Power and Influence A Theoretical Bridge

http://www.jstor.org Power and Influence: A Theoretical Bridge Author(s): David Willer, Michael J. Lovaglia, Barry Markovsky Source: Social Forces, Vol. 76, No. 2, (Dec., 1997), pp. 571-603 Published by: University of North Carolina Press Stable URL: http://www.jstor.org/stable/2580725 Accessed: 05/05/2008 07:19 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=uncpress. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org. Power and Influence: A Theoretical Bridge* DAVIDWILLER, University of South Carolina MICHAEL J. LOVAGLIA, University of Iowa BARRY MARKOVSKY, University of Iowa Abstract Frequently social theorists conflate power and influence; often subsuming influence under a broad conception ofpower. Two contemporary theories separate them. Elementary theory has investigatedpower, status characteristics and expectations states theory has investigated interpersonal influence, and neither theory has considered the phenomenon of the other. We use the two theories to explain howpowerproduces influence and how influence produces power. We develop a theory that shows how the emotional reactions ofgroup members mediate the influence produced by power. We examine some new data and hypothesize that influence produces power. We trace the consequences when power and influence are opposed within a single relationship. Implications outside the limitations of the laboratory are discussed along with new hypotheses to be tested. Conceptions of power and influence are fundamental to the understanding of society. Consider the ways in which power and influence can occur in a social situation. A successful executive with a legendary work ethic asks a salaried employee to stay late to complete an important proposal. The employee agrees and cancels her plans for the evening. We would say that the executive used her influence to convince the employee to stay. Or, the exchange could have been more direct. The executive might have told the employee that if she stayed late, the executive would recommend her promotion. The executive has offered a reward in exchange for the employee's compliance (and implied a threat if she failed to comply). We would say that the *The authors thank The National Science Foundation for grants supporting theory developments and experimental research discussed in this article (SBR 94-23231 to Willer, SBR 94-22974 to Markovsky and Lovaglia, and SBR 95-15364 to Lovaglia and Markovsky). We appreciate comments on an earlier draft by Joseph Berger, Phillip Bonacich, Thomas J. Fararo, Edward J. Lawler, Frans Stokman, and Morris Zelditch, Jr. Direct correspondence to David Willer, Department of Sociology, University of South Carolina, Columbia, SC 29208. ? The University of North Carolina Press Social Forces, December 1997, 76(2):571-603 572 / Social Forces 76:2, December 1997 executive used power to induce the employee's compliance. Of course, power might be operating in the first scenario as well. Because of the executive's position as her superior, the employee might have perceived a threat or promise of reward though none was stated. Conceptions ofpower and influence are also fundamental to sociology. Both terms are used in many ways by diverse researchers. Influence, for example, is sometimes considered an aspect of power. However, more narrow definitions clearly separating the two concepts may have important advantages for social analysis. Recently, research programs have progressed by defining the domain of power phenomena more narrowly and analyzing it more rigorously. Network exchange theories of power have limited their definition to differences in network position that advantage certain actors when negotiating for resources.' Limiting the definition of power in this way excludes most forms of influence. In studies of status characteristics and expectation states, influence derives from expectations that group members have for each other's competence.2 Influence occurs when the advice of competent members is followed. Limiting influence in this way excludes the conditions of power studied by network exchange theories. Although research programs investigating power have developed independently from those investigating influence, some theoretical connections have been made. Here we build on previous work to explicitly link the elementary theory of structural power (Willer 1981a, 1981b; Willer 1987; Willer & Markovsky 1993) and status characteristics and expectations states theory as it applies to interpersonal influence (Berger et al. 1966; Berger & Conner 1974; Berger et al. 1985). Both theories are good examples of cumulative research progams (Szmatka & Lovaglia 1996); both have been tested in experimental programs as extensive as any in sociology; both have also been applied in the field.3 Their unusually high degree of development makes it possible to apply the theories jointly, once explicit links between them have been found. This article does not seek to integrate elementary theory and status characteristics theory into a single theory. Instead we bridge between the theories so that the two can be applied together. These joint applications relate power and influence in ways not possible for either program taken alone. We ask whether power produces influence and whether influence produces power. The answers that we seek have implications that are more general and richer than inferences from either theory taken alone. We offer hypotheses which, if supported, will contribute to the growth of both programs while overcoming limits of each. While not attempting an integration here, we do not reject theory integration as a long-term goal. To the contrary, these joint applications can serve as a feasibility study before the larger task of theory integration is taken up. Power and Influence / 573 Background and Theory We define power as the structurally determined potential for obtaining favored payoffs in relations where interests are opposed. It is the executive's position that gives her power over the employee, rather than anything intrinsic to the person occupying the position. We define influence in a way that clearly distinguishes it from power. Influence is