Psychographic Message Framing Outperforms the Staff-Written Progressive Frame on Conservative Segments
Evidence from a 3,006-person randomized controlled trial, .
union support among religious conservatives, placebo to Personal Safety frame
over the staff-written frame on officer pay, conservative segments
AI frames against the staff-written frame with liberals and Democrats
person randomized controlled trial, 60.9% voter-file match
Summary
In August 2025, Hoplight tested two AI-generated psychographic message frames against a staff-written progressive frame and a placebo group who read an unrelated message, in a 3,006-person randomized controlled trial. On the officer-pay question, the AI-generated frames outperformed the staff-written frame by 10 to 23 points across the conservative values segments. With liberals and Democrats, the AI-generated frames performed about the same as the staff-written frame. On populations Democrats lost in 2024, the staff-written frame drove union support down rather than up. Values segments separated the frames more cleanly than party or race did.
Background
The 2024 election post-mortem split into two surface debates: a channel argument (Joe Rogan, TikTok, podcasts) and a strategic-posture argument (the majoritarian vs. base-maximalist false choice). Both missed the deeper failure, which sits at the message layer. Standard progressive messaging presumes the listener processes the world the way the sender does. When that presumption breaks, the listener hears moralizing as condescension, and the message backfires before it persuades.
Psychographic segmentation groups audiences by values, worldview, and cognitive style rather than demographics. Rather than writing one message for a demographic bucket, psychographic framing matches the message to the listener’s operating system: what they value, what they trust, what they read as a threat. This study tested whether AI-generated psychographic frames could outperform the standard progressive approach on a real policy issue.
Methodology
- Design
- Randomized controlled trial with four conditions: three message treatments and a placebo group who read an unrelated message about a salt company.
- Sample
- 3,006 US adults (weighted to 3,004), fielded August 19–20, 2025. 60.9% voter-file match rate. Weighted on age, gender, race, education, and party identification.
- Partner
- Grow Progress, a progressive research and technology firm with proprietary psychographic segmentation (eight values-based segments cutting across race, class, and party).
- Conditions
- (1) A standard progressive frame, written by senior union communications staff, using familiar progressive messaging: community caregiving, racial solidarity, collective bargaining. (2) Two AI-generated psychographic alternatives, designed to resonate with listeners whose moral architecture prioritizes safety, loyalty, authority, and earned respect. (3) A placebo group who read an unrelated message.
- Outcomes
- Policy agreement (“many security officers are poorly paid and resourced”) and union-support intent (support or oppose the right of security officers to join a labor union).
- Significance
- Significance as flagged by the vendor: 90% confidence versus the placebo. Respondents were general-population adults; the trial measured attitudes, not votes.
AI-generated frames outperformed the staff-written frame by 10 to 23 points on officer pay across conservative segments
On the officer-pay question, the AI-generated psychographic frames outperformed the staff-written progressive frame by 10 to 23 points across the three conservative values segments. With religious conservatives the staff-written frame reached 56% agreement and the AI frames reached 78% and 79%. With enterprising conservatives, the Personal Safety frame reached 82% against 56% for the placebo group, a 26-point gap and the largest in the study. Overall, agreement rose from 59% in the placebo group to 75% under the Personal Safety frame.
The AI frames held the progressive base
The AI-generated frames read like a departure from progressive values, because of their pride and valor tone. With liberals and Democrats they performed about the same as the message progressive human communicators were already using. On union support, the staff-written frame was the only one to lose ground overall, from 65% to 60%, while the AI frames held at 65% and 64%.
The staff-written frame produced backlash on 2024’s decisive populations
On the union-support question, the staff-written progressive frame drove backlash among the populations Democrats lost in 2024. The vendor flagged the first two moves as significant against the placebo:
| Population | Staff-written frame vs placebo | AI frames vs staff-written frame |
|---|---|---|
| Did not vote in 2024 | -10 pts (backlash) | +20 to +21 pts |
| Respondents under 35 | -9 pts (backlash) | +12 to +14 pts |
| Working class (<$50K) | -3 pts | +6 to +10 pts |
The frame the field defaults to is producing the opposite of the intended effect with the audiences that now decide elections.
Psychographic segmentation outperformed demographic segmentation as a predictor
Values segments separated the frames more cleanly than party or race did. Across the eight values segments the three frames moved in opposite directions, by as much as 23 points on officer pay. Cut the same data by party identification and the frames move together. The frame is the lever. Demographics are descriptive shorthand the field has been mistaking for an explanation.
What This Means for Practitioners
The progressive messaging playbook has a structural problem.
It is not a volume problem, a channel problem, or a courage problem. It is a frame problem. The default approach writes messages for people who already process the world through a progressive moral framework, then broadcasts those messages to everyone.
Psychographic framing is not micro-targeting.
It does not require individual-level data or surveillance infrastructure. It works by developing message variants matched to different cognitive styles, then deploying the right variant to the right psychographic segment. The segments cut across race, class, and party.
AI is the scalability mechanism.
Psychographic code-switching is what deep canvassing and relational organizing do at the individual level. Their per-conversation cost confines them to small-scale deployment. AI-generated psychographic framing is the path to that code-switching at the volume and fidelity a national cycle requires.
Study Limitations
This study measured attitudinal shift (policy agreement and union-support intent), not behavioral outcomes (votes, sign-ups, donations). The natural next step is a field deployment testing whether psychographic frames produce measurable behavioral conversion at scale, with voter-file match-back to validate impact on actual turnout and vote choice. The study tested one policy domain (security officer working conditions and unionization). Generalizability to other issue domains requires additional testing, which is underway.
Citation: Pendergast, W. (2026). Psychographic Message Framing Outperforms the Staff-Written Progressive Frame on Conservative Segments: Evidence from a 3,006-Person Randomized Controlled Trial. Hoplight Research Brief. https://hoplight.ai/research