.
Knowing the geography of the Chinese communities of NYC quite well, I do not think this wonderful detailed maps shows conclusively what you say it does. About the same share of Chinese neighborhoods voted for Trump in both elections (across all neighborhoods). This is another mystery worth unpacking. Maybe Chinese origin people living in NYC liked Trump's anti-Chinese government stances.
I think some of this highlights the more practical view of education among certain immigrant groups compared to whites. A greater % of immigrants I would guess view education as more of a tool to use to do well in life financially. The Tiger mom phenomena basically. Whereas whites tend to be more holistic in their views on education, in my experience. I realize what I just said is a broad generalization but this whole topic lends itself to broad generalizations about a lot of things.
The U.S. is an enormously geographically sprawling and culturally diverse nation. Your ideas about deconstructing our system of geographic representation to one base purely on population density, as a way to lord over people out in the hinterlands with different beliefs, make you real enemies to be crushed.
When you talk about the education divide, it sounds very condescending, like you mean people vote left BECAUSE they are educated. You never mention the very genuine economic pain of wage suppression and hollowing out of manufacturing due to (a) importing low wage Hispanic immigrants, or (b) farming out all our jobs to Chinese forced-labor camps.
Conservative here scouting the enemy. Shor's points on messaging seem utterly obvious common sense. Man, you guys are so far left, you really need your own zone.
A bit late to the party here, but this explanation for Bernie's loss isn't all that compelling and sounds more like a media spot for the poll-driven politics that Shor loves and benefits from... he's just rehashing the Angela Nagle/Michael Tracey bait piece published in the conservative American Affairs journal last year.
What was clear from the outset was that Bernie benefitted from a divided field, won the first three primaries in states with large White and Hispanic populations and floundered when he got to South Carolina, the first state with a large Black voter base that he'd lost by large margins in 2016. The rest of the Dems use this performance to rally behind Biden, a plague breaks out, and the rest became history. The key would've been to maintain a divided field, which in retrospect, hinged on a stronger performance in the South and with Black voters, and overcoming a skeptical South that firmly associated Biden with Obama and looked to stalwarts like Clyburn for last-minute direction. For all the optics and talk, the campaign failed to adequately invest, campaign, and prioritize the South and use Biden's weak Iowa performance against him. The "white working class" at this point had already defected to Trump in 2016 after Bernie's 2016 failure and didn't show up to vote in the Dem primaries, and instead we got a sea of moderate, richer conservative Republicans voting blue filling in this vacuum that broke in Biden's favor. I agree with the bread and butter approach but the electorate changed in some meaningful ways this cycle
Re: Starting around 46:00 the discussion about increasing support for gay rights from the early 90s onward. Besides the AP column that David mentioned, another thing that also happened in 1994 was the appearance of Pedro Zamora on the third season of MTV's The Real World. Quoting Wikipedia:
"As the show increased in popularity, Zamora's life as someone living with AIDS gained considerable notice, garnering widespread media attention. Zamora was one of the first openly gay men with AIDS to be portrayed in popular media, and after his death on November 11, 1994 (mere hours after the final episode of his season aired), he was lauded by then-President Bill Clinton. [...] Zamora's conflicts with Rainey were not only considered emotional high points for that season, but are credited with making The Real World a hit show, and with proving that the infant "reality" television format was one that could bring considerable ratings to a network."
It seems like stuff like that might have mattered, especially since unlike a fictional TV series this was presented as something real. The GSS data, at least, does seem to show a big increase in acceptance of homosexuality among the "MTV generation" (born 1970-1979) right around that time: https://twitter.com/xenocryptsite/status/1106566354939846656/photo/1
Zooming out a bit, something that might be worth looking into is the 1992 Cable Act (https://en.wikipedia.org/wiki/Cable_Television_Consumer_Protection_and_Competition_Act_of_1992), and whether that had a significant impact on the content of paid TV. I note that 1992 is the same year that Bruce Springsteen described Cable and Satellite TV as "57 channels and nothing on." Maybe that law did something to change that, by encouraging competition somehow, and perhaps in the process expanded the "effective bandwidth" of the television medium?
