A publisher says manuscript quality improved because its acceptance rate fell. During the period, an advertising campaign caused a flood of low-quality submissions, while the number and quality of accepted manuscripts stayed about the same. What is wrong with the publisher's measure?