The manager of a district's fair price shop network argues: 'Our door-step delivery has worked. In the six months since it began, the number of card holders coming to the shops to draw their quota has fallen by two-thirds.' The manager's reasoning is most open to criticism because

A count of people at the counter goes down when they are served elsewhere and goes down just as far when they are served nowhere, so the fall the manager reports would appear whether his scheme were a triumph or a collapse. Evidence that cannot come out differently under the two rival accounts distinguishes neither, and the figure that would settle the matter — how many quotas were actually handed over at doors — is the one he does not give. The size of the register bears on the scale of the network. What the delivery costs bears on whether the scheme is worth running even if it works. And the fourth option attacks a claim the manager never makes: he speaks of a fall from the earlier figure, not of universal drawing before it.