Simulating Heterogeneous Farmer Behaviors
Introduction
The field of social science attempts to understand and explain society as a function of the behavior people who constitute the society in question . To draw an analogy to mathematics , we may be tempted to model such a system with the central limit theorem , where the sum of many independent identically distributed random variables tends toward a normal distribution . Unfortunately , such an approach to social science is doomed from the start as people rarely make our choices independently of our social networks and we most certainly are not identical in all respects . These two assumptions that make central limit theorems possible are thus immediately and aggressively violated by human behavior and the societies we build , resulting in complex dynamical systems that seem unpredictable or even chaotic . The challenge to social scientists is to construct theories and models that are simple enough to be controlled in experimental settings and yet complex enough to capture the dependent , heterogeneous nature of human behavior .
From a practical perspective , the goal of many such models is to understand the actions that society might take in response to some stimulus as demonstrated in Epstein and Axtell ’ s work on growing artificial societies ( Epstein & Axtell , 1996 ). Governments , in particular , wish their citizens to engage in the political process , participate in national programs , and comply with laws and policies . In current practice , persuasive messaging campaigns designed to elicit such responses are often based on demographic or regional data that result in the use of static approaches to advertising and influencing behaviors ( Gupta & Chintagunta , 1994 ; Kalyanam & Putler , 1997 ). These static approaches necessarily assume that society is homogeneous within the given region or demographic and thus fail to capture the complexity and heterogeneity of real societies .
In this paper , we offer a preliminary step toward modeling how a heterogeneous society might respond to a persuasive messaging campaign . We apply an agent-based model to the question of how persuasive messaging impacts behavior and behavioral intent of a given population . The agents in our model make decisions that are based on a collection of theoretical frameworks of social psychology . The first of these is Ajzen ’ s Theory of Planned Behavior ( TPB ) that postulates intention is an indication of an individual ’ s readiness to perform the behavior of interest . Intention is believed to be determined by one ’ s attitude toward the behavior , one ’ s perceived normative pressure ( from their social connections ) to engage in the behavior , and one ’ s perception of one ’ s own ability to perform the behavior ( Ajzen , 1991 ). However , TPB makes no statements about how intention may change over time . It simply states that intention can be derived from a set of behavioral , normative , and control beliefs . The formation and change of these beliefs depend on our individual experiences , relationships , and exposure to various environmental pressures . To incorporate TPB into a set of virtual agents , we must embed the theory into a framework that allows agents to experience and learn from their environment . To accomplish this we incorporate a cognitive processing model put forth by Jager , Janssen , and Vlek ( 1999 ) and a conditioning model adapted from the approach used by Epstein in Agent Zero ( Epstein , 2014 ).
The remainder of this article is organized as follows . In the literature review , we review the literature relevant to TPB and the additional cognitive frameworks we employ . We formally define the TPB components of attitude , subjective norm ( SN ),
198