The Anxious Liberal — Ideology, Age & Well-Being

The long topline

Generalized happiness in America was remarkably stable for forty years. Across the 1972–2018 GSS waves, the share saying they were very happy sat between 28 and 36 percent, with no clear trend. Then 2021 happened. The pandemic-era reading was 19.4%, the lowest in the series. By 2024 it had partially recovered to 23.6%, still well below the long-run average.

That topline is the only number in this chapter that is uncontroversial. Whether the recent decline is a real attitudinal shift, a mode-effect artifact, a pandemic scar, or some combination of all three is genuinely contested. What is less contested — and is the focus of this chapter — is that the decline has not landed evenly. It has landed harder on the young, and it has landed harder on Liberals.

The ideological gap, three ways

Republicans and Conservatives have reported higher happiness than Democrats and Liberals since the question was first asked. In the 1974 GSS, 42% of Conservatives said they were very happy, against 33% of Liberals — a 9-point gap. Through the 1980s, 1990s, and 2000s, the gap stayed close to that level: usually 5 to 10 points, occasionally narrowing. The proposal that became this book described it, accurately, as a "modest happiness gap."

What changed in the 2010s is the gap's shape. Between 2016 and 2021 it widened sharply — and although it narrowed somewhat in 2022 and 2024, it did not return to the historical baseline.

Three things are worth flagging in Figure 2.

First, the Liberal series falls earliest and falls furthest in the recent leg. The 2021 reading for Liberals is 14.4% — versus 23.4% for Conservatives, a 9-point gap, the largest in the series. The Conservative line also fell, but less, and from a higher base.

Second, the direction is consistent with the original Gallup chart: Liberals down more than Conservatives. The magnitude is different — happiness on GSS dropped roughly 15 points for Liberals between 2016 and 2021, while the Gallup mental-health series in the original chart dropped almost 30 points for the youngest Liberal panel between 2013 and 2023. Some of the magnitude difference is the variable (happiness vs. self-rated mental health "Excellent"), some is the population (all Liberals vs. only 18–29 Liberals), and some is the years (a 2016–2021 leg vs. a longer 2013–2023 arc).

Third, the gap partially closes in the 2022 and 2024 waves. By 2024, Liberals are at 21.2% very happy, Conservatives at 26.0% — a 5-point gap, almost exactly the 1970s baseline. Whether the 2021 spike will persist as a structural feature or fade as pandemic conditions normalize is, on this evidence, an open question.

Two parallel measures show the same direction at smaller magnitudes. Life is exciting (life_rec) — the GSS's "is life dull, routine, or exciting" question — and self-rated poor health show a softer version of the same pattern: Liberals diverge from Conservatives in the recent waves, but the gap is smaller and the timing is less sharp.

The triangulation is partial but pointed. Happiness moves the most. Life-is-exciting moves with it but with smaller amplitude. Self-rated poor health barely registers a partisan gap and doesn't widen — suggesting the divergence is in the subjective-affective register, not in physical health.

A fourth measure goes the opposite way. Financial satisfaction (satfin_rec) — share saying they are satisfied with their financial situation — shows Liberals at parity or slightly above Conservatives in the recent waves, the reverse of happiness. In 2024, 29.0% of Liberals report financial satisfaction against 25.7% of Conservatives. We will return to this contrast at the end of the chapter; it is the central evidence against any reading that says the happiness gap is "really" an income gap.

The age cohort

The other half of the original Gallup chart is age. Within each ideological group, the youngest cohort had moved more than the older ones. In the GSS, the same axis is observable separately, and the same pattern is present.

For most of the GSS's history, age was a weak predictor of happiness. The youngest adults (18–34) and the oldest (65+) reported similar rates of being "very happy," usually within 3 points of each other. The middle-aged groups were not noticeably different.

In 2024, 18–34-year-olds are at 21.0% very happy — versus 28.1% for those 65 and older. A 7-point gap that did not exist before 2018. This is the GSS confirming, on its own data, what the Gallup chart claimed for mental health: the youngest are reporting lower well-being than the oldest, and the divergence is recent.

The same age pattern shows up faintly on life_rec and barely at all on health_rec, consistent with the ideological pattern. Where well-being is most subjective — happiness, life-as-exciting — the youth decline is largest. Where well-being has a stronger physiological floor — self-reported poor health — the gap closes.

Replicating the Gallup chart directly

The original Gallup chart's defining shape was the interaction of age and ideology — Liberals 18 to 29 declining most. The marginal evidence above is consistent with the interaction; running the three-way directly on the raw GSS microdata pins it down. Below is the same chart shape — % very happy by year, twelve cells of age × ideology — pooling 1975 through 2024 GSS waves, weighted by wtssps with strata-and-PSU variance.

