Compliance with bases of power and subordinates' perception of superiors: moderating effect of quality of interaction.

Gupta, Bindu ; Sharma, Narendra K.

Abstract

Social power is critical for superior-subordinate relationship to deliver desirable outcomes. This empirical study examines the power interaction framework in the Indian context by analysing responses of 428 lower- and middle-level executives from 19 organisations through an investigation of the effectiveness of superior's bases of power on the subordinates' willingness to comply. Multivariate analyses revealed that subordinates' compliance with soft bases of power had higher effect than with harsh bases of power on their satisfaction and commitment with superior and global influence of superior on subordinate. The study identifies the moderating role of quality of interaction in the relationship between bases of power and compliance and subordinates' attitude towards superiors.

Keywords: Commitment, Global influence, Harsh bases of power, Power interaction model, Soft bases of power

**********

Although the concept of power often evokes negative impressions, power almost always exists in organisations. Recognising and man aging power can be very healthy for organisations and personnel. McClelland reported that high need for power characterises successful managers (McClelland and Boyatzis, 1982). Kanter (1977, 1979) argued, based on her observation of organisations, that it is powerlessness that impedes an organisation's work. Subordinates expect and want their managers to be powerful so that managers can provide them with the resources necessary to do their job. The process of exercising power serves as one of the key characteristics which define the relationship between supervisor and subordinate (Bennis and Nanus, 1985; Pfeffer, 1992). Research has demonstrated that managerial advancement and success are contingent on effective use of influence (Bass, 1994; Yukl, 1994).

This paper tests the effectiveness of different bases of power and the interactions between social power and quality of interaction between superior and subordinate for gaining compliance. It examines the effect of these interactions on subordinates' attitudes towards superiors.

Social Power

Issues of power and influence have been widely elucidated (for example Hobbes, 1968; Lukes, 1974, Machiavelli, 1984; Nietzsche, 1968; Russell, 1938; Weber, 1948). The social psychological study of power and influence finds its origin in the theorising of Kurt Lewin ( 1941). Following Lewin's initial conceptualisation, French and Raven (1959) defined influence as a force one person (the agent) exerts on someone else (the target) to induce a change in the target, including changes in behaviours, opinions, attitudes, goals, needs, and values. Social power was subsequently defined as the potential ability of an agent to influence a target.

Bases of power can be defined as the resources and characteristics a person has in order to be able to influence others. French and Raven (1959) classified bases of power into five categories, namely reward, coercive, referent, expert, and legitimate power. Later, Raven (1965) added information power as a new basis of power. The taxonomy of social power suggested by French and Raven (1959) has been adopted by several psychology and management researchers (Hinkin and Schriesheim, 1989; Pfeffer, 1981; Podsakoff and Schriesheim, 1985; Nesler, Aguinis, Quigley, and Tedeschi, 1993).

Raven (1992, 1993) extended and reworked the original bases of power into power/interaction model of interpersonal influence. The power/interaction model reclassified the original bases of power by subdividing reward and coercive powers into two types: 'personal' (intangible forms of reward and coercive power) and 'impersonal' (tangible forms of reward and coercive power). Legitimate power was divided into four categories: Legitimate reciprocity-based on the agent having done something positive for the target; Legitimate equity--based on compensating for either (a) hard work or sufferance by the agent or (b) harm inflicted by the target; Legitimate dependence--based on social responsibility, to assist another who is in need; and Legitimate position--attributed to the right one has because of status or position.

In one of the first empirical studies of Raven's new taxonomy, Raven, Schwarzwald, and Koslowsky (1998) developed a scale, the Interpersonal Power Inventory (IPI), for measuring compliance with the 11 power bases. Their analysis revealed that power bases are not independent and they identified two underlying structures or factors, 'soft' and 'harsh'. The soft factor includes expert, referent, information, and dependence legitimacy. The harsh factor incorporates personal and impersonal coercion, personal and impersonal reward, position legitimacy, equity legitimacy, and reciprocity legitimacy. Position legitimacy also loaded high on soft factor. This dichotomy has been supported by further research also (Erchul, Raven, and Ray 2001; Koslowsky, Schwarzwald and Ashuri, 2001; Schwarzwald, Koslowsky, and Agassi, 2001). As not all power bases are available in each situation, comparisons that use the gross measures of harsh and soft tactics, rather than individual tactics, are preferable (Koslowsky and Schwarzwald, 2001).

The first objective of the present study is to examine whether the soft-harsh dichotomy of power bases exists in Indian context, given the differences between the United States and Indian cultures. The US society is characterised by a high level of individualistic orientation and low power distance. Indian culture, on the contrary, is high on power distance and has collectivist orientation.

In his power interaction model, Raven (1992, 1993) argued that the choice of specific influence strategies by an agent is rational but is also affected by situational, personal, and normative factors. Studies have reported that organisational level or position in the firm's hierarchy plays an important role in moderating the effectiveness of influence approaches (Vecchio and Sussmann, 1989; Gupta and Sharma, 2003). Koslowsky and Schwarzwald (1993) argued that the setting determines to a great extent the choice of influence strategy. Their results indicated that power repertoire of higher-status individuals is larger than that of lower-status ones.

In a study examining the effect of gender on the usage of influence strategies (Bui, Raven, and Schwarzwald, 1994), men tended to resort more frequently to harsh bases as their satisfaction with the relationship decreased. In contrast, women tended to resort to soft bases, regardless of satisfaction. Schwarzwald and Koslowsky (1999) found that low self-esteem adolescents resorted more frequently to specific harsh bases as compared to high-self esteem adolescents. One study examined the impact of leadership style on compliance with power bases and reported that officers who worked for high transformational captains reported a significantly greater likelihood to comply with both harsh and soft power bases than their cohorts who worked for low transformational captains (Schwarzwald et al, 2001). This background set the second objective of the study, which is to examine the impact of quality of interaction between superior and subordinates on subordinates' compliance with superior. The study also examined how compliance with given bases of power influences subordinates' attitude towards superior and overall influence of superior on subordinates.

Quality of Interaction

Leader-member exchange (LMX) focuses on the quality of the relationship between a leader and individual members of the group. An essential premise of leader-member exchange theory is that leaders and supervisors have limited amounts of personal, social, and organisational resources (for example, time, energy, role, discretion, and position power), and thus, distribute such resources among their subordinates selectively (for example, Dansereau, Graen, and Haga, 1975; Graen and Scandura, 1987; Graen and Uhl-Bien, 1995). It suggests that leaders do not use the same style in dealing with all subordinates but rather a different type of relationship or exchange develops with each subordinate (Dansereau et al, 1975; Graen and Cashman, 1975; Liden and Graen, 1980; Graen, Novak and Sommerkamp, 1982; Graen and Scandura, 1987). Theoretically speaking, these interactions have been dichotomised by various researchers as 'in- and out-group relationship' (Graen and Cashman, 1975), 'leadership exchange', and 'supervisory exchange' (Liden and Graen, 1980), and high- and low- quality relationships (Graen and Schiemann, 1978).