The pattern is pretty clear: from 1993 to 1998 the increase in acceptance of homosexuality was greater among teenagers and twenty-somethings than it was among people aged 30 and up. I didn't have much relevant direct experience of that period, since I was in elementary school, but I doubt that young people read more newspapers than older people back then. So if you want to answer the question of why this happened, I think you should be looking at media that was targeted at younger demographics - the aforementioned MTV, cable channels with a similar target demographic, maybe certain magazines that were popular with the younger crowd at the time, etc.
We stan Geogism
Hell yeah
Responding to the discussion start around 21:00 on the Hispanic Red Shift in 2020. It's very interesting that apparently the shift was primarily among Hispanics that already identified as Conservative but had previously voted for Democrats. But I have some holes to poke in David's theory that Socialism and Defund The Police caused the shift:
First, AOC's own district is half Hispanic, and she only underperformed Biden in her district by 1.4 percent, after she spent months shouting from the rooftops that defeating Trump was The Most Important Thing Ever. If mere association with the likes of AOC and socialism and defund the police and so on, despite Biden's repeated denials, caused such a big drop, why didn't we see a much larger drop when it came to AOC herself?
Second, David has said on Twitter that "Hispanic turnout in the [2020 Democratic primary] went down by a lot (which probably was a big warning sign that persuasion was already happening)" https://twitter.com/davidshor/status/1343685995208839176 Obviously, this timeline does not at all fit with Defund The Police being a major cause. Why wasn't this mentioned during the interview?
Third, as to the correlations that they found in surveys with views on crime and socialism. Again around 22:30, David said that what happened in 2020 is that Hispanics who identified as Conservative became more likely to vote Republican. Do these correlations persist after you control for self-identified political ideology? If by 2020 more self-identified Conservative Hispanics were identifying with the GOP and consuming right wing media and taking their political cues from that, isn't this exactly what we would expect to see anyway?
I also note that it seems like it wasn't just Hispanics, nor was it just working class or less educated immigrants. Every Chinatown that I've looked up on NYT's precinct-level 2020 election map shifted red, sometimes by double digits. https://www.nytimes.com/interactive/2021/upshot/2020-election-map.html So did Chicago's Little India. It really seems like immigrant groups in general started voting Republican more this election, and I'd be really interested in seeing if the same pattern as in Hispanics is visible with them, IE self-identified Conservatives became more likely to vote Republican.
What's my theory? In 2004, George W. Bush also won an unusually high percentage of the Hispanic vote. Back in March 2016, David did some modeling that predicted that Rubio would do 10% better than Trump among Hispanics (is that 10% vote share or 10% margin?): https://twitter.com/davidshor/status/1012340380178755584 What do Bush 2004, Rubio 2016, and Trump 2020 have in common that's so important? My guess is that Hispanic voters could trust them not to carry out mass deportations and general anti-immigrant campaigns.
That may sound bizarre when talking about Trump, but if you look past his rhetoric and examine the facts on the ground, deportations were lower during Trump's first term than they were during Obama's first term. https://thehill.com/latino/470900-deportations-lower-under-trump-than-obama-report The stuff that made the news mostly involved Central American migrants, and less that 10% of American Hispanics are of Central American descent. And Trump's wall can be bypassed with a ladder, so the most relevant practical impact of building the wall was that Hispanic border communities got some increased economic activity. https://www.youtube.com/watch?v=ToFk0kwrN9k
Furthermore, during his 2020 campaign Trump largely stopped talking about immigration. When the subject came up during the second Presidential debate, Trump responded by first changing the subject, and then when the moderator brought him back on topic, deflected ("who built the cages, Joe?"). See the transcript: https://www.debates.org/voter-education/debate-transcripts/october-22-2020-debate-transcript/
So my theory is that from 2008-2016, a large number of right-leaning Hispanics and Asians voted for Democrats only out of fear that if Republicans got into power they would deport their relatives, round people like them up for "looking like illegals," as happened in Arizona under Sheriff Joe, and so on. Then when even Bad Orange Man failed to bring this about, they decided that Democrats had been crying wolf about this and voted Republican. The fact that, as David put it, the correlation between self-identified political ideology and voting behavior increased among Hispanics in 2020 would seem to be consistent with this.
