Expectations and Actions

Expectancy-Value Models in Psychology

Norman T. Feather

• Publisher: Routledge
• Year: 2021
• ISBN:
  9781000363715
• Edition: 1
• Topic: Psychology
• Subtopic: History & Theory in Psychology
• Index: Psychology

Originally published in 1982, this book examines the current status of expectancy-value models in psychology. The focus is upon cognitive models that relate action to the perceived attractiveness or aversiveness of expected consequences. A person's behavior is seen to bear some relation to the expectations the person holds and the subjective value of the consequences that might occur following the action. Despite widespread interest in the expectancy-value (valence) approach at the time, there was no book that looked at its current status and discussed its strengths and its weaknesses, using contributions from some of the theorists who were involved in its original and subsequent development and from others who were influenced by it or had cause to examine the approach closely. This book was planned to meet this need.

The chapters in this book relate to such areas as achievement motivation, attribution theory, information feedback, organizational psychology, the psychology of values and attitudes, and decision theory and in some cases they advance the expectancy-value approach further and, in other cases, point to some of its deficiencies.

1	Introduction and Overview

Norman T. Feather

The Flinders University of South Australia

This book examines the current status of expectancy-value models in psychology. The distinctive characteristic of this class of models is their attempt to relate action to the perceived attractiveness or aversiveness of expected consequences. What a person does is seen to bear some relation to the expectations that the person holds and the subjective value of the consequences that might occur following the action. We are dealing with a basic question in psychology, the relationship of actions to expectations, where these expectations encompass beliefs about the implications of behavior, and where an important set of these implications consists of consequences that have positive or negative perceived value.

The model that is the center of attention goes by a variety of names. In motivational psychology it is usually referred to as the expectancy-value model, although the alternative and perhaps more accurate term, expectancy-valence model, has also been used. Decision theorists refer to it as the SEU model because the decision rule involves maximizing subjectively expected utilities involved in the choice situation. The terms expectation and expectancy are used interchangeably and they are indexed in terms of the perceived likelihood that an action will be followed by a particular consequence—that is by a subjective probability that the consequence will occur given the response. Similarly, there is a high degree of overlap in the terms used to refer to the subjective value of the expected consequences. Among the concepts that have been employed are incentive values, utilities, valences, and reinforcement values—all of which can be positive or negative.

Some theorists distinguish outcomes from consequences, a given outcome (e.g., success) being associated with a defined set of consequences that may have different levels of perceived likelihood for a person (e.g., more pay, better working conditions). Other theorists distinguish between outcomes at different levels (e.g., level 1 and level 2 outcomes). What is common to all of these analyses is the recognition that actions and their potential consequences are embedded in a complex means-end structure that involves beliefs about the implications of events extending beyond immediate consequences to possible future consequences. People are assumed to possess cognitive structures that concern the implications of their actions, both now and in the future. These implication structures may not always be well-defined, they may be in error, and one would expect them to vary in their details from person to person. But they are assumed to exist and, along with subjective values, valences, or utilities, to be important determinants of goal directed behavior.

In order to provide a concrete image of this cognitive form of analysis I will use a map as an analogy. A map sets out a pattern of interrelationships that schematize the implications of following particular routes. In planning a vacation, passing through some cities and along some roads may be more attractive than other alternatives. Taking some routes may involve additional expense beyond a person’s available resources and income. Some roads may be closed or obstructed or they may be difficult to travel given the condition of a person’s car. Some locations have to be passed through before others can be reached, given the structure of the geographical environment. All these variables would be expected to influence where one decides to spend the vacation and which route one chooses to take. Theoretical approaches that use the concepts of expectancy and subjective value assume that, just as a geographical map can be inspected for a given purpose, so too can people consult their own conceptual or cognitive maps to examine the likely implications of alternative actions and to evaluate in positive and negative terms their anticipated consequences. The decision taken is then assumed to be some function of the person’s expectations and subjective values as they relate to possible courses of action and consequences. The map analogy can be broadened to take account of time perspective because cognitive maps, though available for present use, incorporate the effects of past events and ideas about the future. The cognitive map may also be a distorted impression of reality, affected by one’s wishes and fears. And, of course, it could be in error because of insufficient information.