Studies have reported differences between in- and out-groups with regard to the leader's use of different bases of power (Graen and Cashman, 1975; Thibodeaux III and Lowe, 1996). In-group members reported greater use of expert and referent power bases than out-group members. Out-group members reported a greater use of coercive power than did in-group members. Hunt (1984), in referring to French and Raven's (1959) conceptualisation of power bases, argued that leadership is the use of personal (that is, soft) power bases whereas supervision is the use of organisational (harsh) power bases. Eden (1984, 1990) reported that leaders' expectations and attitude towards followers influence their treatment of subordinates. Leaders with high expectations are likely to behave to convey same expectations by encouraging participation. Leaders with low expectations are likely to convey reduced expectations by exercising close supervision. High quality in-group exchange has been associated with subordinates' job satisfaction (for example, Graen et al, 1982), performance ratings (for example, Liden and Graen, 1980), low turnover (for example, Graen, Liden, and Hoel, 1982) and promotion (for example, Wakabayashi, Graen, Graen and Graen, 1988).

Earlier studies examining the effectiveness of bases of power in terms of subordinates' compliance (for example, Dunne, Sthal, and Malhart, 1978; Rahim, 1989; Rahim and Buntzman, 1989; Rahim, Kim, and Kim, 1994; Sharma and Gupta, 2000; Thamhain and Gemmill, 1974; Yukl and Falbe, 1991) reported that legitimate, expert and referent bases of power are positively associated with compliance. Recent studies using the power interaction model reported more compliance with information, position legitimacy, expert, and dependence legitimacy than with equity legitimacy, impersonal coercion, reciprocity legitimacy, and impersonal reward (Gupta and Sharma, 1999; Koslowsky et al, 2001; Raven et al, 1998). According to power interaction model, an influencing agent who is liked tends to arouse greater compliance even if he/she resorts to unfavourable strategies.

By contrast, a disliked agent arouses resistance as the target attempts to thwart any gains for the former. Raven contends that, in the latter case, the efficacy of power strategies such as expertise and information is diminished, thereby, forcing the disliked influencing agent to resort to harsh strategies. The above evidences suggest that different quality of interaction would evoke different degrees of compliance with bases of power.

We formulated the following hypotheses for power compliance relationship and role of quality of interaction in this relationship:

* H 1: There will be more compliance with soft bases of power than with harsh bases of power;

* H 2: There will be more compliance when subordinates have high quality of interaction with superior than low quality;

* H 3: There will be more compliance with soft bases of power in high quality of interaction than in low quality of interaction; and

* H 4: There will be more compliance with harsh bases of power in low quality of interaction than in high quality of interaction.

Compliance with Power Resources and Subordinates' Attitudes

The power/interaction model describes possible consequences of an influence attempt on the part of target; various effects other than private acceptance or public compliance can occur as a result of an influence attempt, like changes in the target's perceptions of self and of the agent and changes in the power relationships. The third objective of the study is to examine the relationship of subordinates' compliance with their attitudes towards superiors, specifically satisfaction and commitment. Studies have examined the effect of compliance with soft and harsh bases of power on organisational attitudes.

Koslowsky et al (2001) reported that job satisfaction is positively related with compliance with soft power bases and negatively with harsh ones, whereas organisation commitment is positively associated with both. The present study contends that satisfaction and commitment with superior will be more immediate outcomes of the compliance with the bases of power used by the superior. The effect of superior bases of power on organisation attitude, that is,. job satisfaction and organisation commitment may be contaminated by other factors than superior behaviour, which needs further investigation. We hypothesised:

* H 5: Compliance with soft bases of power will positively influence satisfaction with superior;

* H 6: Compliance with soft bases of power will positively influence commitment with superior;

* H 7: Compliance with harsh bases of power will negatively influence satisfaction with superior; and

* H 8: Compliance with harsh bases of power will negatively influence commitment with superior.

Superiors have influence in different areas of subordinates' work life. The present study tried to examine the contribution of power a supervisor has towards overall influence of superiors over subordinates. In the study, overall influence was named as global influence. Here subordinates were asked to indicate the extent of overall influence the superior has on them. With the help of this measure, relative impact of the different bases of power (soft/harsh) can be measured in a given situation. It was hypothesised:

* H 9: Compliance with soft bases of power will positively affect global influence;

* H 10: Compliance with harsh bases of power will negatively affect global influence.

Compliance with Power Resources and Subordinates' Attitudes: Quality of Interaction as Moderator

In this study the quality of interaction has been hypothesised to moderate compliance with given bases of power. It is also assumed that differential quality of interaction will also affect subordinates' attitude towards superiors. According to Sparrowe and Liden (1997), LMX may be viewed as a sponsorship process through which subordinates are either assimilated into the leader's 'inner circle' or become isolated from this important informal network. Thus, an individual with a high quality LMX is likely to be assimilated into the leader's informal network of relationships, whereas a subordinate with low LMX is not.

The following hypotheses were proposed:

* H 11: Quality of interaction will influence the relationship between soft bases of power and subordinates' satisfaction with superior;

* H 12: Quality of interaction will influence the relationship between soft bases of power and subordinates' commitment with superior;

* H 13: Quality of interaction will influence the relationship between soft bases of power and global influence;

* H 14: Quality of interaction will not influence the relationship between harsh bases of power and satisfaction with superior;

* H 15: Quality of interaction will not influence the relationship between harsh bases of power and commitment with superior; and

* H 16: Quality of interaction will not influence the relationship between harsh bases of power and global influence.

Method

Participants

The participants were lower- and middle-level executives (N = 428) from various departments of 19 organisations. The average age of executives was 42.3 (SD = 8.4) years, average experience in current organisation 16.2 (SD = 9.0) years, average experience in present position 6.0 (SD = 6.0) years and average experience in working under the present superior 4.4 (SD = 5.4) years.

Instruments

Interpersonal Power Inventory (IPI) developed by Raven et al, (1998) which includes multiple items tapping the 11 bases of power was used to measure bases of power. The scale has been shown to have high reliability as measured by internal consistency ranging from 0.67 to 0.86 (Raven et al, 1998).

