Scales to Measure Social Power in a Consumer Context

Jonathan E. Brill, Temple University
ABSTRACT - Although the concept of social power is frequently used in social science research, its conceptualization is not consistent. Based on a literature review analyzing benefits and drawbacks of two dominant conceptualizations of power, the argument is made that power may be appropriately measured using attributions made by an involved social actor. It is argued that these attributions have an underlying two dimensional factor structure framed by one or more applicable social contexts. These ideas are extended to a retail shopping context to develop scales for measuring consumer power. The resulting scales are shown to (1) be valid, (2) have acceptable levels of reliability, and (3) be supportive of the two dimensional conception of power advanced. Potential applications for using the scales in consumer research and limitations of this study are discussed.
[ to cite ]:
Jonathan E. Brill (1992) ,"Scales to Measure Social Power in a Consumer Context", in NA - Advances in Consumer Research Volume 19, eds. John F. Sherry, Jr. and Brian Sternthal, Provo, UT : Association for Consumer Research, Pages: 835-842.

Advances in Consumer Research Volume 19, 1992      Pages 835-842

SCALES TO MEASURE SOCIAL POWER IN A CONSUMER CONTEXT

Jonathan E. Brill, Temple University

ABSTRACT -

Although the concept of social power is frequently used in social science research, its conceptualization is not consistent. Based on a literature review analyzing benefits and drawbacks of two dominant conceptualizations of power, the argument is made that power may be appropriately measured using attributions made by an involved social actor. It is argued that these attributions have an underlying two dimensional factor structure framed by one or more applicable social contexts. These ideas are extended to a retail shopping context to develop scales for measuring consumer power. The resulting scales are shown to (1) be valid, (2) have acceptable levels of reliability, and (3) be supportive of the two dimensional conception of power advanced. Potential applications for using the scales in consumer research and limitations of this study are discussed.

INTRODUCTION

It is widely recognized that exchange theory has been a dominant paradigm in the development of marketing thought (Bagozzi 1975, 1979). As power is a core concept inseparable from the ontology of social exchange theory (Blau 1964), it is clearly an important variable to be considered in the study of behavioral relationships between social actors operating within a marketing system. Dixon and Layton (1972), in discussing the role of power in marketing channel systems, have argued that power relationships are necessarily derived from the interdependencies that must exist between marketing actors if the channel system is to survive, and that these interdependencies are responsible for inducing marketing exchange partners, presumably with inherently conflicting objectives, to engage in cooperative behaviors.

Consistent with this reasoning, many researchers have emphasized the role of social power in marketing channel systems. In fact, much of the work in this area has been reviewed by Gaski (1984) who, through a synthesis of research results, has proposed a theory of power and conflict among distribution channel buyers and sellers. Yet, research emphasizing power relationships between retailers and consumers, the buyers and sellers at the end of many marketing channel systems, is relatively uncommon. And, even when these power relationships are considered, the focus is typically on the retailer with little, if any, regard paid to power associated with consumers (e.g., see Porter 1974).

Clearly, then, the potential for explaining interpersonal behavior styles of retailer representatives and consumers through exchange theory remains largely unexplored. It follows that use of a well defined conception of power and appropriate instruments to measure the power construct will be essential to this work. Unfortunately, however, there seems to be little consensus regarding the conceptualization of power and its measurement, a situation that has been lamented by many (e.g., see Cartwright 1959; Pahng 1989), and published scales to measure the social power commanded by consumers appear to be absent from the literature.

TOWARD A THEORETICAL CONCEPTION OF POWER

Definition and Elaboration

Although power has been defined inconsistently, two general approaches to its conceptualization seem to be dominant. These may be referred to as the "power as a resource" and "power as an attribution" perspectives.

Perhaps the more common of these is the power as a resource perspective. Proponents of this view include Dahl (1957), French and Raven (1959), Rollins and Bahr (1976), and Tawney (1931), among others. In this view, power is conceived as an inherent characteristic or quality possessed by a social actor, a definite potential or capacity which may be used to influence or even control the behavior of others in some manner within some context.

Pahng (1989, p. 74) has noted that this conceptualization of power has been attacked in the literature, often quite strongly, being "... criticized as ambiguous, unscientific, and unmeasurable." Cross (1969), one of the more outspoken critics, has provided a particularly clear presentation of the crux of the case against the resource conception. He noted that the measurement of power as a resource relies upon the description of outcome events occurring between social actors. That is, an amount of power is associated with a social actor because the behavior of another is observed to be consistent with the first actor's wishes rather than his or her own. When such measures are then used as predictors of behavior, as is typical in empirical studies involving the power construct, an intolerable dilemna is created; if power is to be attributed in terms of behavioral outcomes and then used to predict behavioral outcomes, what results is tautology.

