Since, by definition, human learning evolves through the individual’s interaction with the environment, it is a process that is always embedded in a situational background, and is therefore influenced by various situational factors. Some of these factors are accounted for in theoretical approaches to learning in economics, but are mostly trivialized, while others are idealized or completely omitted.
These factors are nevertheless known to be important for actual learning behavior in psychology, and, thus, are likely to influence economic (learning) behavior in predictable ways. For instance, existing institutions, or the actual design of a market may induce notable variations in these determinants, and thereby influence (learning) behavior and aggregate outcomes significantly. Hence, learning may be viewed as being contingent on the situational factors as discussed below. Before discussing single determinants, two remarks seem appropriate.
The first concerns the question in what sense learning can be assumed to be influenced by the determinants. In Section 4.2. the effects on learning will be operationalized in terms of outcomes, so that in the following discussion “easier” (or “more difficult”) learning is associated with better (or worse) outcomes. – The second remark concerns the isolation of the effects of single determinants.
As always in the analysis of complex situations, the proposed effects of a single factor may be overlapping with the effects of other factors. Therefore, the following discussion of single factors has to abstract from possible cross-effects of factors. Two groups of learning determinants (situational factors) can be distinguished: One group includes structural influences that exist prior to any (inter)action: (i) the complexity of the environment and the task; (ii) the number of actors involved with the situation, and (iii) the degree of uncertainty associated with the structure of the situation. These factors are labeled structural determinants.
A second group of factors is also induced by the structure but is related to (inter)action: (i) content, quality, and quantity of feedback, and (ii) strategic uncertainty in situations of strategic interdependence; labeled interaction determinants. Table 1 summarizes the determinants and assesses their significance with respect to ideal types of situations.
Complexity of and Information about the Structure
A situation (commonly modeled as an individual decision problem, a market, or a game) can be defined by its structure. The structure consists of the actors, the decisions they face, the information they have when making them (i.e., the actors’ knowledge), how their decisions determine outcomes, and their preferences over outcomes. It also incorporates any repetition, correlating devices, or opportunities for communication.
With respect to the complexity of the structure, two aspects can be distinguished: (i) the complexity of the environment and (ii) the complexity of the task that has to be accomplished within an environment. The environment may vary in complexity in terms of the number and complexity of its elements, its dimensions, and the relations between them. Within a given environment, the task may also vary in complexity depending on (i) the number of the dimensions of the task or problem, (ii) the number of combinations of dimensions, and (iii) the number of outcomes involved with particular actions or potential solutions.
Hence, depending on these factors, the task may involve more or less complex decision making, calculus, etc., and, learning is fostered with decreasing complexity of task and environment. For a given level of structural complexity, information about the structure available to actors may vary. In most cases, at least some of this information is given prior to any (inter)actions. Generally, the higher the level of information about the structure that is initially given, the easier is learning.
But since complexity is not a simple function of the quantity of information available to the actors, the effect of the content of structural information on learning must be analyzed for each situation separately. This analysis may be based on the assumption that the more the content of information reveals about the true nature of the structure, the easier is learning. – If actors know that they do not have all information about the structure, structural uncertainty results. The effects of structural (and strategic) uncertainty will be discussed in Section 4.1.3. Note that in many situations learning about the structure involves experience, i.e., direct or indirect observation of actions and/or their consequences.
Nevertheless, the underlying complexity of the structure, and information about the structure may influence learning significantly, and should therefore be analyzed separately. In traditional economics, the complexity of the environment is often reduced to a minimum, and the task is usually a relatively simple choice among two alternatives.
The effects of the complexity of the structure and information about the structure are commonly not analyzed as a separate variable, or as a determinant for learning processes.– The effects of variations in structural complexity and information on learning discussed in this section, are summarized in the form of hypotheses in Section 4.3.
54 For basic texts on learning and behavior see MAZUR (1994), and CATANIA (1992). Formal models of learning can be found in LUCE, BUSH & GALANTER (1963).
55 Something is mutual knowledge if all actors know it, and common knowledge if all actors know that all actors know it, and so on ad infinitum.
56 See CRAWFORD (1995a, 3). – When behavior is influenced by additional factors, such as how the situation is presented or the social setting in which it takes place, this is sometimes called the context.
57 For example, a simple environment is a room with only a sheet of paper and a pen, whereas a fully equipped modern office is a relatively complex environment.
58 For example, a situation with “few information” may appear complex because many questions are left open, while a situation with “much information” may also appear to be complex because the additional information may be redundant and may obscure what is important.
59 See also HEY (1992, 95), who argues in favor of experiments with varying degrees of complexity because theories and experiments that are too much “stripped-down” fail to handle behavior in complex environments appropriately.
Prof. Tilman Slembeck
A Behavioral Approach to learning in Economics
In economics, adjustment of behavior has traditionally been treated as a “black box.” Recent approaches that focus on learning behavior try to model, test, and simulate specific adjustment mechanisms in specific environments (mostly in games). Results often critically depend on distinctive assumptions, and are not easy to generalize. This paper proposes a different approach that aims to allow for more general conclusions in a methodologically more compatible way.
Prof. Tilman Slembeck