of membership participation: for example, "marginal" members might be motivated from a different set of incentives than "militants." In conceptualizing this variable, we are interested only in the type of incentive that most commonly underlies the behavior of party "militants," who are defined in the preceding variable, 11.02, as those who attend almost every party meeting and constitute a ready source of manpower for performing party activities. For this variable, the concept of party militant applies to parties which have no formal membership as well as those which do.

The concept of material incentives employed in this variable includes instances of government jobs awarded to party members or their relatives or friends in return for services to the party ("patronage" in the classic sense), preferential enactment of laws for monetary gain, and even money payments from the party or government treasury.

Operational Definition. A party is scored according to the extent that party militants seem motivated by material incentives, using the highest applicable code.

0	Few militants, if any, seem motivated by material incentives.
1	About one-third of the militants seem motivated by material incentives.
2	About half of the militants seem motivated by material incentives.
3	About two-thirds of the mililtants seem motivated by material incentives.
4	All or almost all of the militants seem motivated by material incentives.

Coding Results. Needless to say, coding parties across the world on the extent of their reliance on "material incentives" in motivating party militants was hazardous at best. The means for AC 1103 in Tables 12.3a and 12.3b are the lowest reported thus far for any variable, admitting that inferential leaps were taken from scanty items of information. Obviously, the type of data required to pin down the motivational bases of party militants rarely existed in the literature. This information needs to be collected through interviews or questionnaires, and systematic research with such instruments is seldom encountered in the literature of party politics in the 1950s, despite its increasing use today. Given the option of abandoning the variable with dignity or trying to save it at the risk of embarrassment, we chose the riskier course. While we seldom had adequate empirical data to back our estimates of the extent of material incentives as a wellspring of militant behavior, we persisted in forcing judgments on more than 70 percent of the parties. In some cases, we were confident of the judgments even in the absence of good data. This occurred, for example, when we coded the militants of outlawed parties, hounded by government police and troops, as not being motivated by material incentives. But apart from such extremes, our judgments usually constitute only best guesses of the situation, as our AC codes readily admit. The significant correlation of .25 between BV 1103 and AC 1103, moreover, signals our higher confidence in scoring parties for the presence of material incentives rather than their absence.

Temporarily setting aside the issue of data quality for BV 1103, we can consider the results of our coding. According to our scoring, more than half the parties offer their militants little in the way of material incentives to encourage their activity on behalf of the party. At the other extreme, less than one party in ten wins performance solely through promises of personal gain, although another 10 percent or more elicit action from a majority of militants by dangling concrete benefits before them.

go to page 130