Page images

also those imposed by their own differences and deficiencies in behavior. If it is postulated, as in the present paper, that certain drug-produced changes are acceptable to the social deviant, the previously given factor classification is suggestive of some of the alterations which might be desirable. Since some of the behavioral actions of narcotics and alcohol are known, it suggested that the immature, inadequate deviant who has not found independence in solving problems of adult life, may find in alcohol temporary independence from frustration, conflict, anxiety, and monotony, or in opiates nearly complete dissolution of such difficulties. While some of the same actions of these drugs would presumably occur in all deviants, the primary psychopath might be especially susceptible, depending on the degree of socialization, either to enhanced expression of hostility and aggression by alcohol, or to their elimination by opiates.

Framework for an Interactive
Theory of Drug Use

George J. Huba, Ph.D.
Joseph A. Wingard, Ph.D.
Peter M. Bentler, Ph.D.


Our theory of drug use takes the position that drug-taking behavior is caused by several large constellations of intraindividual and extraindividual forces. These domains of influences interact to modify each other while determining the presence or absence of a large variety of lifestyle behaviors, including drug and alcohol use. Many previous theories of drug taking have provided valuable contributions to the field and are correct as far as they go. Flaws in these theories stem less from incorrectness than from incompleteness; they focus on one set of forces to the exclusion of others. In order to provide a more comprehensive view of drug use than is typical, we will discuss the models or domains of influences which form the major subsystems in our larger theory, and then present more specific ideas on how different influences work to modify each other as well as to determine the performance of behaviors.

The detailed theory we will consider is presented graphically in figure 1. This diagram represents sets of influences as large boxes. We should point out quite forcefully that we believe each box represents many different variables, factors, or latent influences, some of which may be largely uncorrelated with one another; that is, what we have presented are relatively abstract domains of influences. In figure 1. we have also drawn a large number of single-headed arrows to signify presumed causal influences. Where no arrow appears, we believe that there is not a strong direct effect. While the mandated length of this chapter precludes a literature review, we should emphasize that most of the links have been substantiated empirically and are recognized as major conclusions by many researchers. The diagram is an abstract summary statement of our theory, which permits detailed empirical tests using a variety of research and analytic techniques, including the new methods for causal modeling with latent variables, a variety of continuous and discrete multivariate methods, and experimentation,

FIGURE 1.-Framework for a theory of drug use

[blocks in formation]

Within the framework, we will try to claim a rather modest role for ourselves, if such a stance is possible given the grandiose nature of the figure.


Having presented the model, we would like to digress somewhat and clarify several points about the framework we are suggesting for the development of a comprehensive theory of drug use. In general, we have attempted to integrate various major themes of research developed by previous workers. While the labels chosen for various domains of influence may not be entirely synonymous with the terminology used by specialists in different fields, we feel that the general set of domains can be differentiated into those variables addressed in studies that span disciplines from psychopharmacology to psychology, sociology, and economics. Second, the act of differentiation and greater specification in various systems is a desirable goal for both current research and future theory. We feel that systems of interest, such as personality, must be successfully charted by determining the major structural and dynamic components. Our current framework is a largely undifferentiated and unelaborated one which should develop naturally as more information about the various domains becomes available through basic and applied research in the major scientific fields of relevance to drug use. As a consequence, we expect that the future elaboration of our framework will possess some degree of ecological validity through empirical derivation rather than theoretical superimposition. Finally, we feel that the current framework allows the kind of differentiation which may permit confirmatory tests with such theory evaluation procedures as causal modeling with latent variables. That is, the current framework is explicitly designed to permit the comparison of various theories within a sophisticated, hypothesis-testing correlational methodology. While experimentation may well provide the best method for clarifying certain specific components of our model (e.g., the effect of certain products on various organismic variables), naturalistic research will be required to interrelate those many components that are not easily or ethically subject to manipulation (e.g., the effects of life stress on drug use) (Bentler 1978).

As this theory goes through successive generations of development there are several paths it must take. As a first task, we feel that within each domain there should be a clarification of major variables that are relevant to understanding drug use. There are certain domains which traditionally have been the province of a given academic specialty, and we feel it is important to combine information from various disciplines so that the sphere may be charted with a consentaneous set of structural referents. Second, we believe that there should be a focus on the development of various submodels within the more general framework. Indeed, there is probably a lifetime of research productivity involved in determining the major structural personality characteristics related to drug use. As information accumulates within each specialty area, we would wish to see further elaboration of the component systems. Third, there should be an attempt to integrate alternative empirical and theoretical systems into our overall concept. While we make no pretense of being able to explain all the phenomena of drug use, we propose the broad framework primarily because we hope that it has some potential for unifying more narrowly based concepts of drug use.


Turning away now from the abstract framework to the more detailed formulations we have chosen in our first attempt at theory, the reader will first note in figure 1 that we have included construct domains that do not directly influence either drug taking or its alternatives. We feel that it is necessary to include these more contextual domains in a theoretical and empirical specification so that we can assess indirect effects as well as derive unbiased estimates of the amount of their influence. Furthermore, we must remember that many different domains are changed directly and indirectly as consequences of drug taking and its alternatives. It is thus critical to consider the dynamic interactions of many different domains when considering drug-taking behavior.

A second major characteristic of our structural model is that the behaviors of drug and alcohol ingestion are embedded in a larger set of preferred behavioral styles which may complement or preclude one another. Indeed, it is necessary to speak of the psychosocial causes of drug taking and its alternatives because many of the alternatives share the same psychosocial causes and may bring the same consequences for the individual. The structural properties of individual behaviors must be considered within the interactive, ecological context of other characteristic behaviors performed and precluded so as to elaborate a theoretical network that has both convergent and discriminant validity.

At this time, we do not pretend to know whether it is more fruitful to approach a domain of behavioral styles from either dimensional or typological viewpoints. That is, we are not sure if there are delineable behavioral types or whether there are some major dimensions of behavioral perference and action. We do believe that it is important to know what other behaviors drug users also perform characteristically and to use co-occurrence with other behaviors as a way of differentiating among drug users. The present approach seems to open an avenue for conceptualizing other habitual behaviors, such as overeating, gambling, or obsessive shopping, in relation to the dynamic causes of drug taking. Our use of the phrase "behavioral styles" is meant to imply that the focus of our theory is on behavior that spans temporal and contextual effects. We are not particularly concerned with ad hoc and fleeting behaviors.

We are continuing to revise and expand the theoretical model. Consequently, the dynamic and structural properties implied by figure 1 should be perceived as a model in the process of evolution. Our goal is to develop and test many of the different submodels implied by the framework.

Proceeding to a detailed consideration of the figure, we have differentiated four major areas of interest at the highest level of abstraction. These are biological, intrapersonal, interpersonal, and sociocultural influences. At the very foundation of the biological area, we would place genetic influences. We also wish to differentiate a domain which we call organismic status and which includes such variables as health or efficient functioning as well as major anatomical and physiological systems. Those aspects of physiology directly confounded with the psychological status of the individual should be specified into a separate domain labeled psychophysiology.

« PreviousContinue »