An additional problem of the conceptualization of power as a characteristic of an actor is that it may stand independent of the crucial idea that power arises as the consequence of dependency. This point has been emphasized by Thibaut and Kelley (1959, p. 124) who have noted that "an individual's power over another derives from the latter's being dependent upon him." Furthermore, Emerson (1962) has observed that power is specific and peculiar to social actor dyads, even when social groups are considered, since it is not uncommon for one individual to act dominantly toward a second party but subserviantly toward a third who may act subserviantly to the second. When power is viewed as a resource with definite magnitude or potential strength, such intransitive relationships must be prohibited for they are logically impossible.

Because of these shortcomings, the other conception, power as a perceived or attributed characteristic of a social actor, seems more sound conceptually. Though similar to the power as a resource conception in that it too posits that power is related to interpersonal or social influence, the power as an attribution perspective entails a key difference: it specifically recognizes that power is an aspect of the relationship between social actors, not a characteristic of a social actor. In other words, power perceptions are derived from the deference one shows another because of his or her dependency or reliance upon the other to behave in a desired manner. Hence, people do not possess power; they are empowered by perceptions of the interpersonal dynamics experienced in relationships.

This conception of power is bolstered by the work of selected social psychologists who have been concerned with power measurement. For example, in studying the processes of social influences in groups, Lippitt, Polansky and Rosen (1952) have not only implied that attributed power is real power, but they have further suggested that attributed power is a valid and reliable measure of social power. Additional empirical support for this proposition has been provided by Tagiuri, Bruner and Blake (1958), while conceptual support is apparent in the literature on negotiation processes (Bacharach and Lawler 1981; Pruitt 1981). Furthermore, it is particularly instructive to note that, when measuring power in studies of channel relationships, it has been common practice (e.g., see El-Ansary and Stern 1972; Etgar 1977, 1978; Hunt and Nevin 1974; Lusch and Brown 1982; Wilkinson 1974) to use measures of influence which the survey respondent has attributed to other parties involved in the channel network.

A critical difference between these two power perspectives is the comparative ease with which the power as an attribution is able to account for the role of social contextuality in power measurement. The power as an attribution view clearly presumes that power has a dynamic property. To be sure, power is often volatile and, sometimes, even fleeting since it is subject to readjustments as circumstances change or the context of a relationship evolves. In contrast, the conception of power as a characteristic resource seems inconsistent with power variances of social actors across relationships and time. Indeed, to retain this theoretical view, one must posit that a social actor has a specific power resource for every conceivable social relationship under every conceivable social circumstance. Hence, the power as an attribution conception offers greater potential than the power as a resource view for the development of parsimonious models incorporating the power construct.

Still, the belief that power in relationships is subject to change does not forbid one to posit that power attributions have a tendency toward stability over time, particularly when a relationship is ongoing. After all, if power attributions in a social relationship are the consequences of interpersonal behaviors, then the stability of power attributions should be similar to the stability of interpersonal behavioral patterns exhibited by the parties involved in the relationship. To suggest stability in power attributions, then, is to rely only upon the idea that people have propensities to exhibit behavioral traits, or personalities, in their interpersonal behaviors. This point, of course, is critical to empirical research concerned with power, for it is only because of its tendency toward temporal stability that power measures useful for studying social relationships may be developed at all.

Consequently, a workable definition of power now may be advanced: power is the perceived ability or potential of a social actor to influence or control the behavior of another within a given relationship or context. Furthermore, it is an outcome of social dependency. It is dependency, not power, that may be viewed as an antecedent condition for predicting behavior. Power, then, can be a useful scientific construct only when it is employed as an indicator for social dependency.

Extending this to a consumer context, a consumer's power over a retailer or its sales agent is derived from the retailer's desire for the use or ownership privileges over the money resources commanded by the consumer. Likewise, a retailer's power over a consumer derives from the consumer's dependency on the retailer as a source of supply of those products and/or services which he or she may desire. As such, the interdependency existing between retail exchange partners leads to their attributed powers with respect to each other.

Furthermore, although much of the ascendancy granted by each party is purely economic in nature, some aspects of this ascendancy are purely social, governed by the institutionalized structure specified by the shopping norms and customs peculiar to the society in which the parties hold membership. Hence, power relationships between retail agents and consumers have multiple contexts, one which is socially instituted in a cultural sense and another which is instituted by the marketing (or economic) system that specifies the rules for exchange transactions.

