& Blackwell Publishers Ltd 2000. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden MA 02148, USA. 297 INFORMATION AND COMMON KNOWLEDGE IN COLLECTIVE ACTION ARIEH GAVIOUS AND SHLOMO MIZRAHI* The paper suggests a theoretical game framework to explain collective action dynamics by learning processes. When a certain fact becomes common knowledge due to a certain event, people accumulate knowledge about the state of the world and act accordingly. We concentrate on a conceptual example of the bandwagon dynamic showing the insights which this approach adds to existing studies. We analyze two other conceptual examples showing that the accumulation of information does not always bring people to know the truth about the state of the world. On the other hand, it may make them aware of their ignorance. Several practical implications follow. 1. INTRODUCTION PROBLEMS ASSOCIATED with collective action are usually explained by the motivation of participants to obtain a ``free ride'' or by information problems in the sense that an individual does not necessarily know everything about others and especially whether or not he/she is a decisive player Downs, 1957; Olson, 1965; Hardin, 1982; Taylor, 1987; Gavious and Mizrahi, 1999). The information parameter and learning processes are therefore central in analyzing solutions for such problems ± particularly, the problem of mobilizing mass collective action. Analyses of learning or Bayesian updating processes explain how a player acquires new information about other players by means of observing their actions Granovetter, 1978; Taylor, 1987; Chong, 1991; Lohmann, 1993, 1994; Kuran, 1995). Traditional threshold models suggest that by observing the number of participants in events of collective action, players update their beliefs concerning the chances of success as well as the costs of action. As the number of participants increases, players infer that the costs of action are expected to be low and the chances of success to be high, a situation which should ®nally lead to mass participation through a more or less complex bandwagon dynamic Schelling, 1978; Granovetter, 1978; Oliver et al., 1985; DeNardo, 1985; Taylor, 1987; Chong, 1991; Karklins and Patersen, 1993; Kuran, 1995). An alternative explanation of the role of information in mass mobilization is provided by the informational cascade model Lohmann, 1993, 1994). Applying ECONOMICS AND POLITICS 0954-1985 Volume 12 November 2000 No. 3 * Corresponding author: Shlomo Mizrahi, School of Management, Ben-Gurion University, P.O. Box 653, Beer-Sheva 84105, Israel. Tel: +972-7-6472218; fax: 972-7-6472896. E-mail: shlomom@ bgumail.bgu.ac.il