We've argued before that the left–right spectrum is a piece of furniture from 1789 doing the work of a theory. That's an opinion. This is the follow-up where we put a number on it.
Here's the experiment. Take everyone who has completed the DNA Scan — four axes, scored 0–100 — and ask a single question: if you had to compress all of that onto one line, what is the best possible line?
Not the traditional left–right line. Not a line we drew. The mathematically optimal one.
The best line the data can draw
Statistics has a standard tool for exactly this question: principal component analysis. Given a cloud of points in four dimensions, it finds the single direction that captures as much of the spread as possible. Whatever "the political spectrum" could ever hope to be, PC1 — the first principal component — is its ceiling. No one-dimensional summary can beat it.
We ran it on 1,168 quality-filtered Scan results (v2.x instrument, straight-lining and speeding responses excluded). The optimal line turns out to be a recognizable left–right axis: it loads toward market economics, national sovereignty, and cultural tradition on one end, and their opposites on the other. The 18th-century seating chart wasn't crazy — there really is a dominant axis, and it looks like the one everyone argues along.
But here's the number that matters:
The best possible left–right line captures 64% of what distinguishes our respondents from one another. The other 36% is structure the line cannot see — no matter how well you draw it.
What 1969 got right
It gets better. Ask the same math for the second-best direction — the most informative axis after left–right — and something delightful happens.
PC2, in our data, separates personal liberty and market economics from their opposites: the direction that distinguishes the libertarian-shaped person from the authority-and-solidarity-shaped person, both of whom the left–right line files in the middle.
If that sounds familiar, it should. It's the diagonal David Nolan added in 1969 when he proposed the two-axis chart that later mutated into every political compass meme you've ever seen. Our respondents' data, fifty-seven years later, independently rediscovers his axis as the single most informative thing to add to the line.
Together, the two axes capture about 85% of respondent variance. The old maps weren't wrong. They were early drafts — each one a real improvement, each one still flattening something.
The usual caveats, stated plainly
Our respondents are people who chose to take a political assessment online — a self-selected sample, not a probability sample of Americans. These numbers describe PoliticalDNA respondents, and we'd expect the exact percentages to shift as the sample grows. The structure of the finding — a dominant left–right axis, a Nolan-shaped second axis, and a large remainder neither can see — is the robust part. The derivation is documented and re-run as the sample grows.
See yourself on the old maps
We've turned this analysis into a feature. Your dashboard now includes You on the old maps: your own result plotted on the classic left–right line and the Nolan chart — including where the line would file you, where you actually sit, and which of your four axes it flattens most.
If you've taken the DNA Scan, your result page has the link. It's free — the panel just needs an account so it has a dashboard to live on.
The point of the exercise isn't that the old maps are useless. It's that you deserve to know what they leave out — and now, for the first time, you can see it.