In this thesis, various aspects of the modeling of social dynamics are discussed. Both theoretical and empirical illustrations and examples are given.
In Chapter 1, formal dynamic models are discussed in general terms. It is argued that a wider application of explicitly dynamic models in sociological theory is needed, and that formal - mathematical and computational - language is a superior tool in constructing theoretical models. In Chapter 2, (with Olof Bäckman) the weak relationship between mathematically formulated theories and statistical models in current sociology is addressed. Potential problems that may distort the theoretical interpretations of statistical tests as a consequence of refraining from an exact specification of the theoretical model are discussed. The Vacancy Competition Model is singled out as an exemplary illustration of an integrative approach for the improvement of quantitative sociology.
A rare process of list transfers at the Stockholm Stock Exchange in 1997 is analyzed in Chapter 3 (with Rickard Sandell). Drawing on a theory of social influence in board interlocks, the phenomenon is modeled as a continuous stochastic process by utilizing event-history techniques. The empirical results support for the idea that social embeddedness is a critical factor in some types of strategic decision making.
In Chapter 4 (with Fredrik Liljeros) the spread of competitive floorball in Sweden in the 1980s and 1990s is modeled as a mixed-influence diffusion process. The prime interest is to examine the relative importance of external and internal change agents. A contribution to this family of models is made, by developing an analytical technique for separating the two sources of influence. The differential equations have exact solutions, and the empirical estimation is undertaken by means of ordinary least square regression. The model provides a very good fit in most of the districts analyzed, and variations in contributions from internal and external influence are observed across districts.
In Chapter 5 (with Fredrik Liljeros), an extended modeling of both internal and external influence is proposed. A model is developed in which the internal influence incorporates the idea of thresholds, and the external influence is modeled as dependent upon parallel diffusion processes in the social system not as a constant parameter. The model is tested empirically on the diffusion of trade union members in the Stockholm area between 1890 and 1940, utilizing non-linear optimization.Chapter 6 is a theoretical analysis of the emergence of organizational rules. In this context, rules are meant to denote something that is the opposite of conscious rationality. It is argued that the presence of such rules can be explained only by means of a dynamic theoretical apparatus. Behavioral rules are then modeled as strategies in a two-by-two game, and evolutionary game theory is applied to model the evolution of organizational rules from such strategies. The essentials of evolutionary game theory is introduced, and some examples are given to demonstrate that a plain and straightforward selection mechanism, already in very simple games, gives rise to interesting and complex dynamics.
Stockholm: Department of Sociology, Stockholm University , 1999. , 151 p.