Dimensionality of the Power Construct

It has become commonplace for researchers to develop operationalized measures of power that posit that power is a five dimensional construct (e.g., see John 1984). This practice would seem to have its roots in the taxonomy of relationships from which social influence may be derived as identified in a landmark essay by French and Raven (1959). According to their taxonomy, there are five bases from which power is derived. These include the ability to reward another ("reward power"), the ability to coerce another ("coercive power"), the occupation of an instituted position ("legitimate power"), the ability to furnish one's expertise to another ("expert power"), and being popular or widely respected so that others wish build associations or connections with you ("referent power").

Nevertheless, use of such a scheme for dimensionalizing and measuring power perceptions seems ill advised. Certainly, the French and Raven taxonomy is useful for understanding the antecedents of dependency which give rise to power perceptions attributed to social actors. But it is essential to recognize that French and Raven (1959) themselves were careful to state that they have only described the bases from which power perception arise, not the dimensions of power itself.

In fact, a scheme for understanding antecdents is not necessarily appropriate for dimensionalizing a construct for measurement purposes. Among the foundations of test and measurement theory is that indicator measures of underlying constructs must have the property of being essentially unidimensional (Anderson and Gerbing 1988). It is this necessary condition, then, that makes adoption of the French and Raven (1959) taxonomy inappropriate as a framework for power measurement; the five bases of power are clearly not mutually exclusive. Reward power, for example, is not logically necessarily distinct from the other power bases -- say, for example, expert power -- inasmuch as the sharing of expertise by one party may rightly be considered reward influence rather than expert influence by the recipient.

An alternative scheme that seems workable is one based on the idea that the outcome of power is some degree of behavioral or fate control (Thibaut and Kelley 1959). Clearly, a social actor may attempt to control outcomes in only one of two ways. Either one may attempt to exert the influence one perceives he or she has over others, or one may attempt to resist or deflect the efforts of others to use the influence that he or she attributes to them. Hence, from this it follows that, within in any given context, power should be conceptualized as having two dimensions, influence and resistance. Influence Power, then, may be defined as the potential or ability that a social actor perceives he or she has, in general and within a given social context, to control the behavior of another. Similarly, Resistance Power is the potential or ability that a social actor perceives he or she has, in general and within a given social context, to deflect influence attempts perceived to be made by another. Integrating this dual dimensionality with the idea that power relationships between consumers and retail agents have both socially instituted and marketing system instituted aspects, it follows that it is appropriate to operationalize the consumer power construct as one consisting of four subscales: Social Influence (SI), Social Resistance (SR), Consumer Influence (CI), and Consumer Resistance (CR).

DEVELOPMENT OF CONSUMER POWER SCALES

In developing these four scales, psychological test construction procedures consistent with those outlined by Anastasi (1988) and Jackson (1967) have been followed. The first step was development of specific test item sets for each of the four scales. Second, statistical analyses evaluating the properties of items included in each set were performed to maximize and assess reliabilities of the scales. Finally, correlational analyses of the resulting scale measures with existing, conceptually related scales were undertaken to assess external validity.

Measures

All test items were constructed using a balanced six-point Likert scale ranging from Agree Strongly (1) to Disagree Strongly (6). Items included in the original SI pool were developed to reflect only the hypothesized influence dimension of power within a nonspecific social (i.e., "purely social") orientation. Care was taken to ensure that the wording of these items did not imply any specific social circumstance. In contrast, items included in the CI pool were created to suggest images of the influence dimension of power within a retail purchase setting. The wording of these items was designed to evoke images of buying products or services and dealing with retail salespeople. The items included in the SR and CR pools were developed in similar manners, except that the focus was placed on the hypothesized resistance, rather than influence, dimension of power within these two contexts. Finally, items intended for each pool were subjected to examination by colleagues and subsequently revised in an effort to ensure their face validity. These procedures resulted in the creation of 60 test items -- 19 for SI, 20 for SR, 13 for CI, and 8 for CR.

The measures chosen for use in the validity check analysis were the Powerful Others Control (POC) and Internal Control (IC) subscales of the 24 items measuring Internal-External Fate Control developed by Levenson (1974). Levenson's (1974) instrument includes three 8-item subscales, the third being Chance Control (CC), and represents a refinement to Rotter's (1966) 29-item Internal-External Control scale in that the items use a six-point (continuous) response set rather than the dichotomous Agree-Disagree/Unsure one chosen by Rotter. In addition, Levenson (1974) provided evidence suggesting that locus of control is a three dimensional construct rather than unidimensional, as Rotter (1966) had conceptualized it to be. The locus of control construct, particularly the POC dimension, has often been closely associated with, or even equated to, social dependency (Lachman 1986).

