Falling Behind Is Not Falling
A claim like "confidence in X has fallen" almost always means one of two different things, and the two are not interchangeable. One is: X got worse, or people's regard for X dropped in absolute terms. The other is: X stayed roughly where it was, but other things moved up around it, so X's rank in the lineup slipped. The first is a statement about X. The second is a statement about everything else. Headlines report the second and let you hear the first, and that substitution is doing a lot of quiet work in how we read the last fifty years.
Take any survey that asks people to rate confidence in a dozen institutions, year after year. If you only look at one institution's line over time, a drop from third place to seventh reads as a fall. But rank is a positional measurement, not a magnitude measurement. It tells you where something sits in a crowd, not how big it is. A sprinter can run the exact same mile split for a decade and still finish last in a field that keeps getting faster. Nothing happened to the sprinter. The field changed. If you only ever saw the finish order, you'd write a story about decline, and you'd be wrong about what declined.
This matters because rank and magnitude respond to completely different interventions. If an institution's absolute standing has fallen, the fix is about the institution, something about it got worse, do something to it. If its rank has fallen while its absolute standing held steady, the fix, if there even needs to be one, is about everything else. You can't repair a sprinter's relative finish by training harder if the problem is that the whole field got faster. You'd be solving the wrong equation, confidently, with real effort, and the number would barely move, because the number was never really about you.
The slippery part is that both stories can be true about the same dataset and look identical from a distance. A chart of "percent expressing high confidence in institution X" sliding down over two decades is consistent with X getting worse, with the question wording changing, with the response scale changing, with other institutions climbing past X, or with some mix of all four. The line itself doesn't tell you which. You have to go stare at the other institutions' lines too, and check whether the question being asked thirty years ago was actually the same question, before you're entitled to a causal story about X specifically.
That checking is annoying and nobody does it by default, because a position chart is so much easier to produce than a magnitude chart. Rank is free: you just sort whatever numbers you already have. Magnitude requires you to ask whether this year's number means the same thing as thirty years ago's number, whether the scale shifted, whether the question wording changed enough to break comparability. Rank-order claims proliferate because they're cheap to generate and visually dramatic, lines crossing each other is a better chart than lines running flat. Magnitude claims are expensive and often boring, which is exactly why they're the ones worth doing.
None of this is an argument that decline never happens, or that institutions are all secretly fine and it's just vibes and league tables. Plenty of absolute decline is real and measurable once you've done the comparability work: wages against the same basket of goods, test scores against the same content standard, trust against the same survey instrument asked the same way. The argument is narrower: that a falling rank is not evidence of absolute decline on its own, and a headline that reports the rank while implying the magnitude is making a claim it hasn't earned.
The sprinter didn't get slower. The race got faster. Those are different sentences, and only one of them is about the sprinter.