Dispositional Effects in Intra-Organizational Influence Tactics: A Meta-Analytic Review

The study of dispositional effects in organizational behavior has been a subject of continued dialogue in the field (Davis-Blake & Pfeffer, 1989; House, Shane & Herold, 1996; Mischel, 1968). In light of the dialogue concerning the utility of studying dispositional effects in organizational behavior research, empirical summaries of the literature are necessary to advance this discourse beyond "armchair philosophy." The need for dispositional research is salient to the on-going dialogue of dispositional effects in the organizational behavior literature.

Over the past twenty years, a number of studies have examined the influence process using categories of proactive behaviors termed "influence tactics" (Kipnis, Schmidt, & Wilkinson, 1980; Schriesheim & Hinkin, 1990; Yukl & Falbe, 1990). A search on PSYCHINFO revealed 341 articles that have examined "influence tactics" or "influence strategies" since 1980. As researchers continue to examine influence tactics, it is important that summaries of findings be calculated to guide future efforts.

Research on influence tactics has examined both situational and dispositional variables. Situational antecedents have included agent's perceived power (Hinkin & Schriesheim, 1990), political behaviors (Harrell-Cook et al., 1999), task objectives (Yukl & Falbe, 1990; Yukl, Guinan, & Sottolano, 1995), direction of influence (Falbe & Yukl, 1992; Yukl & Tracey, 1992), group processes (Guerin, 1995), organizational context (Ansari & Kapoor, 1987), role clarity (Orpen, 2000), and exchange quality (Wayne & Ferris, 1990). Dispositional variables have included work values (Buckle, 2000), impression management (Deluga, 1991; Wayne & Kacmur, 1991), sex (Dubrin, 1991; Thacker, 1995), and motivation (Barbuto, Fritz, & Marx, 2002).

Higgins, Judge, and Ferris (2003) conducted a meta-analysis of outcomes associated with influence tactics and found that the specific strategies used in the influence process have positive effects on work outcomes. Much more is collectively known about the impacts of influence tactics on work-units and organizations than is known about the roots of such strategies.

The antecedents of influence tactics have been a major focus of inquiry in studies using dispositional variables such as impression management, Machiavellianism, self-monitoring, locus of control, social identity, intrinsic motivation, extrinsic motivation, role clarity, gender, age, context, direction of influence, and task objectives. Because such a large number of studies exist in the literature, an examination of the cumulative effects of some of these relationships is necessary to generalize findings and guide future research efforts.

This study tested dispositional effects of influence tactics used, employing literature search strategies consistent with those described by Podsakoff, MacKenzie, and Bommer (1996) and sample-weighted meta-analytic procedures consistent with those of Fuller and Hester (1999), using the Hunter and Schmidt (2004) validity generalization procedures.

Influence Tactics and Their Covariates

Kipnis et al. (1980) extended the work of French and Raven's (1959) power tactics by inductively examining the influence strategies used by employees to influence their supervisors in organizational settings. The resulting work used factor analysis to identify six reliably measured influence tactics. The influence tactics examined in this study included the six tactics originally proposed by Kipnis et al.: ingratiation, rationality, exchanges, coalitions, upward appeals, and assertiveness. These six tactics were replicated by Schriesheim and Hinkin (1990) and Yukl and Falbe (1990) and are evident in most iterations of influence tactic research in the organizational behavior field.

Ingratiation

Ingratiation is the practice of getting others in a good mood before trying to influence them, often taking the form of flattery. In the twenty plus years since ingratiation was first articulated in the Kipnis et al. (1980) framework, its antecedents have been examined extensively. Relationships between person-centered variables and ingratiation influence tactics have been consistently found in the literature. Studies have found positive correlations with impression management (Deluga, 1991; Hochwarter et al., 2000), Machiavellianism (Andersson & Bateman, 2000), self-monitoring (Kumar & Beyerlain, 1991), external locus of control (Ringer & Boss, 2000), social identification (Harrison, Hochwarter, Perrewe, & Ralston, 1998), intrinsic motivation (Barbuto & Scholl, 1999), and extrinsic motivation (Harrison et al., 1998).

