Citizens are continuously inundated with political information. How do citizens process that information for use in decision-making? Political psychologists have generally thought of information processing as proceeding through a series of stages: (1) exposure and attention; (2) comprehension; (3) encoding, interpretation, and elaboration; (4) organization and storage in memory; and (5) retrieval. This processing of information relies heavily on two key structures: working memory and long-term memory. Working memory actively processes incoming information whereas long-term memory is the storage structure of the brain. The most widely accepted organizational scheme for long-term memory is the associative network model. In this model, information stored in long-term memory is organized as a series of connected nodes. Each node in the network represents a concept with links connecting the various concepts. The links between nodes represent beliefs about the connection between concepts. These links facilitate retrieval of information through a process known as spreading activation. Spreading activation moves information from long-term memory to working memory. When cued nodes are retrieved from memory, they activate linked nodes thereby weakly activating further nodes and so forth. Repeatedly activated nodes are the most likely to be retrieved from long-term memory for use in political decision-making. The concept of an associative network model of memory has informed a variety of research avenues, but several areas of inquiry remain underdeveloped. Specifically, many researchers rely on an associative network model of memory without questioning the assumptions and implications of the model. Doing so might further inform our understanding of information processing in the political arena. Further, voters are continuously flooded with political and non-political information; thus, exploring the role that the larger information environment can play in information processing is likely to be a fruitful path for future inquiry. Finally, little attention has been devoted to the various ways a digital information environment alters the way citizens process political information. In particular, the instantaneous and social nature of digital information may short-circuit information processing.
Online processing, and the models arising from it, starts with an optimistic view of the American voter, in which it is supposed that the seeming ignorance of voters does not prevent them from expressing rational attitudes about the very political objects they do not know much about. This means that the seeming ignorance of voters is not necessarily a threat to electoral democracy, but the cognitive structures needed for this sort of rationality also lead to necessary, and sometimes extreme, biases in political information processing. Since information stored in long-term memory is linked, both semantically and affectively (that is, based on the perceived positive or negative valence of the information), affect—understood here as a simple positive or negative valence—colors all steps of information processing. For instance, individuals are likely to avoid, or counter-argue, or simply reject information that is at odds with their existing views. As a result, individuals of different political persuasions may have difficulty coming to agreement on the correct interpretation of relevant facts, or even the facts themselves. Alternative memory-based models, which propose that evaluations are constructed on the spot when a question is asked, may help to explain response instability, but fail to serve as complete replacements for the online processing approach. The bias caused by affect-infused cognition seems to present challenges for electoral democracy just as much as the seeming ignorance it accounts for, but it is argued that such biases are mostly limited to individuals who already hold fairly strong existing attitudes, a group which is unlikely to include most voters. Moreover, some degree of intransigence is likely a good thing, as the alternative is views that shift rapidly with new information.