Does the flat pattern hold up within individual categories, not just across size groups?

An aggregate percentage can be misleading if it is an average of wildly different individual cases. It is possible, in principle, for the overall figure that 684,307 of 752,124 answer sets (91%) named six or more distinct businesses to be true while individual categories vary enormously, some clustering near zero and others near total, with the overall number just landing in the statistical middle without describing any real category well. This finding checks that possibility directly by looking at the spread across individual groups rather than the pooled total.

The study measured 188 individual groups (categories) and recorded, for each, what share of its answer sets named six or more distinct businesses. Across those 188 groups, the median share was 91.8%, meaning half of all groups had a six-or-more rate at or above that figure and half fell below it. The lowest group recorded a rate of 65.3%, and the highest recorded 98.3%.

That is a real range, not a point. A group at 65.3% behaves quite differently from one at 98.3%, and a company operating in a low-rate category should not assume the overall 91% figure describes its own situation. But the range, while real, is still concentrated well above the halfway point. The minimum of 65.3% is still a majority, not a near-zero outlier, and the median of 91.8% sits close to the overall pooled rate of 91%. Contrast this with the mirror-image spread for the five-or-fewer class, where the median across the same 188 groups was only 8.3%, with a maximum of 34.7% and a minimum of 1.7%. In other words, even the group least likely to get a rich, six-or-more answer still produced one at a rate of 65.3%, well above the rate for terse, five-or-fewer answers in that same group.

So the flat pattern across size segments is not an artifact of averaging opposite extremes into a misleading middle. Most individual categories, whatever their size, sit in a similar band, tilted toward naming six or more businesses rather than five or fewer. The variation that does exist, from 65.3% to 98.3%, is worth investigating on its own terms, since something is driving individual categories toward one end or the other. But this study, built around category size as the explanatory variable, finds that size is not that something, since the spread does not track cleanly with the size segments already examined. A category near the low end of the six-or-more range is not reliably a small category; it could as easily be a large one with some other property, unmeasured here, that produces terser answers.