How could it be? Once you've conceded that the difference is small, there is no reason why this factor should come into play at all. Take Georgetown vs. American, for example. Even granting (probably incorrectly) that you'll finish several percentage points higher at American than Georgetown, at every point on the continuum you're better off at Georgetown.
You're mistaken in looking at those two graphs and then stopping. If all that was in consideration were job prospects, then the reputational advantage from the better school far exceeds the indirect employment benefit you'd gain from the big fish factor.
But there are other significant factors, such as the net cost of attendance, which is likelier to be much lower at the lower-ranked schools. These other factors' values may vary considerably depending on the preferences of the applicant. If a applicant was particularly risk-averse, they may be ambivalent between full-ride with stipend at the lower-ranked school or sticker price at a t14.
Now, imagine two worlds. One, the one we are in, in which students from higher ranked schools, on average, do slightly better on tests. Let's say, for the purposes of simplicity, that this will result in a given student to be 10% lower in rank at the higher school
In the other world (one that is implausible but important for illustrative purposes), law schools randomly admit students, and so, there's no reason to presume that any school is better than another in test-taking ability. Better ranked schools still inure the same employment advantages.
So, say there was one applicant with identical preferences in each of these worlds (A in our world, B in example world) who was offered $$$$ at lower-ranked school x and sticker at higher ranked school y.
There seem to be circumstances in which the preferences of A and B (better job prospects, but significant debtload, etc.) could make it such that it is a close-call, such that A would choose the lower-ranked school, but B would choose the higher ranked school.
Perhaps to numerically model the values associated with each choice
School x: 900 (value of debt avoidance) + 420 (value of job prospects)
School y: 1300 (value of job prospects)
In this case, a small value of big fish factor, such as 25 could make the difference between a given student attending higher-ranked school y vs. lower-ranked school x.
I am merely making the modest suggestion that an applicant consider this, not as a be-all-and-end-all, but as another factor in their school selection process.