Rationality

Rationality involves making logical persuasive arguments to influence the decisions or behaviors of others. Kipnis et al. (1980) labeled influence attempts of this nature reason, while Yukl and Falbe (1990) called this tactic rational persuasion. In this study, we used Schriesheim and Hinkin's (1990) term, rationality. Taken together, person-centered variables appear related to the use of rationality influence tactics. Research of the antecedents of rationality has revealed positive correlations with impression management (Deluga, 1991), Machiavellianism (Andersson & Bateman, 2000), self-monitoring (Hochwarter et al., 2000), and intrinsic motivation (Buckle, 2000), and negative correlations with locus of control (Ringer & Boss, 2000).

Exchanges

Exchanges are characterized by the offer of tangible rewards or favors for compliance from others, and may include reminding others of past favors when attempting to influence them. Exchanges were originally conceptualized by Kipnis et al. (1980) and similarly defined later using the same term (Schriesheim & Hinkin, 1990; Yukl & Falbe, 1990). The use of exchange tactics appears to share relations with person-centered variables. Research on antecedents of exchange tactics has revealed positive correlations with locus of control, Machiavellianism, self-monitoring (Hochwarter et al., 2000), and intrinsic motivation (Barbuto, Fritz & Marx, 2001).

Coalitions

Coalitions involve gathering support from several people, creating a peer pressure or a "ganging up" effect to influence others' actions or decisions. Coalitions were originally conceptualized by Kipnis et al. (1980) and similarly defined by later researchers using the same term (Schriesheim & Hinkin, 1990; Yukl & Falbe, 1990). The use of coalition tactics may be preceded by person-centered variables. Research on antecedents of coalitions has suggested positive correlations with impression management and extrinsic motivation (Hochwarter et al., 2000), Machiavellianism (Andersson & Bateman, 2000), self-monitoring (Caldwell & Burger, 1997), intrinsic motivation (Barbuto & Scholl, 1999), and locus of control (Ringer & Boss, 2000).

Upward Appeal

Upward appeal can be described as gaining support from someone's boss in an effort to influence that person, using the hierarchical structure to work the system when attempting to influence. Upward appeal has been studied extensively since Kipnis et al. (1980) originally defined it. Person-centered variables appear to be appropriate antecedents of upward appeal tactics. Research in the antecedents of upward appeal has revealed positive relationships with locus of control and extrinsic motivation (Hochwarter et al., 2000), Machiavellianism (Vecchio & Sussmann, 1989), and self-monitoring (Caldwell & Burger, 1997), and negative relationships with social identity (Hochwarter et al., 2000).

Assertiveness

Assertiveness is the practice of making repeated requests in order to influence another to do something. This tactic features proactive actions designed to make intentions and desires clear, with the aim of getting others to succumb to influence attempts. Kipnis et al. (1980) originally labeled this influence behavior sanctions, Yukl and Falbe (1990) labeled it pressure tactics, while Schriesheim and Hinkin (1990) called it assertiveness. For the purposes of this study, behavior that utilizes threats, demands or intimidation is considered assertiveness. The use of assertiveness influence tactics appears to be preceded by personcentered variables. Research on antecedents of assertiveness revealed positive correlations with Machiavellianism (Vecchio & Sussmann, 1989), extrinsic motivation, impression management, self-monitoring (Hochwarter et al., 2000), and locus of control (Ringer & Boss, 2000); and negative relationships with social identity (Hochwarter et al., 2000) and intrinsic motivation (Barbuto & Scholl, 1999).

Influence Tactic Research Similar to That of Kipnis et al.

Schriesheim and Hinkin (1990) replicated the influence typology of Kipnis et al. (1980) and reported it to be inadequate because of weak factor loading on several dimensions of influence. Schriesheim and Hinkin revised the subscales to produce an 18-item measure (six dimensions with three items each). Yukl and Falbe (1990) contributed to this line of inquiry by adding several influence tactics not operationalized in the other frameworks (consultation, pressure, personal appeals, and inspirational appeals). Yukl and Falbe's work also was unique because it measured both the influence strategies of agents (those attempting to influence others) and the influence behaviors of targets (those who are being influenced). They found that direction of influence had a strong effect on the choice of influence tactic used.