IPI consists of 33 reasons, three for each of 11 bases of power. Using a critical-incident-type technique, subjects were told:

Often supervisors suggest new methods or procedures of working to their subordinates. Sometimes subordinates resist the new procedures, while at other times they do exactly as they are asked to do.

Think of the last time when your supervisor suggested, requested, or commanded that you follow a different procedure or method of working, one which you would not otherwise be inclined to follow. Though initially reluctant, you did exactly what your supervisor asked to do (that is, you complied with him/her). If you have not yet experienced such a situation, please try to imagine such an event.

Here are a number of reasons why you would do what your supervisor requested. Read each of these reasons carefully and decide how likely it would be the reason you would comply with your supervisor in such an instance. Indicate how likely it would be that you complied for this reason. Subjects were provided with seven alternatives ranging from 'extremely unlikely' to 'extremely likely'.

The instrument to measure quality of interaction was based on Tondon (1990) and Liden and Maslyn (1993). The scale consisted of 18 items, representing three dimensions of quality of interaction: perceived contribution, affect, and loyalty. The respondents were instructed to evaluate on a seven-point scale the degree to which each item was true of their interaction with their supervisor/subordinates (1 = not at all; 7 = very much). The reliability coefficient (alpha) of quality of interaction measure was found to be 0.88

A scale consisting of 10 items was developed to measure the subordinates' satisfaction with their immediate superior. Respondents rated these items on a seven-point scale indicating their degree of agreement or disagreement with each item (1 = strongly disagree; 7 = strongly agree). The scale showed the reliability coefficient (alpha) of 0.70.

Commitment was measured by using O'Reily and Chatman (1986) framework. O'Reily and Chatman (1986) adopted Kelman's (1958) processes of attitude change to specify three bases of psychological attachment to the organisation: Compliance, intemalisation, and identification. In the present study commitment with superiors was measured with 12 items related to subordinates' compliance, intemalisation, and identification with superior. The scale has high reliability coefficient (alpha) of 0.86.

Seven items were used to measure global influence of superior in different work behaviour and outcomes of subordinates. Respondents were asked to judge, on a seven-point scale, the degree to which each statement was true of their relationship with immediate superiors (1 = never; 7 = always). The scale showed a reliability coefficient of 0.75.

The questionnaire also included general questions on rank, gender, age, education, and experience with present organisation, with present superior and total experience.

Procedure

Participants were contacted personally with due permission of their organisations. They were briefed about the purpose of the study, and instructions regarding how to fill the questionnaire were given.

Analyses

Data were analysed using descriptive and inferential statistical techniques. Data obtained through Interpersonal Power Inventory were factor-analysed. Descriptive statistics, correlations, and reliability coefficients were computed for each measure. Hypotheses were tested using Manova and regression analysis. Separate analyses were conducted for each of the three dependent variables, namely satisfaction with superior, commitment with superior and global influence.

Results

Data Reduction

Before conducting factor analyses, sampling adequacy test was computed using Kaiser-Meyer-Olkin (KMO) statistics. All the items were found to have KMO value higher than 0.60. The KMO statistics for the sample data set was 0.86 and Bartlett test showed that non-zero correlations existed at the 0.000 significance level.

We initially factor analysed the 11 bases of power to examine whether the dichotomy 'harsh-soft' exists in the Indian context also. The average score for 11 bases of power were subjected to principal-component factor analysis with varimax rotation using SPSS (version 10.0). The analysis yielded a two-factor solution (Table 1). The first factor, which explained 31.34 per cent of the variance, represented 'soft power' and included six bases of power (information, referent, position legitimacy, personal reward, dependence legitimacy, and expert).

The second factor, representing 'harsh power', explained 29.22 per cent of the variance and included five bases of power (impersonal coercive, reciprocity legitimacy, equity legitimacy, personal coercive, and impersonal reward). Based on these results, two factor scores were calculated for each participant. The first represented mean compliance to soft sources (alpha = 0.84) and the second, to harsh sources (alpha = 0.82). These factor scores were used for further analyses.

Degree of Compliance

Table 2 presents the means and standard deviations for the likelihood that compliance can be attributed to each of the 11 bases of power, ordered by size of the means. The table indicates that some bases of power (such as information power, dependence legitimacy, and expert power) were more likely to be the reason for compliance, while others (equity legitimacy, personal coercive and impersonal coercive) were less likely. Table 2 also shows that subordinates attributed more compliance to soft bases of power than harsh and support H 1.

Quality of Interaction and Compliance with Bases of Power

To examine the impact of quality of interaction on compliance with bases of power, a one-way Manova was conducted. For quality of interaction, individuals' item scores were summed and the totals divided at the median of 5.17, with high scores indicating in-group membership (n = 209) and low scores indicating out-group membership (n = 219). Scores on the IPI were used as the dependent measures. Significant results were obtained for the main effect, of compliance with bases of power: F (2,426) = 675.38, p < 0.001 and quality of interaction: F (1,426) = 12.56, p < 0.001; as well as interaction effects, the interaction between compliance with bases of power and quality of interaction: F (1,426) = 28.38, p < 0.001.

The results of follow-up univariate analysis, presented in Table 3, indicate the significant impact of quality of interaction for soft power sources F (1, 426) = 41.07, p <0.001. Subordinates reported greater compliance with soft bases of power when quality of interaction was high (M = 5.17, SD = 0.82) than when it was low (M = 4.65, SD = 0.86) which is in agreement with stated H 3. Subjects reported no significant difference in compliance with harsh bases of power in low or high quality of interaction. This finding does not support H 4.

Univariate analysis for 11 bases of power and quality of interaction indicated significant difference in compliance for all bases of power in low and high quality of interaction except equity legitimacy and personal coercive. For impersonal coercive basis of power, subordinates reported more compliance in low quality of interaction (M = 3.49, SD = 1.43) than in high quality of interaction (M = 3.01, SD = 1.35). For other bases of power, subordinates indicated more compliance in high quality of interaction than in low.

Compliance with Bases of Power and Subordinates' Attitude and Moderating Effect of Quality of Interaction

Table 4 displays zero-order correlations among the study variables. The soft bases of power are positively correlated with satisfaction with superior, commitment with superior, and global influence with correlation coefficients ranging from 0.34 to 0.47. In addition, harsh bases of power showed a significant positive correlation with commitment with superior (r = 0.16) and global influence (r = 0.19) and a non-significant negative correlation with satisfaction with superior.

We conducted moderated hierarchical regression analyses for soft and harsh bases of power separately for each dependent variable, namely, satisfaction with superior, commitment with superior, and global influence. The results are reported in Tables 5 and 6.