The interaction is real but more layered than the Gallup chart implies. Three things stand out from the cell-by-cell run.

The 18–34 Liberal cell shows the largest post-2018 shift, but not a long arc of decline. Pre-2018 the line tracks roughly 28% — close to the population mean for the entire post-1975 series. Then 2021 hits 10.5% (the post-COVID, post-mode-shift wave), 2022 partially recovers to 20.4%, and 2024 sits at 18.2%. The pre-2018 to post-2018 shift is 9 points; the lasting "structural break" magnitude is closer to that 9-point figure than to the 28-point peak-to-trough drop the 2021 reading suggests in isolation.

Older Conservatives also fell. The 65+ Conservative cell — the comparison group that was supposed to be "stable" in the Gallup mental-health data — moved from a 1990s mean of 40.8% to a 2018-2024 mean of 31.4%, a 9.4-point decline of its own. The cell with the biggest recent drop is young Liberals, but the broader "very happy" decline is not unique to them. The 2021 wave dragged everyone down; the post-2021 recovery is partial across all twelve cells.

The relative gap between young Liberals and old Conservatives is roughly where it was. In the 1980s and 1990s, 65+ Conservatives averaged about 14 points above 18-34 Liberals on "very happy" (with high year-to-year noise). In 2024, the same gap is about 10 points. What changed is that both cells dropped — and the youngest-Liberal floor is now visibly lower than it has been at any point in the prior fifty years, even as the Conservative-elder ceiling has dropped from a 1990s average near 41% to about 28% in 2024.

That third point is the most important caveat to the Gallup interpretation. The chart that went viral framed the youngest Liberal cohort's decline as a story unique to that cell. On GSS happiness data — the longer, harder-counted instrument — the decline is concentrated there, but it is not exclusive to it. The right framing is: a broad-based well-being decline that has hit the young harder, and within the young, the Liberals harder still.

What the literature says

The post-2010 ideological well-being gap has accumulated a small but rapidly growing academic literature, much of it published in the last five years. The findings are broadly consistent with what we see above — and they help adjudicate which of several mechanisms might be doing the work.

The empirical anchor. Catherine Gimbrone and colleagues at Columbia, working with Monitoring the Future data on 12th graders from 2005 to 2018, established the modern version of the finding: depressive affect rose for all adolescents after 2012, but rose substantially faster for those identifying as liberal — particularly liberal girls with college-educated parents. The paper appears in SSM – Mental Health, an Elsevier peer-reviewed open-access journal, and is the closest thing to a definitive empirical paper on the topic.

Internalizing-symptom trends in U.S. adolescents diverged by political beliefs after 2010, with liberal-identifying youth reporting the largest increases.

— Gimbrone, Bates, Prins, & Keyes, The Politics of Depression (2022)

The mental-health-specific extension. Zach Goldberg's 2025 Manhattan Institute report, Mental-Health Trends and the "Great Awokening," extends Gimbrone's finding to adults using Pew, MTF, ANES, and NSDUH data through the 2020s. Goldberg's headline finding is that self-identified liberal women under 30 show the largest post-2012 rise in self-reported diagnosed mental illness — the cell-level claim the Gallup chart visualizes. The Manhattan Institute is a right-of-center think tank, so the framing is openly contested, but the underlying time series Goldberg charts are publicly accessible and reproduce in the GSS happiness data above.

The framing books. Jonathan Haidt's The Anxious Generation (2024) frames the youth decline as a story about "phone-based childhood" — heavy social-media use displacing in-person interaction during a critical developmental period — with disproportionate impact on adolescent girls. Greg Lukianoff and Haidt's earlier Coddling of the American Mind (2018) had argued that progressive campus norms encourage "cognitive distortions" — catastrophizing, black-and-white thinking — that overlap with the cognitive style associated with anxiety and depression. Jean Twenge's iGen (2017), built directly on GSS-style panel comparisons, is the canonical statement of the "born after 1995" cohort thesis. None of these are peer-reviewed; they synthesize peer-reviewed work and provide the dominant interpretive frame.

The credibility caveat. The first GSS-specific paper on this topic was Emil Kirkegaard's "Mental Illness and the Left," published in 2020 in Mankind Quarterly — a journal widely characterized in the academic literature as a fringe venue with race-science-adjacent editorial history. We mention the paper because it is sometimes cited online and the empirical claim (an ideological gradient on GSS mental-health proxies) is real, but we do not draw on it for this chapter and recommend readers who encounter the citation flag the venue concern. The same author had a coauthored 2024 paper retracted from Scandinavian Journal of Psychology in 2025.