My selection of the map analogy was deliberate and not accidental. Both Tolman and Lewin conceived of the environment as structured and a person’s cognitive representation of it as having map-like qualities (Tolman, 1948; Lewin, 1936). Both were concerned with the structured representation of events— Tolman with the concept of expectation and the development of cognitive maps; Lewin with the topological characteristics of both the person and the psychological environment represented in terms of a finite set of interconnected regions. These seminal contributions of both Tolman and Lewin have had a lasting influence on the development of expectancy-value models in psychology. Both theorists emphasize cognitive structures that incorporate a coherent set of events interconnected across space and time.

The concern with cognitive concepts that capture the structured nature of knowledge and experience has a long history in psychology, going back at least to the Gestalt psychologists (Koffka, 1935; Kohler, 1927). In an earlier article I noted some of the concepts that have been used, especially as they relate to the cognitive and behavioral effects of structural discrepancies (Feather, 1971, pp. 14–15). As well as the concept of expectation, other structural concepts that come to mind are hypothesis, schema, prototype, category, belief, naive theory, cell assembly, abstract structure, personal construct, image, plan, script, and balanced structure. All of these concepts depart from a simple stimulus-response associationist analysis of behavior and recognize the individual’s capacity to process information so that it becomes organized and set within a context of meaning. These organized residues of experience then become important filters for future information processing, serving as benchmarks, criteria, or reference frames against which new information can be tested. They also guide the form that behavior takes. Such a cognitive analysis moves one far beyond a machine like view of the person reacting on the basis of stimulus-response connections that are either innately programmed or learned. Instead, the model of the person is that of an active processor of information, organizing and constructing experience into meaningful internal representations, and behaving not as an automaton but as a thoughtful, purposeful being.

There is little doubt that, from the present vantage point in history, the old debate in the 1940s and 1950s between the stimulus-response associationists and the cognitive psychologists has ended in a resounding victory for the advocates of cognitive theories. Yet there is a problem that continues to plague some cognitive approaches—an old problem that has drawn comment (e.g., Brunswik, 1943: Guthrie, 1935) and which needs to be raised again. Let us consider an early statement of the problem. In a well known criticism of Tolman’s theory, Guthrie (1935) wrote:

Signs, in Tolman’s theory occasion in the rat realization, or cognition, or judgment, or hypotheses, or abstraction, but they do not occasion action. In his concern with what goes on in the rat’s mind, Tolman has neglected to predict what the rat will do. So far as the theory is concerned the rat is left buried in thought; if it gets to the food-box at the end that is his concern not the concern of the theory [p.172].

The point is that some cognitive theories have left a vacuum between cognition and action. Although there are theorists who have tried to fill this vacuum (see Miller, Galanter & Pribram, 1960, for a notable example), others have continued to ignore it, so that the organism still remains buried in thought, trapped within its own knowledge structures.

To be fair to Tolman, he did respond to Guthrie’s criticism by developing principles of performance that acknowledged the importance of needs and valences as well as expectations in the prediction of a rat’s behavior in a Skinner box (Tolman, 1955). And Lewin (1951) in some of his later work attempted to go beyond the encapsulated life-space to draw links between the “real” world, the cognitive representation of this reality, and behavior in real-life contexts (Cartwright, 1959, 1978).

Tolman’s (1955) solution to the problem was in effect to develop an expectancy-value model, incorporating valences or subjective values as well as expectations. Miller, Galanter, and Pribram (1960, Chapter 4) in their seminal work on images, plans, and behavior also pointed to the importance of taking evaluation into account, although their approach was cast in terms of TOTE units (Test-Operate-Test-Exit units) using cybernetic and computer simulation principles in the context of an information-processing analysis. Miller et al. (1960) endorsed the view that “… knowing is for the sake of doing and that doing is rooted in valuing … [p. 71]” and argued that any psychology that provides less than a recognition of the connection between knowledge, evaluation, and action, that “… allows a reflex being to behave at random, or leaves it lost in thought or overwhelmed by blind passion—can never be completely satisfactory [p. 71].”

Expectancy-value models may be seen as one possible way of filling the theoretical vacuum between knowledge and action by way of bringing in evaluation as an important component. They have in common a reference to higherorder cognitive structures that represent the interconnectedness of events. In expectancy-value models the events concern actions and their expected consequences and these consequences may have different degrees of positive and negative subjective value or valence.

