An Illustrated Guide to Generational Shifts in Psycho-Political Attitudes

For the first time last year, Millennials outnumbered Boomers among American adults. Already in 2014, they had outnumbered GenXers. This relentless generational change is the point of departure for grappling with the contemporary revolution in psycho-political attitudes and beliefs.

Computed from single age population size using 1980 and 1964 as the cutoffs to define generations.

We have been paying attention to the class gradients in psycho-political attitudes. But, as we shall see, cohort or generational gradients cannot be ignored at all. In fact, generational change provides a neutral theory of the woke counterrevolution. This theory says that people’s attitudes haven’t changed per se. Rather, every year brings a fresh crop of 18-year-olds who are woker than the ones who came before them. And this process, ultimately driven by the machinery of the prestige schools, is an efficient explanation for the present upheaval in psycho-political attitudes.

In this empirical note, we document race, class and generational gradients in politically-relevant beliefs. Ultimately, the goal is to understand the political consequences of this generational shift in psycho-political attitudes. All data is from the 2018 module of the General Social Survey. So the data is very recent, having been collected two years into the Trump presidency. All the response variables we consider are ordinal. We code these as taking integer values. We shall be looking at conditional means to get a handle on race, class and generational gradients in the responses.

We begin with belief in God, where the response ranges from ‘don’t believe’ (0) to ‘know that God exists’ (5). Millennials are dramatically less god-fearing than Boomers or GenXers.

We also note that there is a significant class gradient in faith. The working class is more god-fearing than the middle class. And the pattern extends to the minor classes at the extremes.

Note that we are using self-identified class. In class identification too, we find a generational shift. Millennials and GenXers are much less likely than Boomers to think of themselves as middle class. Whereas 51 percent of Boomers think of themselves as middle class and only 35 percent as working class, the odds are reversed for the younger generations. This speaks directly to the vanishing of the American Dream. Recall also that 44 percent identify as working class and middle class each, while 9 percent identify as lower class and 3.6 percent as upper class. These proportions have been virtually constant since GSS started asking respondents to self-identify their class in 1972.

We formally compute race, class and generational gradients using OLS. We report heteroskedasticity-robust standard errors since we are looking at cross-sectional variation. We find that, relative to white middle-class Boomers, African-Americans and the lower classes are more god-fearing and Millennials are considerably less god-fearing.

How certain are you that God exists?
coefstd errt-StatP
Race[T.BLACK]0.5980.0797.560.000
Class[T.Lower]0.3190.1212.640.008
Class[T.Working]0.2130.0772.760.006
Class[T.Upper]-0.4130.214-1.930.053
Generation[T.GenXers]-0.0080.087-0.090.925
Generation[T.Millennials]-0.6030.087-6.910.000
Adj. R-squared:0.055
F-statistic:20.34
No. Observations:1855
Source: GSS (2018), author’s computations. OLS estimates with robust std errors. Intercept included but not shown. Response is an ordinal variable encoding degree of belief in God.

GSS asked respondents how often they feel disrespected. The younger generations report feeling disrespect considerably more often than Boomers.

There is also a significant class gradient in perceived slight. People who self-identify as working class or lower class report feeling disrespected more often than people who identify as middle class or upper class.

As before, we regress our ordinal variable on race, class and generational dummies. Surprisingly, we find that, relative to white middle-class Boomers, Black respondents do not report higher frequencies of feeling disrespected once we control for class and generation. By contrast, the working class, and especially the lower class, report feeling more disrespected. This pattern goes to the heart of elite-mass relations. The younger generations also report higher rates, suggesting major generational shifts in the propensity to take slight.

How often do you feel disrespected?
coefstd errt-StatP
Race[T.BLACK]0.0150.1190.130.897
Class[T.Lower]0.8900.1924.650.000
Class[T.Working]0.3130.0893.520.000
Class[T.Upper]0.0750.2540.300.768
Generation[T.GenXers]0.4510.1074.220.000
Generation[T.Millennials]0.5470.1035.330.000
Adj. R-squared:0.052
F-statistic:11.7
No. Observations:1246
Source: GSS (2018), author’s computations. OLS estimates with robust std errors. Intercept included but not shown. Response is an ordinal variable encoding frequency of feeling disrespected.

The frequency with which people report being provided poor service also varies strongly by generation. Millennials report receiving shoddy service more often than GenXers, who in turn report receiving shoddy service more often than Boomers.

The class gradient is also robust. People who identify as working class report being poorly served more often than the middle class.

Once we control for race and generations, the gradient for the working class vanishes. Instead, we find that African-Americans, people who identify as lower class, and the younger generations, all report receiving shoddy service more often than white middle-class Boomers.

How often do you get poor service?
coefstd errt-StatP
Race[T.BLACK]0.3760.0963.920.000
Class[T.Lower]0.3220.1522.130.034
Class[T.Working]0.0730.0711.020.308
Class[T.Upper]-0.0230.194-0.120.906
Generation[T.GenXers]0.2180.0862.520.012
Generation[T.Millennials]0.2940.0813.630.000
Adj. R-squared:0.031
F-statistic:7.164
No. Observations:1249
Source: GSS (2018), author’s computations. OLS estimates with robust std errors. Intercept included but not shown. Response is an ordinal variable encoding frequency of receiving poor service.

Yet another variable that encodes perceived slight the frequency of being treated as “not smart.” Millennials report much higher frequencies than GenXers, who in turn report higher frequencies than Boomers.

The class gradient is also significant. The lower classes report being treated as stupid more often than the upper classes.

African-Americans too, report being treated as stupid more often than white middle-class Boomers. GenXers, and especially Millennials, perceive being treated as stupid much more often than Boomers.