Step 1 of the analysis includes bases of power and test of hypotheses from H 5 to H 10, and Step 2 includes hypothesised bases of power and interactions between bases of power and quality of interaction and test of hypotheses from H 11 to H 16. The interaction term variables were mean-centered to reduce potential multi-collinearity effects (Aiken and West, 1991; Jaccard, Wan, and Turisi, 1990).

Main Effect of Bases of Power on Satisfaction, Commitment and Global Influence

Results of regression analysis for soft bases of power support hypotheses H5, H6, and H9. H5 posited a positive relationship between compliance with soft bases of power and satisfaction with superior and it was supported ([beta] = 0.34, p < 0.01, Table 5). H6 and H9, stating positive influence of soft bases of power on commitment with superior and global influence, also got support ([beta] = 0.45, p < 0.01; and [beta] = 0.36, p < 0.01 respectively, Table 5).

Hypothesis H7 predicted that harsh bases of power will negatively influence satisfaction with superior. As shown in Table 6, this hypothesis was not supported in the study. Further, harsh bases of power significantly and positively influenced commitment with superior ([beta] = 0.16, p < 0.01) and global influence ([beta] = 0.19, p < 0.01), which is contrary to the direction of relationship stated in hypotheses (H8 and H10).

To obtain further insight into the relationship between bases of power and dependent variables, individual bases of power within each category were entered in regression equations for soft and harsh bases of power separately (see Table 5 and Table 6). Among six soft bases of power only expert power was found to have positive effect on satisfaction with superior ([beta] = 0.23, p < 0.01, Table 5).

Referent power, personal reward, and expert power were significantly related to commitment with superior ([beta] = 0.17, p < .01; [beta] = 0.25, p < 0.01 and [beta] = 0.24, p < 0.01 respectively, Table 5). For global influence, significant effects of referent power ([beta] = 0.16, p < 0.01) and expert power ([beta] = 0.23, p < 0.01) were obtained. From harsh power category, impersonal coercive and impersonal reward were found to be significantly related with satisfaction with superior ([beta] = -0.43, p < 0.01; [beta] = 0.21, p < 0.01 respectively, Table 6). For commitment with superior, significant effects of impersonal coercive ([beta] = 0.32, p < 0.01), legitimate reciprocity ([beta] = 0.14, p < 0.05), legitimate equity ([beta] = 0.11, p < 0.05) and impersonal reward ([beta] = 0.28, p < 0.01) were obtained. Impersonal coercive, personal coercive and impersonal reward were found to have significant effect on global influence ([beta] = -0.12, p < 0.05; [beta] = 0.14, p < 0.05 and [beta] = 0.16, p < 0.01 respectively; Table 6).

Moderating Effect of Quality of Interaction

Model 2 tests moderating effects. The addition of terms involving interaction between soft bases of power and quality of interaction increases [R.sup.2] in satisfaction with superior, commitment with superior, and global influence by 4 per cent, 4 per cent, and 1 per cent respectively (see Table 5).

As postulated (H11, H12, and H13), quality of interaction moderated the relationship of compliance with soft bases of power with satisfaction with superior ([beta] = -0.19, p <.01), commitment with superior ([beta] = -0.19, p < .01), and global influence ([beta] = -0.11, p <.05). Interaction of individual soft bases of power with quality of interaction increases [R.sup.2] by 6 per cent for satisfaction with superior, by 6 per cent for commitment with superior, and by 3 per cent for global influence. For satisfaction with superior, significant interactions were found between information power and quality of interaction ([beta] = -0.11, p < .01) and personal reward, and quality of interaction ([beta] = -0.14, p < .05). Interactions of referent power, personal reward and dependence legitimacy with quality of interaction were found to be significant for commitment with superior ([beta] = -0.14, p < .05; [beta] = -0.13, p < .05; and [beta] = 0.14, p < .01 respectively, Table 5). For global influence, significant interactions were found between information power and quality of interaction ([beta] = -0.12, p <.05) and referent power and quality of interaction ([beta] = -0.14, p<.05).

For compliance with harsh bases of power and subordinates' attitude, the moderating effects of quality of interaction were not significant which also support the stated hypotheses

(H14, H15, and H16). Interaction of individual harsh bases of power with quality of interaction increase [R.sup.2] by 6 per cent for satisfaction with superior, by 5 per cent for commitment with superior, and by 1 per cent for global influence. Interaction between impersonal reward and quality of interaction was significant for satisfaction with superior ([beta] = -0.18, p < .01) and between personal coercive with quality of interaction for global influence ([beta] = -0.15, p < .05).

Discussion

One of the objectives of the present study was to examine the soft-harsh power-bases dichotomy in Indian context. The results of factor analysis indicated that the 11 bases of power discussed in power interaction model fall into two major categories, harsh and soft bases. These findings are in congruence with the findings of studies conducted in other cultural context (Koslowsky et al, 2001; Raven et al, 1998; Schwarzwald et al, 2001), which indicates that definitions of soft and harsh bases of power are similar across cultural settings. However, the percentages of variance explained by soft and harsh categories in other cultural context and in the present study were at variance. Previous research conducted on American and Israeli students/co-workers (Raven et al, 1998), Israeli nurses (Koslowsky et al, 2001) and Israeli police captains (Schwarzwald et al, 2001) reported more variance accounted for by harsh category (35 per cent, 40 per cent, 42 per cent and 46.8 per cent respectively) than by soft (25 per cent, 20 per cent, 16 per cent, and 15.3 per cent respectively). On the contrary, the present study revealed that more variance was explained by soft bases (31.34 per cent) than by harsh bases (29.22 per cent). These variations may be because of work settings differences or culture variations, which indicate the need of further research. Cultural diversity may be reflected in the form of high versus low cultural context in which organisations work. Members of high-context cultures (for example, China, Japan, and South Korea) desire to establish trust first, value personal relations, and their negotiations tend to be slow and ritualistic (Munter, 1993). In contrast, members of low context cultures (for example, Canada, Germany, and the United States) desire to get right down to business, value performance and expertise, and strive to make their negotiations as efficient as possible (Munter, 1993). India being in particular a high-context culture, supervisors and subordinates might adopt communication strategies that make soft bases of power consistent with the cultural context.