Hispanics, more than any other group, supported bringing in the feds to stop the looting and rioting during the summer’s protests. They associate DTP and BLM with democrats, as a result Latinos in Texas/Florida defected to the right. The left continues to believe immigration is important to Hispanic, despite polls showing jobs and healthcare as their top issues.
Its easy to see why Latinos are defecting. Most Latinos (Cuban, Mexico, Puerto Rican) in the USA are mostly of European ancestry (based on genetic tests) and are becoming “white”. They have high marriage rates with whites. Pew released research that 1 in 5 Hispanic children has a white parent. Since they tend not to have college degrees they are starting to vote Republican, like most non college educated whites. Same happened with Italians. The WSJ interviewed Hispanics in Texas and many didn’t not consider themselves POCs. Latinos are more Henry Cuellar and Marco Rubio than AOC.
Prop 16 in CA died, because of the Hispanic vote. Point is, none of its shocking. Aoc’s district has a lot of white progressives.
https://www.pewresearch.org/fact-tank/2020/09/11/hispanic-voters-say-economy-health-care-and-covid-19-are-top-issues-in-2020-presidential-election/
https://www.pewresearch.org/fact-tank/2017/06/06/the-rise-of-multiracial-and-multiethnic-babies-in-the-u-s/
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4289685/
Still working my way through this, but I want to first respond to, at 38:42, David saying that increasing pro-immigration sentiment was caused by the increasing college-educated share of the population/electorate.
I'm pretty sure that this is completely wrong. Pew has a graph of immigration attitudes over time broken down by generation (https://www.pewresearch.org/wp-content/uploads/2019/01/FT_19.01.29_ImmigrantsBurden_Generationaldifferences.png?resize=420,394), and the trendline for Generation Xers starts increasing around 2010 at the exact same time that it does for Millennials, and tracks the Millennial trendline almost exactly from that point on, even though Gen-Xers (defined in this context as people born from 1965-1980) were long past their college years by then. Boomers and Silents also show the same general post-2010 trend of increasing pro-immigrant sentiment, though they started from a lower base.
The pre-Trump trend (I agree that when Trump came onto the scene there was an additional pro-immigration backlash to his rhetoric and actions, but the trend was already visible before 2015) of increasing pro-immigration sentiment was, in my opinion, almost certainly primarily caused by the end of net illegal immigration with the Great Recession (https://www.pewresearch.org/fact-tank/2019/06/12/5-facts-about-illegal-immigration-in-the-u-s/). There's an inflection point around 2007-2010 in basically every immigration public opinion graph that I've looked at. Opinions changed because the situation changed.
Also, as for the general question of whether a college education makes people more pro-immigration, a longitudinal study in Switzerland tracked people from before they entered college until age 30, and found that people's opinions on immigration stayed basically the same between entering college and graduating, and that highly educated individuals actually became a bit more anti-immigration after they entered the labor market. Quoting the study's abstract: "This suggests that differences between educational groups are mostly due to selection effects, and not to the alleged liberalizing effect of education." https://academic.oup.com/esr/article-abstract/31/4/490/496810
Similar longitudinal studies in the US have found similar results: https://www.economist.com/graphic-detail/2020/01/09/are-left-wing-american-professors-indoctrinating-their-students
Here’s my go to on the education/cosmopolitanism link being causal: https://twitter.com/davidshor/status/1035985692529225728?s=21
There are similar studies using similar strategies in South Korea and the US that find similar results.
The mechanisms are a bit unclear. I *don’t* think it’s due to colleges indoctrinating people (the Euro study I linked to isn’t even about college! Though the Korean one was). My main guess is that as an age cohort gets better educated, there are both partial and general equilibrium effects on attitudes that are driven by changes in media consumption, partly because middle/high-brow cultural production is produced by very left wing people.