Some 20 years ago I noted that there were a number of expectancy-value models in existence and that they had emerged from diverse areas of psychology (Feather, 1959a)—level of aspiration (Lewin, Dembo, Festinger, & Sears, 1944), the analysis of a rat’s bar-pressing performance (Tolman, 1955), the theory of decision making (Edwards, 1954), social learning and clinical psychology (Rotter, 1954), risk-taking behavior and achievement motivation (Atkinson, 1957), and from my own research into committed and wishful choices between alternative goal objects (Feather, 1959b). All of these models used similar concepts and it seemed to me at that time that they presented a major alternative to the drive theories that were then very much in vogue (Hull, 1943; Spence, 1956). They took a different point of view from drive theory. Instead of attempting to explain behavior in terms of stimulus-response connections along with intervening constructs such as drives and incentives, the expectancy-value approaches emphasized the cognitive structuring of reality by an active organism and the purposive quality of its behavior within a means-end framework—a structured set of relationships in which various means or alternative courses of action could be seen as relevant to the attainment of positive goals or the avoidance of negative consequences.

Each of these models assumed that the behavior that occurred in a given situation could be related to a resultant that maximized the combination of expectations and subjective values. The organism was assumed to pursue that course of action associated with a maximum subjectively expected utility, resultant weighted valence, resultant motivation, behavior potential, choice potential, performance vector, or whatever the combination was called. Obviously the general approach assumed that the organism had the capacity to process information so that cognitive expectations about the implications of actions could be formed and be elicited on the basis of past experience and present perception, that it could also assign positive and negative values to the various possible outcomes or consequences of these actions, and that in some sense it could also combine particular expectations and subjective values so as to arrive at some meaningful resultant. In these models, action was therefore assumed to be related to the interplay of knowledge and evaluation within the perceived situation.

In this book we look at expectancy-value models as they have developed in the 20 or so years since my earlier review (Feather, 1959a). How have they weathered these intervening years? In what areas have they been used? What are their strengths and weaknesses? How generally can they be applied to the analysis of action? These are some of the questions that we hope to answer. The chapters in this book cover theoretical and empirical contributions from a number of different areas, and they are international in their character spanning North American, European, and Australian research. In organizing the contents of the book I have not taken the role of the protective advocate. There are positive statements about expectancy-value models, to be sure, but this volume also includes a considerable amount of critical comment and counterpoint—deliberately so, for that is the way we ultimately move forward.

Let us now set the scene by briefly describing the organization of this book and summarizing some of the distinctive contributions of each chapter.

OVERVIEW OF CHAPTERS

The chapters in this book discuss expectancy-value models in relation to seven different contexts: achievement motivation, attribution theory, information feedback, social learning theory, values and attitudes, organizational psychology, and decision making. Psychologists working in each of these contexts have used expectancy-value models to account for important characteristics of motivated behavior such as changes in the direction of behavior following choice or decision, the latency of response and the persistence of behavior in a particular direction, effort expended, and the level or amplitude of performance. Taken together the chapters relating to each of these contexts are intended to provide a wide ranging and scholarly account of the current status of the expectancy-value approach.

The Context of Achievement Motivation

The model of risk-taking behavior proposed by Atkinson (1957) has been an important source of hypotheses about how people with different motives might be expected to behave in achievement situations where both success and failure are possible outcomes. Atkinson’s model was an early example of an expectancy-value model and it had considerable influence on subsequent theorizing and research in the achievement motivation area.

The first section of this book contains four chapters that describe developments since the initial statement of the Atkinson model. In Chapter 2, Atkinson discusses the background to the development of his model, bringing to the fore some important historical issues. He outlines his earlier conception of how expected consequences influence actions (the risk-taking model or theory of achievement motivation) and examines in detail how the theoretical statement forced a new and different conception of the motivational significance of an inhibitory tendency—implying that it should be regarded as a tendency not to perform an act where the expected consequence is negative. Atkinson then argues that the earlier theory of achievement motivation was limited because it took an episodic view of behavior. He describes new developments that follow from a theory about the dynamics of action that assumes that the person is always engaged in some activity, and that theories of motivation should relate to the ongoing stream of behavior and changes in the stream rather than to situationally bound episodes (Atkinson & Birch, 1970, 1978). He redefines concepts from the early risk-taking model in terms of the new concepts of instigating and inhibitory forces. This redefinition has important implications, one of which is that in accounting for choice or preference, one should look at ratios that involve fo...