Bennett is confused. That we advocate voter registration reform does not mean we are single-factor analysts. Bennett does not understand the distinction between our analysis of the complex causes of low turnout, on the one hand, and our analysis of the limited options for reform, on the other.
As to the first, we wrote Why Americans Don't Vote because we thought the “party behavior,” “legal institutional,” and “social-psychological” schools were each inadequate, taken alone, as explanations of low turnout. The party system explanation rests too much of its historical account on the impact of the single election of 1896 in stifling party competition and increasing internal oligarchy and fragmentation, and it is evasive on the question of why the restoration of party competition in the 20th century did not raise turnout in 19th century levels. Legal-institutional explanations fail to provide an account of the political origins of legal arrangements, and of the uneven implementation of these rules by party organizations. And social-psychological explanations do not confront the anomaly that correlations between social traits and non-voting are unique to the contemporary United States. In fact, turnout was high in the 19th century when education levels were low, and then fell in the 20th century as education levels rose.