Simulating Heterogeneous Farmer Behaviors
Figure 3 . The dimensions of cognitive processing .
number of domains , such as health-related behaviors and consumer decision-making . However , there is significant variation in the predictive power of TPB from one particular behavioral decision to the next , and there are a number of criticisms of TPB . One of the criticisms involves the concept of “ inclined abstainers ,” who form what is measured to be an intention to act , but then subsequently do not act ( Sniehotta et al ., 2014 ). Central to this practical problem with TPB is the passage of time in an ever-changing environment . TPB was meant to equate intention to act with behavioral action in a relatively short interval of time . Although Sniehotta et al . advocate abandoning TPB altogether , we contend that when embedded in a larger dynamic framework , like the one set forth by Janssen et al ., the usefulness of TPB becomes more apparent . Specifically , it provides us with a conceptual model for simulating the potential impact of persuasive messaging on behavioral intent as it evolves over time . The presence of inclined abstainers , in turn , gives us a potential measure of the threshold behavioral intent must reach in order to translate into behavioral action . In the next section , we describe how we combined these two conceptual models
into a preliminary step toward simulating the collective response to persuasive messaging campaigns .
The Model
Model Components
Our objective is to create a dynamic model that can ultimately be used to demonstrate the relative effectiveness of a planned persuasive messaging campaign . As such , we require a messenger system that is representative of advertising media and a population of virtual agents who receive periodic messages from this system . Additionally , we would like to model different strategies of persuasive messaging . One such strategy often entails the use of celebrities or community leaders who speak out in support of the campaign ’ s interests . These influential figures often have much larger social networks ( e . g ., Twitter followers ) than normal . This also implies that our model needs a network structure for connecting individuals both to each other and to the messaging system . Finally , the agents are being encouraged by the persuasive messaging campaign to take some sort of action . When their behavior-
204