Given this, the nature of correlational results among the SR, CR, SI, CI, POC, and IC scales can be anticipated. As its name implies, the POC subscale is a measure of the influence that one perceives others have over him or her. Consequently, it is expected that this measure would be strongly associated with measures of one's perceptions regarding one's ability to resist influence attempts by others (e.g., the SR and CR scales) and weakly associated with measures of one's perceptions about one's ability to influence others (e.g., the SI and CI scales). Reasoning in a similar fashion, since the IC scale is intended to measure one's perceived ability to influence or direct one's own behavior, it should be more strongly related to the SI and CI than the SR and CR scales. Hence, empirical results consistent with these expectations would provide evidence of discriminatory and convergent validity of the SR, SI, CR, and CI scales.

TABLE 1

CHARACTERISTICS OF RESPONDENTS

Sample

Because of the preliminary nature of scale development work, respondent selection and recruitment relied upon convenience sampling techniques. First, the cooperation of 88 undergraduate and graduate students enrolled in the business and education schools of a major northeastern university was enlisted. Second, an additional 96 community dwelling adults were recruited to participate using a snowballing method. Selected characteristics of the total sample of 184 respondents are reported in Table 1.

Analyses and Results

The analyses began with an exploratory factor analysis having two purposes, to confirm the hypothesized dual dimensionality of power and to eliminate items that failed to reflect only a single power dimension. To do this, the item pools associated with each social context were combined, and separate exploratory analyses were run for the purely social and consumer item pools. Specifically, the CI and CR items were combined to form a pool totaling 21 items, and SI and SR items were combined to form a pool of 39 items.

In both analyses, results of the scree test and forced factor trials suggested the presence of two factors. For the consumer items pool analysis using a two factor solution with a varimax (orthogonal) rotation, 17 of the 21 items loaded primarily on one factor and, in all of these cases, the pattern was consistent with the hypothesized influence and resistance dimensions. Likewise, for the social items pool analysis, 26 of the 39 items loaded primarily on one factor and again, among these, the hypothesized pattern was observed. Those items not loading clearly on a single underlying factor were eliminated from further analysis to ensure unidimensionality of each scale, and the items corresponding to the influence and resistance dimensions were again separated for further analysis. This procedure and its corresponding results, like the earlier item drafting and revision process, provide strong evidence of the substantive validity of four distinct scales, each representing one of two power dimensions in one of two social contexts.

Further reductions in the number of items included in each scale were managed by computing corrected item-total correlations and eliminating those items with insignificant correlations, a procedure recommended when item and total score measures are both continuous (Norusis 1988). Each scale was reduced by removing the item with the smallest correlation coefficient in an iterative manner until all items exhibited corrected item-total correlations of .30 or greater. Because this criterion is four times the correlation standard error and it is standard practice to accept all items exhibiting a corrected item-total correlation greater than twice this standard error (Crocker and Algina 1986), the degrees of interitem consistency for each scale have been nearly maximized without risking wholesale eliminations of items.

These procedures resulted in the creation of a six item CI scale, an eight item instrument for CR, an eleven item scale for SI, and a five item instrument for SR. Computed values of Cronbach's alpha were .66 for CI, .75 for CR, .87 for SI, and .62 for SR. Although the reliabilities of the CI and SR scales are somewhat low, each scale has the desirable characteristic of only including items with relatively large standard deviations. Jackson (1967) has noted scales incorporating items featuring wide variances are desirable because the observed reliabilities are not a consequence of the inability of the items to indicate good separation among subjects. Consequently, these two scales would seem to offer relatively high discriminatory power. The items included in the final versions of each scale, along with selected reliability statistics, appear in the Appendix.

TABLE 2

CORRELATIONS BETWEEN SCALES

Finally, external validity of each scale was examined in order to confirm that the scales measure what they are intended to measure. Convergent and discriminant validity was assessed by computing correlational results of all scale measures with each other and with the POC amd IC subscales. Table 2 shows these results.