In his power interaction model Raven (1992, 1993) argues that the choice of specific influence strategies by an agent is also affected by situational, personal, and normative factors. The focus here was on the quality of interaction between superior and subordinate as potential correlate of subordinates' compliance. The quality of these relationships determines the amount of physical or mental effort, material resources, information, and/or social support exchanged between the supervisor and subordinate (Liden, Sparrowe, and Wayne, 1997). Dienesch and Liden (1986) reported that high LMX members consistently received more informal rewards than low LMX members. We expected more compliance with soft bases of power in high quality of interaction and more compliance with harsh bases of power in low quality of interaction. The findings of the study clearly support the hypothesis regarding the advantage of high quality of interaction over low quality of interaction in obtaining compliance. Subordinates reported greater compliance in high quality of interaction (M = 4.48) than in low quality of interaction (M = 4.19). More compliance was also found with soft bases of power in high quality of interaction than in low quality of interaction.

Overall, the present study indicates that subordinates attributed more compliance to soft basis of power than harsh even in low quality of interaction. These findings may be explained in the light of the origin of these bases of power. Harsh bases of power are available to the superior due to the position held in the organisation and usage of these may be perceived by subordinates as somewhat arbitrary, low task relevant and utilising one's positional advantage. While soft bases of power gain compliance through personal rather than organisational resources and are perceived by subordinates as more task relevant.

Power interaction model also states that, an influencing agent who is liked tends to arouse greater compliance even if he/she resorts to unfavourable strategies and a disliked agent arouses resistance as the target attempts to thwart any gains for the former. Raven (1992, 1993) argues that, in the latter case, the efficacy of power using soft strategies is diminished, thereby, forcing the disliked influencing agent to resort to harsh strategies. The findings of the present study partially support the above argument. Subordinates reported more compliance with harsh bases of power in high quality of interaction than in low. But they also reported more compliance with soft bases of power than with harsh bases of power in low quality of interaction.

The reason for this may be that quality of interaction is a mutual linking and involves both 'on' and 'off the job' dimensions, while liking may not be necessarily linked with task-related interaction. From a practical perspective, these findings indicate that in high quality of interaction, leaders, even when using harsh bases of power, can expect compliance in situations demanding immediate action and out-group members may respond more positively to leaders' soft approach than a harsh one based on their position. Subordinates' compliance is not only related to bases of power of leader but also influenced by the way they relate to each other. These relationships have to be managed to provide superior performance and continued productivity.

The final part of the analysis focuses on the relationship between reported compliance to bases of power and attitude towards superior, and on the moderating effect of quality of interaction on these relationships. The distinction between soft and harsh power bases was used to understand these relationships. According to Avolio and Bass (1988), compliance that is attributed to fear of punishment, promise of reward, or the desire to fulfil contractual obligations (harsh bases) emphasises subordination rather than voluntary acceptance. The findings of the present study are in congruence with this contention. Present findings reveal that when subordinates attributed compliance with soft bases of power they reported increase in their satisfaction and commitment with superior and global influence. Compliance with harsh bases of power was positively related to commitment and global influence but unrelated to satisfaction. These findings are in congruence with the results of a study examining the relationship between compliance and commitment and satisfaction with organisation (Koslowsky et al, 2001).

At the individual level, expert power and impersonal reward positively influence satisfaction with superior and coercive basis was found to have negative effect on satisfaction. These findings are in some way congruent with the results of earlier studies (Carson, Carson, and Roe, 1993; Schriesheim, Hinkin and Podsakoff, 1991), which demonstrated that expert and referent bases of power are strongly related with satisfaction and performance, whereas coercive basis of power shows a negative or neutral relationship.

Podsakoff and Schriesheim (1985) mentioned that expert power is the only source related to commitment and it was observed in other studies as well (Schriesheim et al, 1991, Koslowsky et al, 2001). The present study also indicates the contribution of expert power for commitment. Along with expert power, referent power and personal reward significantly contributed to commitment with superior. From harsh power category, impersonal and personal coercive, legitimate equity, and impersonal reward had significant impact on the commitment with superior.

Compliance with soft bases of power, especially referent and expert power, also positively contributed to overall influence of the superior over subordinates. From harsh power category, impersonal reward positively contributed to global influence but impersonal coercive reduced overall influence (as negative beta weight indicates). These findings indicate that soft bases of power are related to more long-term impact than harsh bases of power.

Previous research has indicated the superiority of high quality of interaction in several aspects such as organisational commitment, satisfaction with supervision, and satisfaction with work (Duchon, Green, and Taber, 1986; Schriesheim and Gardiner, 1992; Vecchio and Gobdel, 1984). Our findings provide further insight into the advantage of quality of interaction in influence process. Subordinates who have high-quality exchanges with the leader enjoy relationships based on mutual contribution, loyalty, trust, and liking. Quality of interaction significantly moderated the relationship between compliance with soft bases of power and subordinates' satisfaction and commitment with superior but not compliance with harsh bases of power and subordinates' attitude towards superior.

From the soft category, the quality of interaction significantly moderated the compliance with information and personal reward and satisfaction with superior. For commitment with superior, significant interac-tions were found for referent power, personal reward, and dependence legitimacy with quality of interaction. These findings add further insight that harsh bases of power may be effective for immediate action, but continuous use of power bases from this category with high in-group members would not be desirable as it may lead to resistance in relationships.

This study has certain limitations. It considered only subordinates' perception about bases of power and quality of interaction. We could have more faith in results and possibility of common method variance could have been reduced by involving superiors also. In addition, we hypothesised causal relationships, but it was a correlational study.

In conclusion, our findings have implications for both researchers and practitioners. Present findings support the assumption that soft power sources are more effective than harsh sources inherited by the superior due to his position in the organisation. Furthermore, the quality of interaction between superior and subordinates plays a pivotal role which emphasises developing high quality of interaction between both, leading to positive outcome for both parties. These findings have implications for training programmes aimed at creating positive work culture in organisations. Future research needs to examine the link between subordinates' compliance with superior and organisation's performance along with moderating effects of cultural factors.

References

Aiken L and S West, 1991. Multiple Regression: Testing and Interpreting Interactions. Newbury Park, CA: Sage.

Avolio BJ and BM Bass, 1988. "Transformational Leadership, Charisma, and Beyond" in JG Hunt, BR Baliga, HP Dachler and CH Schriesheim (eds), Emerging Leadership Vistas, pp 29-50, Lexington, MA: Lexington Books.

Bass BM, 1994. Improving Organisation Effectiveness through Transformational Leadership, Thousand Oaks, CA: Sage.

Bennis WG and B Nanus, 1985. Leaders: The Strategies for Taking Charge. New York: Harper and Row.

Bernard CI, 1938. The Functions of the Executive. Cambridge: Harvard University Press.