The mechanism literature. Several non-exclusive mechanisms have been proposed:

• Catastrophizing and cognitive distortion (Haidt; Lukianoff & Haidt, 2018). Liberals are increasingly socialized into a worldview that emphasizes systemic injustice and existential threat (climate, democracy). These framings overlap with cognitive distortions clinically associated with anxiety and depression. The mechanism is psychological.
• External locus of control. Conservatives more often report internal locus of control — believing they can shape their own lives. Liberal worldviews more often emphasize external structural forces. Across the psychology literature, internal locus correlates with subjective well-being. The mechanism is dispositional.
• Political news engagement as chronic stressor (al-Gharbi, We Have Never Been Woke, 2024). Highly-engaged news consumers show elevated stress markers; news-cycle engagement skews liberal in the modern American media environment.
• Differential reporting and stigma. Conservatives may underreport mental-health symptoms due to stigma; Liberals may overreport because they are more comfortable framing distress in clinical terms. This mechanism predicts the gap is partly an artifact of self-report mode.

The GSS data above cannot adjudicate between these. The financial-satisfaction reversal (Liberals slightly higher in 2024) is more pointed. If the mechanism were primarily about Liberals being economically worse off, financial satisfaction should track happiness; it doesn't. That points away from material conditions and toward the subjective-affective register that the cognitive-distortion and locus-of-control theories operate in.

What the data adds (and doesn't)

Three things the GSS data above adds to the existing literature.

First, the partisan-happiness gap is older than the post-2012 widening. Conservatives have been roughly 5 points happier than Liberals since the 1970s. Anything that frames the gap as a phenomenon of the social-media era is incomplete. The widening is recent; the gap is not.

Second, the recent widening is real but small in marginals. On GSS happiness, the Liberal–Conservative gap moved from about 5 points pre-2016 to about 9 points in 2021, partially reverting to 5 points by 2024. The 2021 reading is contaminated by the mode shift and the pandemic; the 2024 reading is the cleanest post-shock estimate, and it shows a gap close to the historical baseline, not a structural break.

Third, the youth decline and the ideology widening are the same finding when you look at the cells. Figure 5 — the direct three-way replication — shows the 18–34 Liberal cell taking the largest post-2018 hit (a 9-point structural drop from the long-run mean), but it also shows the 65+ Conservative cell falling 9.4 points over the same window. The unique-to-young-Liberals reading the Gallup chart suggests is too strong on GSS happiness data. The right reading is: a broad-based well-being decline, concentrated harder on the young, and within the young harder still on Liberals. The interaction the Gallup chart visualizes is present in GSS — just smaller in magnitude and accompanied by a partial decline among older Conservatives that the Gallup chart did not show.

Implications

Three follow naturally.

The book's class-and-education frame still applies, but partially. Earlier chapters argued that variation in attitudes is dominated by education and income, with party providing a secondary effect. Happiness fits that frame poorly. The financial-satisfaction reversal — Liberals slightly above Conservatives in recent years — argues against a strict income mechanism. The youth × ideology pattern argues for something more subjective: a worldview-and-affect story, not a class story.

The mechanism question matters for what the gap implies. If the gap is reporting bias plus stigma asymmetry, the policy and clinical implications are limited. If the gap is real distress driven by ideological framing or news-cycle stress, it is a genuine well-being problem with a recent ideological etiology. The GSS data is consistent with either, but the financial-satisfaction reversal and the consistency across life_rec make a pure-reporting story harder to defend.

The 2021 reading is methodologically central, not a footnote. It is the single point where every subsequent claim about "the gap widened" has to defend itself against the mode-shift confound. Honest reporting of this chapter's finding requires distinguishing between the 2016→2021 movement (large, partly artifactual) and the 2016→2024 movement (smaller, robust). The narrowing in the 2022 and 2024 waves should not be ignored.

The youngest-liberal decline the original Gallup chart depicts is, on the direct three-way GSS evidence, real but more layered than the chart implied. The 18-34 Liberal cell dropped from a long-run mean of 28% very happy to 18.2% in 2024 — a 9-point shift below trend, and the largest post-2018 movement of any of the twelve age × ideology cells. But the 65+ Conservative cell — the comparison group the original chart used to contrast against — fell almost identically, from 41% to 31% over the same window. Both cells dropped; the gap between them is roughly where it has always been. What's new is not that young Liberals have fallen below the reference; it's that the entire happiness floor has dropped, with the young falling further than older cohorts and Liberals further than Conservatives within each age band. The cleaner finding from this chapter remains the age axis: for 40 years, age barely predicted American happiness. In the last six, it has begun to. That is the change worth following.