How often do people act as if you are not smart?
coefstd errt-StatP
Race[T.BLACK]0.3080.1232.500.013
Class[T.Lower]0.9060.1954.660.000
Class[T.Working]0.1940.0852.280.023
Class[T.Upper]0.2380.2271.050.293
Generation[T.GenXers]0.3200.1003.210.001
Generation[T.Millennials]0.6370.0976.550.000
Adj. R-squared:0.068
F-statistic:15.27
No. Observations:1248
Source: GSS (2018), author’s computations. OLS estimates with robust std errors. Intercept included but not shown. Response is frequency of being considered stupid.

Respondents were also asked how often they feel threatened. Threat perception is dramatically conditioned by generation. Millennials report much higher frequencies of feeling threatened than GenXers, and the latter report higher frequencies than Boomers.

By contrast, the class gradient here is weak. Only the lower class reports a higher frequency of being threatened — probably because it is.

These patterns are reflected in the formal estimates. Relative to white middle-class Boomers, Millennials and GenXers report higher frequencies, as do people who identify as lower class. The explanation for the two is probably different. The lower class may be physically more at risk, whereas the younger generations are quicker to say they feel threatened. Recall the cancellation of Bennet, the Op-Ed page editor at the New York Times. The mob led by Hannah Jones claimed that Senator Cotton’s Op-Ed threatened the safety of Black reporters at the newspaper. That story has been repeated often enough. Perhaps what explains it is this generational shift in threat perceptions.

How often do you feel threatened?
coefstd errt-StatP
Race[T.BLACK]-0.0550.082-0.670.506
Class[T.Lower]0.3910.1572.500.013
Class[T.Working]0.0030.0600.050.957
Class[T.Upper]-0.0510.148-0.350.730
Generation[T.GenXers]0.1610.0712.270.023
Generation[T.Millennials]0.2900.0694.190.000
Adj. R-squared:0.02
F-statistic:4.558
No. Observations:1253
Source: GSS (2018), author’s computations. OLS estimates with robust std errors. Intercept included but not shown. Response is frequency of feeling threatened.

Millennials are dramatically more supportive of affirmative action than the older generations.

But the class gradient in support for racial preference is weak.

Race, however, is extremely salient. Meanwhile, the generational gradient vanishes once we control for class and race. This is astonishing considering the general perception that the white working class is less supportive of affirmative action than the white middle class. The only gradient that does not vanish is of the Black dummy.

How much do you favor affirmative action?
coefstd errt-StatP
Race[T.BLACK]0.9110.1297.050.000
Class[T.Lower]0.2590.1641.580.114
Class[T.Working]-0.0520.090-0.580.565
Class[T.Upper]0.0530.2390.220.823
Generation[T.GenXers]-0.0110.101-0.110.911
Generation[T.Millennials]0.1320.1051.260.210
Adj. R-squared:0.061
F-statistic:10.48
No. Observations:1146
Source: GSS (2018), author’s computations. OLS estimates with robust std errors. Intercept included but not shown. Response is support for affirmative action.

Unlike affirmative action, both class and generational differences are large and significant for what political scientists call “symbolic racism.” The question posed to respondents is as follows.

Irish, Italians, Jewish and many other minorities overcame prejudice and worked their way up. Blacks should do the same without special favors.

Responses to this question are supposed to contain information on racial resentment. But what they actually mine is the working class script of the ‘disciplined self’ identified by Michele Lamont. So we should expect a class gradient for this response.

We document a massive and recent generational shift in attitudes towards race-relations. Millennials are dramatically more opposed to the idea embodied in the “symbolic racism” statement than the older generations. This generational shift in attitudes towards race-relations is the principal driver of what has been called the Great Awokening by Zach Goldberg and others.

The class gradient on the symbolic racism question is robust as expected.

Formally, we find that, relative to white middle-class Boomers, African-Americans and Millennials fall on one side of this question and the lower classes fall on the other. This is the key to the class-partisan realignment and the present confrontation over race-relations.

Do you think that Blacks should work their way up like others?
coefstd errt-StatP
Race[T.BLACK]-0.6870.112-6.130.000
Class[T.Lower]0.2860.1441.990.047
Class[T.Working]0.3000.0823.650.000
Class[T.Upper]-0.2620.244-1.070.283
Generation[T.GenXers]-0.0820.093-0.880.379
Generation[T.Millennials]-0.4470.095-4.730.000
Adj. R-squared:0.063
F-statistic:13.02
No. Observations:1205
Source: GSS (2018), author’s computations. OLS estimates with robust std errors. Intercept included but not shown. Response is degree of agreement with the symbolic racism statement.

What is the political valence of these patterns? Note first that, of all the response variables considered above, only the first, belief in God (r=0.15, P < 0.001), and the last two questions on race-relations, are associated with partisan identity. Support for affirmative action predicts support for Democrats (r=-0.33, P < 0.001), while agreement with the gotcha “symbolic racism” statement predicts support for Republicans (r=0.34, P < 0.001). But even if they don’t contain predictive information, all the differences in psycho-political attitudes we have looked at condition the political affinities in important ways.

What emerges from this analysis is a portrait of generational change. We have documented that Millennials are less god-fearing, quicker to take offense, and much more liberal on race relations than Boomers. GenXers, by contrast, are no less god-fearing and no more liberal on race-relations than Boomers. In terms of propensity to take offense, they are midway between Millennials and Boomers. The real confrontation is thus between Millennials and their Boomer parents. Now that 40 percent of people in the late 20s have moved back in with their parents, we should picture millions of heated debates and the generations getting cross with each other as they watch antiracists battle police officers on American streets.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s