But intuitively, I am just very very skeptical of the selection hypothesis for two reasons:
1) Younger cohorts are way more cosmopolitan than older ones, here and everywhere else in the West, and a very significant fraction of age effects decline when you adjust for education. That wouldn’t by itself be enough, but:
2) Education divides in attitudes are roughly equal in magnitude among both younger and older people. Education levels are much higher among younger cohorts, so if it was just selection then we’d see much smaller college-noncollege gaps in support for younger voters and we do not.
This reminds me a bit of the debate about the impact of schooling. If you took panel data on cognitive performance, it would tell you that college students being smarter is entirely selection (there is actual literal selection on test scores) and that schools don’t make you smarter.
But that’s not true! We know from a variety of quasi-experimental and cross-sectional and experimental studies that school funding boosts cognitive performance, even if most variation in individual cognitive performance is not due to schools.
None of this even attempts to address the fact that ~2009-2014 saw a large increase in pro-immigrant sentiment among generations too old to go to college, so clearly other factors must be at work here. But leaving that aside...
Page 5 of the study that you linked shows that the latest reform studied took place in 1972. I have serious doubts that we can generalize that to the post-Internet era. Maybe marginal students were made more likely by the reform to move to a big city, which made a big difference in the information environment back then.
"My main guess is that as an age cohort gets better educated, there are both partial and general equilibrium effects on attitudes that are driven by changes in media consumption, partly because middle/high-brow cultural production is produced by very left wing people."
You can read a PDF of the Swiss study here: https://pdfs.semanticscholar.org/cc4c/f0f75aa9d6e9fbfd74ce12e439525598077b.pdf?_ga=2.147674036.2054985278.1608707216-1414935208.1607114379 The graph on page 7 shows that the study followed subjects all the way up until the age of 30, and found no statistically significant individual-level effects.
"2) Education divides in attitudes are roughly equal in magnitude among both younger and older people. Education levels are much higher among younger cohorts, so if it was just selection then we’d see much smaller college-noncollege gaps in support for younger voters and we do not."
In a vacuum, yes you would expect that, but a lot of other stuff has changed over the past few decades. Suppose, for instance, that a lot of the increase in social liberalism particularly among younger generations over the last few decades was driven by improvements in communications tech. One would expect that households that send their kids to college would tend to get cable TV, then internet, then broadband internet, before households that did not send their kids to college.
And indeed, that's exactly what we see. Click the "Income" and "Education" tabs above the "Who uses the internet" and "Who has home broadband" graphs here: https://www.pewresearch.org/internet/fact-sheet/internet-broadband/ You see pretty large income and education gaps during the period that the internet was being rolled out. In 2004, for instance, 86% of college graduates used the internet and 47% had home broadband, but among adults with only a high school diploma the numbers were 53% and 18%, respectively. As late as 2019, 93% of college graduates had home broadband compared to just 59% of adults with only a high school diploma.
We'd see similar effects if, for whatever reason, poltical polarization in general increased between demographics that are likely to go to college and demographics that are not.
I have my own reason for intuitively favoring these sorts of explanations: if college education was the primary driver of "cosmopolitan" liberal attitudes, then high school students should be comparatively very conservative. Political Tik Tok should resemble a junior version of Fox News or Newsmax. I see no evidence that this is the case, and the one study that I've found that examined the political beliefs of Generation Z, and included a fairly large sample of under-18s, found them to be quite similar to Millennials: https://www.pewresearch.org/social-trends/2019/01/17/generation-z-looks-a-lot-like-millennials-on-key-social-and-political-issues/ 61% of Millennials and 62% of Zoomers believe that increasing racial and ethnic diversity is good for society, for instance.
"This reminds me a bit of the debate about the impact of schooling. If you took panel data on cognitive performance, it would tell you that college students being smarter is entirely selection (there is actual literal selection on test scores) and that schools don’t make you smarter."
Do longitudinal studies really not find that the same individuals score higher on cognitive tests (after taking into account renorming for age and the Flynn Effect and so on) after they graduate than they did before they entered college?
Correction: actually the last reform studied took place in 1974, but the general point stands.
Three things for David Shor (assuming he is still looking at these comments):
- Do you have any opinion/critique of Steven Bram's ideal voting system: https://www.youtube.com/watch?v=W9Q0azZOyww (19:45 mark) ?
