Political elites spend considerable resources on recruitment activities. Existing research generally concludes that these activities are effective, but also suggests that political elites recruit strategically But strategic recruitment may undermine the impact of recruitment on participation if political elites use past participation as an indicator of the ability and willingness of individuals to respond to recruitment efforts since recruitment may then be directed toward individuals who would have participated without recruitment. Furthermore, the existing research fails to capture many recruitment efforts. It also has employed recruitment variables measuring requests for a political act different from the act used as the dependent variable. As a result, the actual effect of recruitment on participation is still in doubt. Using more appropriate data we demonstrate that recruitment of all types is shaped by past participation. However, while controlling for past participation does reduce the impact of recruitment on political participation-being asked to take part or give money-does expand political participation.
Recruitment, the explicit request by political elites for individuals to perform some political act, has figured prominently in two recent seminal works on political participation: Steven J. Rosenstone and John Mark Hansen's Mobilization, Participation, and Democracy in America (1993) and Sidney Verba, Kay Lehman Schlozman, and Henry E. Brady's Voice and Equality: Civic Voluntarism in American Politics (1995). Rosenstone and Hansen (1993) view recruitment as a type of political mobilization and political mobilization, they assert, is necessary for political participation. Verba, Schlozman, and Brady (1995) postulate that recruitment is one of three major causes of political participation. Both of these analyses demonstrate that recruitment is positively related to a variety of political acts and their findings are consistent with a long series of studies dating back 70 years documenting the impact of recruitment on turnout (Berelson, Lazarsfeld, and McPhee 1954; Blydenburgh 1971; Calderia, Clausen, and Patterson 1990; Eldersveld 1956; Eldersveld and Dodge 1954; Gosnell, 1927; Huckfeldt and Sprague 1992; Krammer 1970; Lupfer and Price 1972; and Wielhouwer and Lockerbie 1996). Huckfeldt and Sprague (1992) and Wielhouwer and Lockerbie (1996) also found that party contact is positively related to political acts other than voting.
But Rosenstone and Hansen's (1993) and Verba, Schlozman, and Brady's (1995) estimate of the impact of recruitment on political participation may be flawed for several reasons. First, Rosenstone and Hansen's analysis missed many recruitment requests. Given the lengthy time span of their analysis they were forced to rely on the only consistent measure of recruitment available in the NES surveys, the party contact question. The wording of this question is as follows: "The political parties try to talk to as many people as they can to get them to vote for their candidates. Did anyone from one of the political parties call you up or come around and talk to you about the campaign?" While respondents may confuse recruitment requests from candidates with requests from parties it is clear that requests from many political actors, such as interest groups, are missed by this question.'
The second problem with Rosenstone and Hansen's (1993) analysis of recruitment is a mismatch between the recruitment request that is captured by the party contact question, and the political acts they used as the dependent variables. While the exact request that the party worker is making is not explicitly specified in the party contact question, the most likely request is for the respondent to register if he or she is not already registered and then to cast his or her ballot for the party's candidates. This kind of request should impact electoral participation. While some of these conversations may contain a request for campaign participation or money and while a request for electoral participation may have some spillover effect on these other political acts, the effects on these other kinds of participation should be muted, at least compared to that generated by more explicit requests for these other political acts. …