Dispositional Variables and Hypotheses

House, Shane, and Herold (1996) provide a useful distinction for clearly defining dispositional variables. They argue that dispositions are psychological (as opposed to physical or other objectively assessed characteristics of individuals). Therefore, dispositions may include such psychological constructs as personality characteristics, need states, attitudes, preferences, and motives. This meta-analysis examines dispositional antecedents of the influence tactics using impression management, Machiavellianism, self-monitoring, locus of control, social identity, intrinsic motivation, and extrinsic motivation. Other dispositional variables were considered but inadequate sample sizes prevented their inclusion.

Impression Management

Goffman (1955) provided the first notion of impression management in the behavioral sciences with the notion of dramaturgy; described as the process by which individuals acted out the roles they were trying to convey. Impression management has since been described as the ability to portray a desired image or attitude to convey desired identities (Higgins, et al., 2003; Schlenker, 1980).

Individuals with a high degree of impression management often take great care to come across in a positive light (Deluga, 1991; Stevens & Kristof, 1995). Several scholars have examined the role played by impression management in leadership settings (e.g., Sosik, Avolio & Jung, 2002). Others have examined impression management and influence tactics in several studies (Andersson & Bateman, 2000; Deluga & Perry, 1994; Hochwater et al, 2000). In this meta-analysis, impression management is tested as a covariate with each of the six influence tactics. Given the results of past studies examining these variables, we expect positive relationships between self-reported use of impression management and their use of ingratiation, coalitions, upward appeals, and assertiveness.

Hypothesis 1: Impression management reported by agents will be positively related to their use of ingratiation, coalitions, upward appeals, and assertiveness influence tactics.

Machiavellianism

Machiavellianism is the willingness to do whatever it takes to get what one wants. Grounded in the principles of power and economic theory, Machiavelli (1513/1988) described the political savvy leaders must demonstrate to maintain their standing and be effective. Individuals high in the Machiavellianism disposition are seen as politically savvy, highly influential, and, at times, morally questionable (Christie & Geis, 1970). Individuals low in Machiavellianism are seen as nave or unrealistic, overly trusting, and easily taken advantage of. Several researchers have examined the role Machiavellianism plays in organizational behavior research, but fewer looked at influence tactics (see Harrell-Cook et al, 1999; Harrison et al., 1998; Hochwater et al., 2000; Kumar & Keverlain, 1991; Vecchio & Sussmann, 1989). In this meta-analysis, Machiavellianism is examined as a self-reported disposition with each of the six influence tactics. Based on the results of these studies, we expect Machiavellianism to share a positive relationship with exchange tactics, ingratiation, assertiveness, upward appeals, and coalitions.

Hypothesis 2: Machiavellianism of agents will be positively related to their use of exchange, ingratiation, assertiveness, upward appeals, and coalition influence tactics.

Self-Monitoring

Self-monitoring is the extent to which people are willing and able to control their mannerisms and attitudes. Snyder (1974) operationalized self-monitoring behavior as a means for understanding why some are able to adapt well to different settings, audiences, and expectations, while others seem less adaptable to these differences. Individuals low in self-monitoring tend to be more consistent across situations and less adaptable to change.

The role of self-monitoring in organizational behavior has been studied extensively over the past thirty years (see Day, Unckless, Schleicher, & Killer, 2002). Several recent studies have examined self-monitoring as an antecedent to leadership (see Becker, Ayman, & Korabik, 2002; Dobbins, Long, Dedrick, & demons, 1990; Turnley & Bolino, 2001). Many researchers have examined self-monitoring and specific influence tactics (Caldwell & Burger, 1997; Harrell-Cook, Ferris, & Dulebohn, 1999; Harrison et al, 1998; Hochwater et al., 2000; Kumar & Beyerlain, 1991). In this study, we test the relationship between self-reported self-monitoring and each of the six influence tactics. From results of these reported studies, we anticipate similar positive relationships between self-monitoring and ingratiation, coalitions, exchanges, and assertiveness.

Hypothesis 3: Self-monitoring of agents will be positively related to their use of ingratiation, coalitions, exchanges, and assertiveness influence tactics.