Brass DJ and ME Burkhardt, 1993. "Potential Power and Power Use: An Investigation of Structure and Behavior", Academy of Management Journal, Vol 36, pp 441-470.

Bui KV, BH Raven and J Schwarzwald, 1994. "Interpersonal Satisfaction and Influence Tactics in Close Heterosexual Relationships", Journal of Social Behavior and Personality, Vol 9, pp 429-442.

Carson PP, KD Carson and Roe CW, 1993. "Social Power Bases: A Metaanalytic Examination of the Relationships and Outcomes', Journal of Applied Social Psychology, Vol 23, pp 1150-1169.

Dansereau F Jr, G Graen and WJ Haga, 1975. "A Vertical Dyad Linkage Approach to Leadership with Formal Organisations--A Longitudinal Investigation of Role Making Process', Organisational Behavior and Human Performance, Vol 13, pp 46-78.

Dienesch RM and RC Liden, 1986. "Leader-member Exchange Model of Leadership: A Critique and Further Development", Academy of Management Review, Vol 11, pp 618-634.

Duchon D, Green and TD Taber, 1986. "Vertical Dyad Linkage: A Longitudinal Assessment of Antecedents, Measures, and Consequence", Journal of Applied Psychology, Vol 71, pp 56-60.

Dunne EJ Jr, MJ Stahl and LJ Jr Melhart, 1978. "Influence Sources of Project and Functional Managers in Matrix Organisations", Academy of Management Journal, Vol 21, pp 135-140.

Eden D, 1984. "Self-fulfilling Prophecy as a Management Tool: Harnessing Pygmalion", Academy of Management Review, Vol 9, pp 64-73.

--, 1990. Pygmalion in Management: Productivity as a Self-fulfilling Prophecy, Lexington, MA: Lexington Books.

Erchul WP, BH Raven and AG Ray, 2001. "School Psychologists' Perception of Social Power Bases in Teacher Consultation", Journal of Education and Psychological Consultation, Vol 12 No 1, pp 1-23.

French JRP Jr and BH Raven, 1959. "The Bases of Social Power', in D Cartwright (ed), Studies in Social Power, pp 150-167, Ann Arbor, MI: Institute for Social Research.

Graen GB and J Cashman, 1975. "A Role-making Model of Leadership in Formal Organisations: A Development Approach" in JG Hunt and LL Larson (eds), Leadership Frontiers, pp 143-165, Kent, OH: Kent State University Press.

Graen GB, RC Liden and W Hoel, 1982. "The Role of Leadership in the Employee Withdrawal Process", Journal of Applied Psychology, Vol 67, pp 868-872.

Graen GB, MA Novak and P Sommerkamp, 1982. "The Effects of Leader Member Exchange and Job Design on Productivity and Satisfaction: Testing a Dual Attachment Model', Organisational Behavior and Human Decision Process, Vol 30, pp 109-131.

Graen GB and Scandura TA, 1987. "Toward a Psychology of Dyadic Organising", Research in Organisational Behavior, Vol 9, pp 175-208.

Graen GB and Schiemann W, 1978. "Leader-member Agreement: A Vertical Dyad Linkage Approach", Journal of Applied Psychology, Vol 63, pp 206-212.

Graen GB and Uhl-Bien M, 1995. "Relationship Based Approach to Leadership: Development of Leader-member Exchange (LMX) Theory of Leadership Over 25 Years: Applying A Multi-level Multi-domain Perspective', Leadership Quarterly, Vol 6, pp 219-247.

Gupta, Bindu and NK Sharma, 1999. "Effectiveness of Bases of Power: An Experimental Study', Psychological Studies, Vol 44, pp 35-42.

Gupta B and NK Sharma, 2003. Bases of Power as a Function of Organisation Type and Managerial Level. Behavioural Scientist, Vol 4 No 2, pp 73-78.

Hinkin TR and Schriesheim CA, 1989. "Development and Application of New Scales to Measure the French and Raven, 1959 Bases of Power', Journal of Applied Psychology, Vol 74, pp 561-587.

Hobbes T, 1968. Leviathan. Middlesex: Penguin. (Original work published 1651).

Imai Y, 1989. "The Relationship between Perceived Social Power and the Perception of being Influenced', Japanese Psychological Record, Vol 31, pp 97-107.

Hunt JG, 1984. "Leadership and Managerial Behavior" in JE Rosenzweig and FE Caste (eds), Modules in Management. Chicago: Science Research Associates.

Hunt SD and JR Nevin, 1974. "Power in a Channel of Distribution: Sources and Consequences", Journal of Marketing Research, Vol 11, pp 186-193.

Ivancevich J and J Donnelly, 1970. "Leader Influence and Performance", Personnel Psychology, Vol 23, pp 539-549.

Jaccard J, CK Wan and R Turisi, 1990. "The Detection and Interpretation of Interaction Effects between Continuous Variables in Multiple Regression", Multivariate Behavioral Research, Vol 25, pp 467-478.

Kanter RM, 1977. Men and Women of the Corporation. New York: Basic Books,

--, 1979. "Power Failure in Management Circuits", Harvard Business Review, Vo1 57, pp 65-75.

Kelman HC, 1958. "Compliance, Identification and Internalisation: Three Processes of Attitude Change", Journal of Conflict Resolution, Vol 2, pp 51-60.

Koslowsky M and J Schwarzwald, 1993. "The Use of Power Tactics to Gain Compliance: Testing Aspects of Raven's (1988) Theory in Conflictual Situations", Social Behaviour and Personality, Vol 21, pp 135-144.

Koslowsky M and J Schwarzwald, 2001. "The Power Interaction Model: Theory, Methodology and Empirical Applications", in AY Lee-Chai and JA Bargh (eds), Use and Abuse of Power, pp 195-214, Philadelphia, PA: Psychology Press.

Koslowsky M, J Schwarzwald and S Ashuri, 2001. "On the Relationship between Subordinates' Compliance to Power Sources and Organisational Attitudes", Applied Psychology: An International Review, Vol 50 No 3, pp 455-476.

Lewin K, 1941. Analysis of the Concepts Whole, Differentiation, and Unity, University of Iowa Studies in Child Welfare, Vol 18, pp 226-261.

Liden RC and GB Graen, 1980. "Generalisability of the Vertical Dyad Linkage Model of leadership', Academy of Management Journal, Vol 23, pp 451-465.

Liden RC and JM Maslyn, 1993. "LMX-MDM: Scale Development for Multidimensional Measure of Leader-member Exchange', Paper Presented at the National Meetings of the Academy of Management, Atlanta, Georgia (Revised version).