- Do you think there is such a thing as a generational voting block (controlling for all other demographics)? If so, how has this influenced policy in the past and how will it influence policy in the future? For background, I was shocked to see Baby Boomer's made up nearly 40% of eligible voters during the 2000 election cycle: https://www.pewresearch.org/social-trends/2019/01/30/an-early-look-at-the-2020-electorate-2/.... I'm just trying to figure out if I'm reading too much into this data point.
- In regards to class, what is your understanding of the polling on tax cuts and the Iraq War? My understanding is support correlated positively with respect to amount of education and amount of income.
you know someone is interesting if you've heard him interviewed by R. Khan, Julia Galef, Politico, NYMag already and I still watched:) I was a big fan of "Breaking the Two Party Doom Loop." great book that I think is a key idea after seeing what has happened after RVC was implemented in other countries. Had to laugh at this one though: https://www.theatlantic.com/politics/archive/2020/12/democrats-election-ranked-choice-voting-new-york/617461/ You can't please everyone!
I truly enjoy and learn a lot from the discussion. I agree with David on pushing for economic issues that are common concerns to many undecided voters. I am an immigrant and it took me decades to become center left democrat. My friend who is a highly educated immigrant (Phd) did not vote in the 2020 election b/c his family is upset about the leftest leaning agenda and defund policy slogan in the Democratic Party. They are from China and the idea of socialism, eve misperception, is a big turn off for them. Luckily, we are from Massachusetts so the votes from his family did not count. Massachusetts is a solid blue state. It took months before I can talk to them b/c they skipped the most important election in recent history.
Noah, excellent interview with Shor (and I've read quite a few Shor interviews by this time!) He makes a lot of important points in this one that deserve wide circulation. I wonder if you have thought about posting a transcript of the conversation--that might help bring the very enlightening discussion to more people. Btw, great job with the substack in general; I consistently find that your newsletters provide useful analysis and data about stuff I'm interested in. Keep up the good work!
Any chance you could upload the audio of this to your Hexapodia podcast channel so I could listen on Apple Podcasts?
I think he really hits on something when he starts talking more broadly (on context of the Hispanic shift) about the working class/educated divide. Remember: it used to be that the working class was not a good fit at all for Republicans. They were seen as “fat cats” who didn’t give a shit. When you thought about the Republicans you knew in real life, perhaps as a working class person you’d think of the big dog bosses at your company who were too good to talk to you or ever really acknowledge you.
Now.... my how things have changed, and I think Trump helped accelerate the cultural shift that was already happening with the working class. Look, here’s the deal: I know lots of working class people, and if you want them to look at you like you have two heads tell them how great of a President you think Hilary Clinton would have been. You’ll get laughed at probably 9 times out of 10. And since Hispanics are much more likely to be working class vs whites in this country, and the churches they attend also don’t have the traditional Dem backing that blacks get, Hispanics are getting it: they increasingly have more in common with working class GOP voters than the college educated Dems. This is going to be tough to be reverse I think, but you never in politics. Things can change quickly with the right candidates/message.
Highly educated immigrant groups also moved right in 2020 Look up any Chinatown on NYT's precinct-level election 2020 map: https://www.nytimes.com/interactive/2021/upshot/2020-election-map.html One precinct in San Francisco's Chinatown shifted its vote margin right by 22 points. The margin in Chicago's "Little India" shifted right by 11 points.
Knowing the geography of the Chinese communities of NYC quite well, I do not think this wonderful detailed maps shows conclusively what you say it does. About the same share of Chinese neighborhoods voted for Trump in both elections (across all neighborhoods). This is another mystery worth unpacking. Maybe Chinese origin people living in NYC liked Trump's anti-Chinese government stances.
I think some of this highlights the more practical view of education among certain immigrant groups compared to whites. A greater % of immigrants I would guess view education as more of a tool to use to do well in life financially. The Tiger mom phenomena basically. Whereas whites tend to be more holistic in their views on education, in my experience. I realize what I just said is a broad generalization but this whole topic lends itself to broad generalizations about a lot of things.
Great discussion!! Anxiously awaiting part II.