Locus of Control

Locus of control can be described as the extent to which individuals believe they can control the events around them (Rotter, 1966). Persons may be classified as having internal loci of control (meaning they believe they control their own destiny), or external loci of control (meaning they believe others or "chance" determine their fate) (Spector, 1982). Rotter was among the first to operationalize the construct of locus of control with a developed measure, which is still used today in organizational research. Locus of control has been used to examine such organizational behavior phenomena as job satisfaction and performance (Judge & Bono, 2001) and reactions to promotions and being passed over (Lam & Schaubroeck, 2000). Several researchers have examined locus of control as an antecedent to agents' influence tactics (Harrison et al., 1998; Hochwater et al., 2000; Ringer & Boss, 2000). In this meta-analysis, locus of control is examined as a covariate with each of the six influence tactics. Consistent with past results, we anticipate that external locus of control (high scores in locus of control) will be negatively related to rationality and positively related to exchanges, upward appeals, and ingratiation influence tactics.

Hypothesis 4a: External locus of control will be negatively related to rationality influence tactics.

Hypothesis 4b: External locus of control will be positively related to exchanges, upward appeals, and ingratiation influence tactics.

Social /Referent Identification

Social identity has its roots in social exchange theory (Blau, 1974; Hollander, 1979) and referent social power (French & Raven, 1959). Social identity is characterized by a desire for acceptance and affiliation with a group or organization, as well as a strong need for connectedness and group cohesion. Ashforth and Mael (1989) advanced the field's understanding of social identity theory by explicating variables of interest and delineating the role of socialization in the workplace.

Several studies have examined theories conceptually similar to social identity (Barbuto et al., 2002; Farmer & Maslyn, 1999; Gavin et al., 1995). In this meta-analysis, social identity is examined as a covariate with each of the six influence tactics. Consistent with the results reported in these studies, we expect social identity to share a negative relationship with assertiveness, upward appeals, and coalition influence tactics. We also expect a positive relationship between social identity and the use of ingratiation and exchange influence tactics.

Hypothesis 5a: Social identity of agents will be negatively related to their use of assertiveness, upward appeals, and coalition influence tactics.

Hypothesis 5b: Social identity of agents will be positively related to their use of ingratiation and exchange influence tactics.

Intrinsic Motivation

Intrinsic motivation comes from an inner drive to succeed; the energy source for performance comes from within (Deci, 1975; Katz & Kahn, 1978). Intrinsic motivation has been divided into two general categories - process and outcomes. Process-based intrinsic motivation involves derivation of pleasure and rewards during the engagement of activities. Outcome-based intrinsic motivation involves a desire to pursue outcomes and reach challenging goals.

Several studies have examined internal motives as predictors of influence tactics used (Barbuto et al., 2002; Barbuto & Scholl, 1999; Blickle, 2000). The general category of "intrinsic" motivation describes both types of internal motivations. In this meta-analysis, intrinsic motivation is examined as a covariate with each of the six influence tactics. Consistent with past results, we anticipated positive relationships between intrinsic motivation and both ingratiation and rationality influence tactics.

Hypothesis 6: Intrinsic motivation of agents will be positively related to ingratiation and rationality influence tactics.

Extrinsic Motivation

Extrinsic motivation is characterized by an external drive to be rewarded by others (tangible or not) - that is, the energy source for performance comes from others (Barnard, 1938; Katz & Kahn, 1978). This motive typically features material or social inducements. Two general categories of extrinsic motivation have been delineated in the literature (Leonard et al., 1999). The first involves an external tangible reward-seeking motive; the second involves an external reputation-enhancing motive. Several studies have examined external motives as predictors of influence tactics used (Barry & Shapiro, 1992; Benson & Hornsby, 1988; Kumar & Beyerlain, 1991). In this study, no distinction is made between social or material extrinsic motivation - the general category "extrinsic" motivation encompasses both. Extrinsic motivation is examined in this metaanalysis as a covariate with each of the six influence tactics. Consistent with past results, we anticipate a positive relationship between extrinsic motivation of agents and their use of assertiveness, ingratiation, exchange, upward appeals, and coalition influence tactics.

Hypothesis 7: Extrinsic motivation of agents will be positively related to their use of assertiveness, ingratiation, exchanges, upward appeals and coalition influence tactics.