Liden RC, RT Sparrowe and SJ Wayne, 1997. "Leader-member Exchange Theory: The Past and Potential for the Future', Research in Personnel and Human Resource Management, Vol 15, pp 47-119.

Lukes S, 1974. Power." A Radical View. London: MacMillan.

Machiavelli N, 1984. The Prince. Oxford, England: Oxford University Press. (Original work published 1532).

McClelland DC and BE Boyatzis, 1982. "Leadership Motive Pattern and Long-term Success in Management", Journal of Applied Psychology, Vol 67, pp 737-743. McClelland DC and BE Boyatzis, 1982, "Leadership Motive Pattern and Long-term Success in Management", Journal of Applied Psychology, Vol 67, pp 737-743.

Munter M, 1993. "Cross-cultural Communication for Managers", Business Horizons, May-June, pp 69-78.

Nesler MS, H Aguinis, BM Quigley and Tedeschi, JT, 1993. "The Effect of Credibility on Perceived Power", Journal of Applied Social Psychology, Vol 23, pp 1407-1425.

Nietzsche F, 1968. The Will to Power, New York: Vintage Book. (Original work published between 1883 and 1888).

Pfeffer J, 1981. Power in Organisations. Boston: Pitman.

--, 1992. Managing with Power." Politics and Influence in Organisations. Boston, MA: Harvard Business School Press.

O'Reilly CA and J Chatman, 1986. "Organisational Commitment and Psychological Attachment", Journal of Applied Psychology, Vol 71, pp 492-499.

Podsakoff PM and CA Schriesheim, 1985. "Field Studies of French and Raven's Bases of Social Power: Critique, Reanalysis, and Suggestions for Future Research', Psychological Bulletin, Vol 97, pp 387-411.

Rahim MA, 1988. "The Development of a Leader Power Inventory", Multivariate Behavioral Research, Vol 23, pp 491-502.

--, 1989. "Relationship of Leader Power to Compliance and Satisfaction with Supervision: Evidence from A National Sample of Managers", Journal of Management, Vol 15, pp 545-556.

Rahim MA and GF Buntzman, 1989. "Supervisory Power Bases, Style of Handling Conflict with Subordinates, and Subordinate Compliance and Satisfaction", Journal of Psychology, Vol 123, pp 195-210.

Rahim MA, NH Kim and JS Kim, 1994. "Bases of Leader Power, Subordinate Compliance, and Satisfaction with Supervision: A Cross Cultural Study of Managers in the U.S and S Korea", The International Journal of Organisational Analysis, Vol 2 No 2, pp 136-154.

Raven BH, 1965. "Social Influence and Power" in ID Stenier and M Fishbein (eds), Current Studies in Social Psychology. New York: Holt, Rinehart, Winston, pp 371-381.

--, 1992. "A Power/interaction Model of Interpersonal Influence: French and Raven Thirty Years Later", Journal of Social Behavior and Personality, Vol 7, pp 217-244.

--, 1993. "The Bases of Power: Origins and Recent Developments", Journal of Social Issues, Vol 49 No 4, pp 227-251.

Raven BH, J Schwarzwald and M Koslowsky, 1998. "Conceptualising and Measuring a Power/Interaction Model of Interpersonal Influence", Journal of Applied Psychology, Vo1 28 No 4, pp 307-332.

Russell B, 1938. Power: A New Social Analysis. London: Allen and Unwin.

Scholl W, 1995. "Effective Teamwork--A Theoretical Model and a Test in the Field", in J Davis and E Witte (eds), Understanding Group Behavior. Hillsdale, N J: Lawrence Erlbaum Associates Inc.

Schriesheim CA and CC Gardiner, 1992. "An Exploration of the Discriminant Validity of the Leader-member Exchange Scale (LMXT) Commonly Used in Organisational Research", in M Schnake (ed), Proceedings of the Southern Management Association, (pp 91-93). Valdosta, GA: Southern Management Association

Schriesheim CA, TR Hinkin and PM Podsakoff, 1991. "Can Ipsative and Single Item Measures Produce Erroneous Results in Field Studies of French and Raven's (1959) Five Bases Power? An Empirical Investigation", Journal of Applied Psychology, Vo1 76, pp 106-114.

Schwarzwald J, M Koslowsky, 1999. "Gender, Self-esteem, and Focus of Interest in the Use of Power Strategies by Adolescents in Conflict Situations", Journal of Social Issues, Vol 55, pp 15-32.

Schwarzwald J, M Koslowsky and V Agassi, 2001. "Captain's Leadership Type and Police Officers' Compliance to Power Bases", European Journal of Work and Organisational Psychology, Vol 10 No 3, pp 273-290.

Sharma NK and Gupta Bindu, 2000. "Reasons for Compliance: Leaders' Bases of Power and Subordinates' Perceptions', Psychological Studies, Vol 45, No 1 and 2.

Sparrowe RT and RC Liden, 1997. "Process and Structure in Leader-member Exchange", Academy of Management Review, Vol 22 No 2, pp 522-552.

Thamhain HJ and GR Gemmill, 1974. "Influence Styles of Project Performance Correlates", Academy of Management Journal, Vol 17, pp 216-224.

Thibodeaux III HF and Lowe RH, 1996, "Convergence of Leader-member Exchange and Mentoring: An Investigation of Social Influence Patterns", Journal of Social Behavior and Personality, Vol 11 No 1, pp 97-114.

Tondon K, 1990. Quality of Interaction in Leader-member Dyads: Measurement, Antecedents, and Consequences, Unpublished doctoral dissertation, Kanpur, India: Indian Institute of Technology.

Vecchio RP and BC Gobdel, 1984. "The Vertical Dyad Linkage Model of Leadership: Problems and Prospects', Organisational Behavior and Human Performance, Vol 34, pp 5-20.

Vecchio RP and M Sussmann, 1989. "Preferences for Forms of Supervisory Social Influence", Journal of Organisational Behavior, Vol 10, pp 135-143.

Wakabayashi M, GB Graen, MR Graen and MG Graen, 1988. "Japanese Management Progress: Mobility into Middle Management", Journal of Applied Psychology, Vo173, pp 217-227.

Weber M, 1948. From Max Weber: Essays in Sociology. London: Routledge.

Yukl GA, 1994. Leadership in Organisations (3rd ed). Englewood Cliffs, NJ: Prentice Hall.

Yukl G and CM Falbe, 1991. "Importance of Different Power Sources in Downward and Lateral Relations', Journal of Applied Psychology, Vol 76, pp 416-423.