The relatively high positive correlations observed between the CR and SR scales (r = .53) and the CI and SI (r = .56) support the contention that these scale pairs measure the resistance and influence dimensions of power, respectively. Furthermore, in keeping with the weak positive relationship observed between POC and IC (r = .17), considerably weaker relationships between all pairs of resistance/influence scales (SR-CI, r = .29; CR-SI, r = .28; SR-SI, r = .36; CR-CI, r = .25) were observed as well.

In addition, both influence scales were positively correlated with the IC scale (CI, r = .24; SI, r = .34). It is also noteworthy that the relationship between the IC and SI scales was found to be stronger than that between the IC and CI scales; this too was to be expected in that the IC scale is a measure of social, rather than consumer, influence. Furthermore, little or no relationship was to be expected between the POC and the SI and CI scales because the correlation between POC and IC is weak (r = .17). Once again, the results were supportive (POC-CI, r = -.004; POC-SI, r = .13). In summary, then, these observations all lend support to claims of convergent and discriminant validity of the four power scales developed.

DISCUSSION

Research Applications

The results of this study support two important ideas about social power. First is that power is measurable when it is conceptualized as an attributed, as opposed to an inherent, characteristic of a social actor. Second is that, within any given social context, power may be viewed as consisting of two underlying dimensions, which might be labeled "influence" and "resistance".

Furthermore, it has been argued that power is an attributed outcome deriving from a social actor's dependency on another within a specific context or social situation. Specifically, it is dependency, not power, that has been posited to influence interpersonal behavior. This conceptual orientation, in combination with the present empirical results indicating adequate scale reliabilities and external validity, suggests that the power scales developed here may be useful for operationalizing the power construct in analyses of latent variable models of consumer interpersonal behavior.

Although one or more measures of locus of control (e.g., Levenson 1974; Rotter 1966) may be used to operationalize social dependency, such measures are general in nature and lack a contextual sensitivity. The General Systems Theory paradigm (e.g., see Berrien 1968; Miller 1955; von Bertalanffy 1968) reminds us of the importance of accounting for environmental influences and contextuality in the construction of theory; hence, it is proper that models of consumer behavior employ measures that reflect the influences of the social circumstances peculiar to the shopping and consumption planning experience. The scales developed here offer the opportunity to incorporate measures of power for use as a context sensitive indicator of a consumer's social dependency within a retail shopping environment and, in this respect, they represent an important addition to the social scientist's model building tool box.

APPENDIX

Study Limitations

It is essential to recognize that, due to the complexity and breadth of the research issue, the focus of this study was intentionally limited in two respects. First, the focus has been placed on measuring power attributed to the consumer with no consideration of retail agent power. Secondly, scale development efforts have been limited to consumer power perceptions attributed by the consumer. Certainly, the development of measures for power attributed to consumers by retail agents and for power attributed to retail agents by both consumers and retail agents are relevant to interpersonal behavior studies of retail market exchanges. Hence, further scale development efforts to measure these perceptions of retailer-consumer power relationships are to be encouraged in future studies.

The use of convenience sampling is also a limitation of this study. The failures to have obtained a probability sample and to have achieved consistently high levels of internal consistency among all four subscales threaten the generalizability of the findings. In view of the facts that power is an essential concept in the ontology of the exchange paradigm and that exchange is central to marketing behavior theory, further research and scale refinement efforts to improve or extend the present scales seem warranted.

REFERENCES

Anastasi, Anne (1988), Psychological Testing, Sixth Edition, New York: Macmillan Publishing Company.

Anderson, James C. and David W. Gerbing (1988), "Structural Equation Modeling in Practice: A Review and Recommended Two-Step Approach," Psychological Bulletin 103 (May), 411-423.

Bacharach, Samuel B. and Edward J. Lawler (1981), Bargaining, San Francisco: Jossey-Bass Publishers.

Bagozzi, Richard P. (1975), "Marketing as Exchange," Journal of Marketing 39 (October), 32-39.

Bagozzi, Richard P. (1979), "Toward a Formal Theory of Marketing Exchanges," in Conceptual and Theoretical Developments in Marketing, eds. O. C. Ferrell, Stephen W. Brown, and Charles W. Lamb Jr., Chicago: American Marketing Association, 431-447.

Berrien, F. Kenneth (1968), General and Social Systems, New Brunswick, NJ: Rutgers University Press.

Blau, Peter M. (1964), Exchange and Power in Social Life, New York: John Wiley & Sons.

Cartwright, Dorwin (1959), "Power: A Neglected Variable in Social Psychology," in Studies of Social Power, ed. Dorwin Cartwright, Ann Arbor: University of Michigan Press, 1-14.