Method

Literature Search

The meta-analytic procedures of Podsakoff et al. (1996) for reviewing the literature and selecting studies to include in the analysis were used. To obtain the broadest sample possible, we conducted a search for studies that included empirical findings involving "influence tactics" or "influence strategies" through computer and manual methods. Computerized searches were conducted using Psychlnfo, Social Sciences Index, and Business Source Elite from 1980 to 2004. The year 1980 was chosen as a starting point because it corresponds with the seminal work of Kipnis et al. (1980), who devised the first empirical measure of intraorganizational influence tactics. In addition, we manually searched the subject index for any other references to influence tactics. Finally, references cited those articles published on the topic were manually searched for other potential articles missed in the other search procedures.

Overall, the literature search identified 237 published studies, dissertations, or book chapters that included citations of the influence tactics model. Of these initial articles, 58 could be eliminated because they were not examining the domains of interest to this analysis. The remaining 179 published studies, dissertations, or book chapters were obtained for manual examination and possible inclusion in the metaanalytic procedures.

However, upon further examination, most of these articles proved inappropriate for inclusion in this study, for several reasons. First, many of these studies were not empirical - those that were theoretical or qualitative were eliminated. Second, this meta-analysis was limited to those studies that examined the taxonomic categories of influence tactics described and measured (Kipnis et al., 1980; Schriesheim & Hinkin, 1990; Yukl & Falbe, 1990). Those articles examining constructs other than these (e.g., Wayne & Ferris, 1990) also were eliminated from further analysis. Third, many of these empirical studies had to be excluded from the study because of inadequate reporting of empirical results (correlations), which in many cases included psychometric properties (coefficient alphas) for the subscales. Studies were included in the generalization procedures if "r" was included or calculable in the reporting, regardless of whether psychometric properties were also provided.

Assessment of Measure - Pursuant to our key research objective, the articles were examined to assess the overall reliabilities of the subscales of interest. A total of 28 studies using these three measures were removed from the assessment of measures portion of the metaanalysis because of insufficient reporting of subscale reliabilities. However, these 28 empirical studies were not removed from the meta-analytic procedures testing the relationships between person-oriented antecedents and influence tactics used.

This process resulted in the identification of 39 empirical samples/studies for inclusion in the assessment of measures. Of these studies, 20 reported psychometric properties based on the measure defined by Kipnis et al. (1980), 11 reported psychometric properties using the measure developed by Schriesheim and Hinkin (1990), and eight reported psychometric properties for Yukl and Falbe's (1990) measure (see Appendix for description of studies included).

Antecedents of Influence Tactics - To examine antecedents of influence tactics used, those studies examining the influence tactics described above were included in the analysis. Several journal articles were found that examined such antecedents as impression management, Machiavellianism, self-monitoring, locus of control, social identity, intrinsic motivation, extrinsic motivation, role clarity, gender, age, context, direction of influence, and task objectives. While multiple studies existed for each of the variables listed, the desired cumulative sample size necessary to find small (.15 to .25) meta-effects was 800 (Rosenthal & Rosnow, 1991); however, this manual examination yielded large enough cumulative samples to warrant further analysis for only the first seven antecedent variables. In total, 31 independent samples were included in this portion of the meta-analysis.

Meta-Analytic Methods Used

A sample-weighted assessment of the three measures and correlates was used, as recommended by Fuller and Hester (1999). Using a sample-weighted assessment is preferred to unweighted analysis because it better controls for mean sampling variance. It was demonstrated that sample-weighted measures in meta-analysis better controlled for homogeneity/heterogeneity issues, sampling variance, large sample outliers, and rater biases. Analyses included weighted average estimates of true validity; estimates of the standard deviation of validities; and corrections made for sampling and measurement error in both the independent and dependent variables (impression management, Machiavellianism, self-monitoring, locus of control, intrinsic motivation, extrinsic motivation, and social/referent identification). Specifically, the interactive procedure for artifact distribution meta-analysis (Harter, Schmidt, & Hayes, 2002; Law, Schmidt, & Hunter, 1994a; 1994b) was used, because it results in improved accuracy of meta-analytic statistics over earlier procedures (see Schmidt & Hunter, 1999).

