Here's how those numbers are plausible: https://en.wikipedia.org/wiki/Three_cueing

Each of these quotes is from a different part of that Wikipedia page:

> It involves guessing the meaning of unfamiliar words based on context clues, as opposed to phonics, the more traditional method of sounding out words.

> Three cueing methods have been criticized for misunderstanding how reading is acquired and for potentially damaging children's reading abilities over the long term.

> In response to the example of children failing to distinguish between "pony" and "horse", Goodman argued that it was irrelevant whether children understood the specific word, as "pony" and "horse" are similar concepts, and a reader failing to distinguish between them would still understand the meaning of the story as a whole.

> Some researchers and educators have attributed part of three cueing's popularity in classrooms to its ease of teaching relative to phonics.

> As of 2020, an estimated 75% of American teachers used three cueing.

> Over 12 states have explicitly banned educators from teaching three cueing. A lawsuit was filed by families in Massachusetts whose children had been taught three cueing; one of the plaintiffs stated that her son had difficulty reading once classroom materials transitioned into using chapter books.

You're making the assumption that the change in absolute terms is entirely driven by deficient additions to the population. It's just as possible that some portion of the population lost their skill by allowing it to atrophy from underuse.

That assumes the shift was entirely due to high school students becoming adults. There are also people who haven't read much over those years and have had their skills declined to the point of failure. Or, also likely, sample size issues.