The U.S. is an enormously geographically sprawling and culturally diverse nation. Your ideas about deconstructing our system of geographic representation to one base purely on population density, as a way to lord over people out in the hinterlands with different beliefs, make you real enemies to be crushed.
When you talk about the education divide, it sounds very condescending, like you mean people vote left BECAUSE they are educated. You never mention the very genuine economic pain of wage suppression and hollowing out of manufacturing due to (a) importing low wage Hispanic immigrants, or (b) farming out all our jobs to Chinese forced-labor camps.
Conservative here scouting the enemy. Shor's points on messaging seem utterly obvious common sense. Man, you guys are so far left, you really need your own zone.
A bit late to the party here, but this explanation for Bernie's loss isn't all that compelling and sounds more like a media spot for the poll-driven politics that Shor loves and benefits from... he's just rehashing the Angela Nagle/Michael Tracey bait piece published in the conservative American Affairs journal last year.
What was clear from the outset was that Bernie benefitted from a divided field, won the first three primaries in states with large White and Hispanic populations and floundered when he got to South Carolina, the first state with a large Black voter base that he'd lost by large margins in 2016. The rest of the Dems use this performance to rally behind Biden, a plague breaks out, and the rest became history. The key would've been to maintain a divided field, which in retrospect, hinged on a stronger performance in the South and with Black voters, and overcoming a skeptical South that firmly associated Biden with Obama and looked to stalwarts like Clyburn for last-minute direction. For all the optics and talk, the campaign failed to adequately invest, campaign, and prioritize the South and use Biden's weak Iowa performance against him. The "white working class" at this point had already defected to Trump in 2016 after Bernie's 2016 failure and didn't show up to vote in the Dem primaries, and instead we got a sea of moderate, richer conservative Republicans voting blue filling in this vacuum that broke in Biden's favor. I agree with the bread and butter approach but the electorate changed in some meaningful ways this cycle
Re: Starting around 46:00 the discussion about increasing support for gay rights from the early 90s onward. Besides the AP column that David mentioned, another thing that also happened in 1994 was the appearance of Pedro Zamora on the third season of MTV's The Real World. Quoting Wikipedia:
"As the show increased in popularity, Zamora's life as someone living with AIDS gained considerable notice, garnering widespread media attention. Zamora was one of the first openly gay men with AIDS to be portrayed in popular media, and after his death on November 11, 1994 (mere hours after the final episode of his season aired), he was lauded by then-President Bill Clinton. [...] Zamora's conflicts with Rainey were not only considered emotional high points for that season, but are credited with making The Real World a hit show, and with proving that the infant "reality" television format was one that could bring considerable ratings to a network."
https://en.wikipedia.org/wiki/The_Real_World_(TV_series)#History
It seems like stuff like that might have mattered, especially since unlike a fictional TV series this was presented as something real. The GSS data, at least, does seem to show a big increase in acceptance of homosexuality among the "MTV generation" (born 1970-1979) right around that time: https://twitter.com/xenocryptsite/status/1106566354939846656/photo/1
Zooming out a bit, something that might be worth looking into is the 1992 Cable Act (https://en.wikipedia.org/wiki/Cable_Television_Consumer_Protection_and_Competition_Act_of_1992), and whether that had a significant impact on the content of paid TV. I note that 1992 is the same year that Bruce Springsteen described Cable and Satellite TV as "57 channels and nothing on." Maybe that law did something to change that, by encouraging competition somehow, and perhaps in the process expanded the "effective bandwidth" of the television medium?
Here's GSS data broken down by year of birth in a more readable table format: https://twitter.com/xenocryptsite/status/1106570403084976128/photo/1
The pattern is pretty clear: from 1993 to 1998 the increase in acceptance of homosexuality was greater among teenagers and twenty-somethings than it was among people aged 30 and up. I didn't have much relevant direct experience of that period, since I was in elementary school, but I doubt that young people read more newspapers than older people back then. So if you want to answer the question of why this happened, I think you should be looking at media that was targeted at younger demographics - the aforementioned MTV, cable channels with a similar target demographic, maybe certain magazines that were popular with the younger crowd at the time, etc.