Crocker, Linda and James Algina (1986), Introduction to Classical and Modern Test Theory, New York: Holt, Rinehart and Winston.

Cross, John G. (1969), The Economics of Bargaining, New York: Basic Books.

Dahl, Robert A. (1957), "The Concept of Power," Behavioral Science 2 (July), 201-218.

Dixon, Donald F. and Richard A. Layton (1972), "Initiating Change in Channel Systems," in New Essays in Marketing Theory,, ed. G. Fisk, New York: Allyn and Bacon, 315-332.

El-Ansary, Adel I. and Louis W. Stern (1972), "Power Measurement in the Distribution Channel," Journal of Marketing Research 9 (February), 47-52.

Emerson, Richard M. (1962), "Power-Dependence Relations," American Sociological Review 27 (February), 31-41.

Etgar, Michael (1977), "Channel Environment and Channel Leadership," Journal of Marketing Research 14 (February), 69-76.

Etgar, Michael (1978), "Selection of an Effective Channel Control Mix," Journal of Marketing 42 (July), 53-58.

French, John R. P. Jr. and Bertram Raven (1959), "The Bases of Social Power," in Studies of Social Power, ed. Dorwin Cartwright, Ann Arbor: University of Michigan Press, 150-167.

Gaski, John F. (1984), "The Theory of Power and Conflict in Channels of Distribution," Journal of Marketing 48 (Summer), 9-29.

Hunt, Shelby D. and John R. Nevin (1974), "Power in a Channel of Distribution: Sources and Consequences," Journal of Marketing Research 11 (May), 186-193.

Jackson, Douglas N. (1967), Personality Research Form Manual, Goshen, NY: Research Psychologists Press.

John, George (1984), "An Empirical Investigation of Some Antecedents of Opportunism in a Marketing Channel," Journal of Marketing Research 21 (August), 278-289.

Lachman, Margie E. (1986), "Personal Control in Later Life: Stability, Change, and Cognitive Correlates," in The Psychology of Control and Aging, ed. Margret M. Baltes and Paul B. Baltes, Hillsdale, NJ: Lawrence Erlbaum Associates, 207-236.

Levenson, Hanna (1974), "Activism and Powerful Others: Distinctions Within the Concept of Internal-External Control," Journal of Personality Assessment 38 (August), 377-383.

Lippitt, Ronald, Norman Polansky, and Sidney Rosen (1952), "The Dynamics of Power: A Field Study of Social Influence in Groups of Children," Human Relations 5 (1), 37-62.

Lusch, Robert F. and James R. Brown (1982), "A Modified Model of Power in the Marketing Channel," Journal of Marketing Research 19 (August), 312-323.

Miller, James G. (1955), "Toward a General Theory for the Behavioral Sciences," American Psychologist 10, 513-531.

Norusis, Marija J. (1988), SPSS/PC+ Advanced Statistics V2.0, Chicago: SPSS Inc.

Pahng, Sukbaum (1989), "A Theoretical Approach to the Study of Channel Negotiation Process," in AMA Proceedings, eds. Paul Bloom et al., Chicago: American Marketing Association, 73-78.

Porter, Michael E. (1974), "Consumer Behavior, Retailer Power, and Market Performance in Consumer Goods Industries," The Review of Economics and Statistics 56 (November), 419-436.

Pruitt, Dean G. (1981), Negotiation Behavior, New York: Academic Press.

Rollins, Boyd C. and Stephen J. Bahr (1976), "A Theory of Power Relationships in Marriage," Journal of Marriage and the Family 38 (November), 619-627.

Rotter, Julian B. (1966), "Generalized Expectancies for Internal Versus External Control of Reinforcement," Psychological Monographs: General and Applied 80 (No. 1, Whole No. 609), 1-28.

Tagiuri, R., J. Bruner, and R. Blake (1958), "On the Relations Between Feelings and Perceptions of Feelings Among Members of Small Groups," in Readings in Social Psychology. eds. E. Maccoby, T. Newcomb and E. Hartley, New York: Holt, Rinehart, and Winston, 110-116.

Tawney, R. H. (1931), Equality, New York: Harcourt, Brace.

Thibaut, John W. and Harold H. Kelley (1959), The Social Psychology of Groups, New York: John Wiley & Sons.

von Bertalanffy, Ludwig (1968), General Systems Theory, New York: George Braziller.

Wilkinson, Ian F. (1974), "Researching the Distribution Channels for Consumer and Industrial Goods: The Power Dimension," Journal of the Market Research Society 16, 12-32.

----------------------------------------