Bindu Gupta

Institute of Management Technology, Ghaziabad, India

Narendra K Sharma

Indian Institute of Technology Kanpur, India Table 1: Factor Analysis Loadings for 11 Power Sources Loadings Power sources by factor Factor 1 Factor 2 Soft Bases of power Information Power 0.77 0.41 Referent Power 0.68 0.41 Position legitimacy 0.68 0.34 Personal Reward 0.68 0.50 Dependence legitimacy 0.68 0.15 Expert Power 0.73 0.20 Harsh Bases of power Impersonal Coercive -0.21 0.81 Reciprocity legitimacy 0.43 0.68 Equity legitimacy 0.34 0.67 Personal Coercive 0.20 0.77 Impersonal Reward 0.37 0.69 Explained Variance (%) 31.34 29.22 Internal Consistency 0.84 0.82 Table 2: Means and Standard Deviation for Bases of Power Std Mean Deviation Information power 5.11 1.09 Dependence legitimacy 5.04 1.03 Expert power 5.00 1.29 Position legitimacy 4.97 1.23 Personal reward 4.73 1.25 Referent power 4.54 1.16 Reciprocity legitimacy 4.12 1.36 Impersonal reward 4.04 1.53 Equity legitimacy 3.72 1.32 Personal coercive 3.67 1.28 Impersonal coercive 3.26 1.41 Soft basis of power 4.89 0.88 Harsh basis of power 3.76 1.05 Table 3: Means and Standard Deviations of Quality of Interaction (High and Low) and Bases of Power Measures Std Bases QI Mean Deviation F of power SOFT Low 4.65 0.86 41.07 ** High 5.17 0.82 HARSH Low 3.74 0.98 0.27 (ns) High 3.79 1.12 Note: ** = p <.001; ns = not significant Table 4: Correlation between Bases of Power Factors and Study Variables Satisfaction Commitment Quality of with with Global Interaction Superior Superior Influence Soft Bases of Power 0.37 ** 0.34 ** 0.47 ** 0.36 ** Harsh Bases of Power 0.009 -0.07 .016 ** .018 ** Note: ** = p <.001 Table 5: Hierarchical Regression Analyses for Testing Main Effect of Soft Bases of Power and Moderating Effects of Quality of Interaction for Satisfaction, Commitment, and Global Influence Commitment Independent Satisfaction with variables with Superior Superior Model Model 1 2 1 2 Soft Bases of Power 0.34 ** 0.31 ** 0.45 ** 0.43 ** Soft Bases of Power and Quality of Interaction -0.19 ** -0.19 ** [R.sup.2] 0.11 0.15 0.21 0.25 F 53.69 ** 37.65 ** 113.71 ** 69.75 ** Information Power (IP) 0.09 0.09 -0.04 -0.05 Referent Power (RP) 0.11 0.11 0.17 ** 0.16 ** Position Legitimacy (PL) -0.11 -0.10 0.04 0.05 Personal Reward (PR) 0.11 0.08 0.25 ** 0.21 ** Dependence Legitimacy (DL) 0.03 0.04 -0.06 -0.04 Expert Power (EP) 0.23 ** 0.21 ** 0.24 ** 0.23 ** IP * QI -0.11 * -0.07 RP * QI -0.08 -0.14 * PL * QI 0.00 -0.03 PR * QI -0.14 * -0.13 * DL * QI 0.08 0.14 ** EP * QI 0.01 0.02 [R.sup.2] 0.15 0.21 0.27 0.33 F 12.11 ** 8.78 ** 25.30 ** 17.13 ** Independent Global variables Influence Model 1 2 Soft Bases of Power 0.36 ** 0.35 ** Soft Bases of Power and Quality of Interaction -0.11 * [R.sup.2] 0.13 0.14 F 64.81 ** 35.77 ** Information Power (IP) 0.10 0.11 * Referent Power (RP) 0.16 ** 0.16 ** Position Legitimacy (PL) -0.03 -0.01 Personal Reward (PR) 0.10 0.07 Dependence Legitimacy (DL) -0.05 -0.03 Expert Power (EP) 0.21 ** 0.19 ** IP * QI -0.12 * RP * QI -0.14 * PL * QI 0.10 PR * QI -0.01 DL * QI 0.05 EP * QI 0.01 [R.sup.2] 0.16 0.19 F 13.61 ** 8.33 ** Note: Entries are beta weights; ** = p <0.01; * = p <0.05 QI = Quality of Interaction Table 6: Hierarchical Regression Analysis for Testing Main Effect of Hash Bases of Power and Moderating Effects of Quality of Interaction for Satisfaction, Commitment, and Global Influence Commitment Independent Satisfaction with variables with Superior Superior Model Model 1 2 1 2 Harsh bases of -0.07 -0.07 0.16 ** 0.16 ** Power Harsh Bases of Power and Quality of Interaction 0.05 -0.05 [R.sup.2] 0.01 0.01 0.03 0.03 F 1.93 (ns) 1.49 (ns) 11.23 ** 6.19 ** Impersonal Coercive (IC) -0.43 ** -0.39 0.32 ** -0.28 ** Reciprocity Legitimacy (RL) 0.07 0.09 0.14 * 0.15 * Equity Legitimacy (EL) 0.01 0.03 0.11 * 0.08 Personal Coercive (PC) 0.01 0.03 -0.04 -0.03 Impersonal Reward (IR) 0.21 ** 0.19 0.28 ** 0.26 ** IC * QI 0.09 0.09 RL * QI 0.01 -0.03 EL * QI -0.09 -0.10 PC * QI 0.06 0.08 IR * QI -0.18 ** -0.12 [R.sup.2] 0.16 0.22 0.18 0.23 F 15.94 ** 11.41 ** 18.98 ** 12.67 ** Independent Global variables Influence Model 1 2 Harsh bases of Power 0.19 ** 0.19 ** Harsh Bases of Power and Quality of Interaction -0.04 [R.sup.2] 0.04 0.04 F 15.51 ** 8.04 ** Impersonal Coercive (IC) -0.12 * 0.12 * Reciprocity Legitimacy (RL) 0.10 0.06 Equity Legitimacy (EL) -0.04 -0.02 Personal Coercive (PC) 0.14 * 0.15 * Impersonal Reward (IR) 0.16 ** 0.17** IC * QI 0.03 RL * QI 0.09 EL * QI -0.01 PC * QI -0.15 IR * QI -0.05 [R.sup.2] 0.09 0.10 F 7.86 ** 4.85 ** Note: Entries are beta weights; ** = p <0.01; * = p <0.05 QI = Quality of Interaction