You misunderstand. Of course it wasn't "a special review unit". It was the randomly selected review unit. As in any test methodology, if your random sample has

You misunderstand. Of course it wasn't "a special review unit". It was the randomly selected review unit. As in any test methodology, if your random sample has problems, it's a bad sign for failure rates for the batch.