the socially induced modification of a belief, attitude, or expectation effected without recourse to sanctions. The theoretical distinction between power and influence may or may not be warranted. Wrong (1979) adapts Russell's (1938) philosophical ideas to fashion a definition of power that encompasses influence: "Power is the capacity of some persons to produce intended andforeseen effects on others" (Wrong 1979:2, italics in original).4 Psychological definitions of power can be even more inclusive. For Heider (1958) power is a person's ability to accomplish something, to alter the environment whether human or nonhuman - in some way, while social power is a person's ability to cause another to do something. On the other hand, the concept of influence can include power. Zimbardo and Leippe (1992:2) define social influence as "the changes in people caused by what others do." Wrong (1979:4) asserts that "Power is identical with intended and effective influence" and French and Raven (1968:260) "define power in terms of influence and influence in terms of psychological change." These uses result in a diffuse concept expressed in two different ways depending on the context. When power and influence are identical, that A influences B's activity by changing B's beliefs is an example of power, whereas influence would be A changing B's behavior through threat of force.5 Although a delimited conception of power has proven easier to approach empirically, the broader view of power continues to spawn some research.6 Other theorists have sought to demarcate power and influence. For Parsons, power derives from "positive and negative sanctions" through which "ego may attempt to change alter's intentions" (1963a:338)7 whereas "influence is a way of having an effect on the attitudes and opinions of others" (1963b:38). This distinction is like that drawn earlier by Bierstedt, for whom "influence and power can occur in relative isolation from each other." For Bierstedt (1950), Karl Marx was influential upon the twentieth century, but he was not powerful. "Stalin, on the other hand, is a man of influence only because he is first a man of power" (1950:732). Zelditch (1992:995) draws the distinction more sharply, "What distinguishes power is that it involves external sanctions ... Influence, on the other hand, persuades B that X is right according to B's own interests." Mokken and Stokman's distinction is similar: "The exercise of influence takes place mainly by means of persuasion, information and advice" (1976:37), but, for power, "force, coercion and sanctions are sufficient" (1976:35, italics in original).8 A theoretical distinction between power and influence will only prove useful if it entails empirical consequences. Table 1 shows the different antecedents and consequences of power and influence as the terms are used in the theories upon 574 / Social Forces 76:2, December 1997 which we focus, and the empirical tests in which those theories have been applied. Briefly, elementary theory locates power in the structure of exchange networks and, when power differences occur, predicts different payoffs for exchanging actors. Status characteristics theory locates interpersonal influence in the status (prestige) order of a group. When status differences occur, low status actors alter their behavior to conform with advice of high status actors because that advice is expected to be competent and beneficial to the group. The two theories do not. cover all kinds of influence and power. For example, status characteristics theory does not deal with the effects of persuasion, and elementary theory does not predict power from negotiation styles. Furthermore, here we will focus only on the network exchange theory component of elementary theory,9 not considering elementary theory's applications to coercive relations with negative sanctions. Nevertheless, the definitions that the two theories provide for power and influence apply to a broad base of literature. In defining power, Weber ([1918] 1968) and Aron (1988) focus on the gain to the high power actor, while Dahl (1957, 1968) and Lukes (1974) focus on the loss to the low power actor. Because the larger payoffs of high power actors result from low power actors' smaller payoffs, power in exchange networks links gains to losses. Wrong's and French and Raven's definitions of power quoted earlier emphasize control. Elementary theory is more specific and deals with control only as it relates to valued outcomes. In regard to the belief change brought about by influence, we mean an expectancy regarding a property of an object or event (Rotter 1972). By implication, expectations need not be consciously recognized by those who hold them. To detect influence it is sufficient to note behavior change in the absence of sanctions for that behavior. In experiments, the acceptance or rejection of influence is measured by the difference between the initial and final decision of a subject given disagreement with another actor when that actor is not capable of rewarding or punishing the subject (Berger et al. 1977). Having two hitherto independent theories, elementary theory and status characteristics theory, which have not investigated the central phenomena of the other, we begin with power and influence sharply demarcated. Our initial treatment of power and influence as distinct phenomena will not dictate our conclusions, however. Because we bridge between the two, a convergence is possible. For example, if power and influence freely produce each other from similar conditions, the terms should be merged. Alternatively, relations may be asymmetric requiring power and influence to be kept distinct but understood as related phenomena. As our argument unfolds it will become apparent that the distinction between power and influence is useful and that the relationship between them is complex. For example, whether power confers influence depends in part on emotional reactions and we offer new formulations to explain these mediating factors. The ideas of the article are developed in the following sequence: (1) elementary theory as it is applied to power in exchange networks; (2) influence as formulated by status Power and Influence / 575 TABLE 1: Structures and PredictedEvents for the Two Theories Theory Type of Structure Events Predictions ElementaryTheory Exchange Network Power Exercise High vs. Low Payoffs Status Characteristics Status Order Interpersonal Stable vs. Changed Theory Influence Beliefs characteristics theory; (3) theoretical work explicating the effects of structural power on interpersonal influence and new data on the mediating role of emotional reactions; (4) theoretical work explicating the effects of status expectations and influence on power in exchange networks; and (5) implications when power and status are opposed in social relationships. EXCHANGE STRUCTURES AND POWER Power occurs in relations where the interests of actors - individual or corporate - are opposed but complementary (Wilier & Markovsky 1993). Exchange relations are mixed motive games and interests are opposed in that agreements with higher payoffs to one actor necessarily have lower payoffs to the other (Willer 1985). Interests are complementary in that neither actor can benefit without reaching some agreement. In many experiments, exchange is operationalized as the mutual transfer of rights to portions of valued resource pools to actors A and B (Markovsky, Willer & Patton 1988). In such cases, actors' interests are opposed in the sense that the greater the proportion of the resource pool that is gained by one actor, the less the other can receive.10 At the same time, A and B have complementary interests in cooperating with one another: Agreements for resource allocations must be mutual. Power occurs, for example, when A can demand more from B without fear of loss - as when A has a partner who serves as an exclusive alternative to B, but B has no such alternatives. A's profit in the relation will be greater than B's. Network exchange theories distinguish "power structure" which refers to network conditions like exclusion from "power exercise," the observable consequences of structural power. Exchange researchers measure power exercise by actors' relative payoffs: the actor gaining more exercises power (Bienenstock & Bonacich 1992; Cook & Emerson 1978; Markovsky et al. 1988). In experimental settings, the situations of two actors are hypothesized to be identical except for their position in the exchange network. This allows the inference that the power structure produced different payoffs for the actors, such that some exercised power over others. Individual differences such as negotiating style are controlled by randomly assigning subjects to network positions. 576 / Social Forces 76:2, December 1997 Elementary theory predicts power from two conditions: an actor's best expected payoff called "Pmax," and the actor's expected payoff at confrontation, i.e., when negotiations break down, called "Pcon" (Willer & Anderson 1981; Willer & Markovsky 1993). Consider the exchange relation between actors A and B where they divide ten resources that are lumpy in units of one, only bundles of whole units can be exchanged. Here PAmax = 9: A hopes at best for nine, leaving the minimum of one as the inducement for B to exchange. If neither A nor B has a source of profit outside the relation, the relation is symmetrical so Pmax is also 9 for B. If the two fail to agree on a division, they are in confrontation and both gain zero. That is to say, PAcon = 0 and PBcon = 0. Neither A nor B has a power advantage and equal division of the resource pool is predicted. But now assume A has a fixed alternative of six units that can only be realized if A does not exchange with B. In this case A need not accept less than six from B, so A's payoff at confrontation is now six, in contrast to B's zero. That is to say, PAcon = 6 and PBcon = 0. Furthermore, now B can hope for no more than four from A. So now Pmax for B is 4, but Pmax for A remains at 9. When A and B allocate the resource pool, A has an advantage and will acquire more resources than B. The difference in acquired resources indicates A's power exercise over B. The research problem is to predict that resource division. RESISTANCE THEORY Resistance theory is the branch of elementary theory that systematically relates expected best payoffs and payoffs at confrontation to predict resource divisions and thus the amount of power exercised in exchange relations (Wilier 1981b, 1987). When PA is A's predicted payoff from exchange and, as explained above, PA max is A's best expected payoff and PAcon is A's payoff at confrontation, A's resistance is: RA - PA max-PA (1) PA -PA con Exchange occurs when the resistance of two actors is equal. Then PA max- PA PB max- PB RA- =- = RB (2) PA PA con PB-PBcon For the example above, A and B have a 10-point resource pool to divide, while A has a fixed offer of 6 as an exclusive alternative to exchanging with B such that PAmax = 9, but PB max = 4 while PA con = 6 but PB con = 0. It follows that 9-PA- 4-P8B3 RA = = RB (3) PA-6 PB Since for any agreement PA + PB = 10, we solve for PA = 7.71 and PB = 2.29. The inequality of this resource division is the amount of power A exercises over B. Power and Influence / 577 Power Structures Power in exchange networks results when some positions are more likely than their potential exchange partners to be excluded from exchange and from receiving the profits that thereby accrue (Markovsky et al. 1988; Markovsky et al. 1993). In the 2-branch of Figure 1, A is connected to B and C, and both relations contain ten valued resource units. Assume that only one agreement is possible in any exchange round. If A reaches an agreement with B, then C is excluded from exchange and profit. If A reaches agreement with C, then B is excluded. On the other hand, A need never be excluded. In a series of exchange rounds, the agreement reached by B or C with A in the previous round determines best hopes and payoff at confrontation for the actors in the current round. The production of power differences in exchange networks is modeled by a series of calculations similar to the analysis of the A - B exchange above where A had a fixed alternative. Now any offer from either B or C affects the other relation as did the alternative payoff. For example, let B offer A an equal division, PA