Because reliability information was not available for all subscales of every study, we used artifact distribution meta-analytic methods (e.g., Hunter & Schmidt, 2004) to correct for measurement error in the six criterion subscales. The reliability distributions for the criterion measures consisted of rater reports of influence tactic use.

The reliability of the independent variable was similarly estimated by using artifact distribution methods. Harter et al. (2002) argued that because the independent variables are used to predict criterion variables in practice, the practitioner must live with the reliability of measures. However, correcting for random response and transient error in the independent variable answers the theoretical question of how highly the underlying constructs relate to each other. Therefore, for the purposes of theoretical discussion, true score correlations (p estimates) also were calculated (see Tables 2 & 3).

In considering corrections for range variation or range restriction, fundamental theoretical questions arise as to whether such corrections are necessary (see Harter et al, 2002). In field studies examining organizational interpersonal influence processes such as those examined in this meta-analysis, it can be argued that range restrictions are unnecessary because we are studying populations as they exist in the workplace, and it is the workplace to which we wish to generalize our results. For this reason we have conducted the meta-analysis without range restrictions - recognizing that the findings will be more modest as a result.

We provide two meta-analytic indices for each of the 42 relationships among the seven antecedent variables and six influence tactics used by agents: (a) observed correlations with no corrections; and (b) true score correlations (p) with correction for measurement error of independent and dependent variables (see Hunter & Schmidt, 2004). Estimate (a) is downwardly biased since it ignores the role of human error and response bias in psychometric instrumentation. Estimate (b) is the true score correlation of the constructs for units across populations, corrected for error in the criterion and independent measures. No range restriction corrections were used in this validity generalization procedure.

Results

Assessment of Measures

In total there were six influence tactics for which subscale calculations were made. The number of studies ranged from two to 18, with cumulative samples ranging from 434 to 3751. Kipnis et al. (1980) and Schriesheim and Hinkin (1990) each included six influence tactics: ingratiation, rationality, exchanges, coalitions, upward appeals, and assertiveness. Yukl and Falbe (1990) measured ten influence tactics, excluding assertiveness, but including additional influence tactics not specifically measured by the previously mentioned scales. For the purpose of this analysis, only the six influence tactics measured in Kipnis et al. and Schriesheim and Hinkin were examined. The decision to not include the additional influence tactics was based on the lack of cumulative sample size and the minimal number of studies that had used Yukl and Falbe's taxonomy. For the purpose of establishing archived reliability information, several studies not used in the correlation analysis were included in the assessments of measures section of this meta-analysis, and, conversely, several studies used in the correlation analysis were not included in the reliability assessments. Decisions to include or exclude studies were always based on the variables selected and pertinent information reported in the studies retrieved. Sample-weighted assessments of the seventeen subscales, measuring six influence tactics were calculated:

For ingratiation tactics, the measure developed by Kipnis et al. (1980) earned the highest sample-weighted coefficient alpha (? = .71) of the three subscales, while sampling the largest population (3751) as well. Schriesheim and Hinkin's (1990) measure, which had set out to advance the work of Kipnis et al. (1980), achieved the lowest coefficient alpha (? = .65). Yukl and Falbe's (1990) measure achieved an acceptable reliability score (? = .70). The poor performance in this meta-analysis of Schriesheim and Hinkin's (1990) ingratiation subscale brings into question its reliability.

For rationality, the measure developed by Kipnis et al. (1980) earned the highest sample-weighted coefficient alpha (? = .72) of the three subscales and sampled the largest population (3674). Schriesheim and Hinkin's (1990) and Yukl and Falbe's (1990) measures both achieved acceptable reliability scores (? = .71 and ? = .72, respectively). All three measures achieved strong reliability coefficients (greater than .70).

For exchange tactics, the measure developed by Yukl and Falbe (1990) earned the highest sample-weighted coefficient alpha (? = .74), followed by the measures of Kipnis et al. (1980) (? = .72) and the measure developed by Schriesheim and Hinkin (1990) (? = .71). All three measures achieved strong reliability coefficients (greater than .70). Based on its coefficient alpha, Yukl and Falbe's subscale was most reliable; however, the standard deviation of these reliabilities indicates the most variation in reliability across samples.

For coalition tactics, the measure developed by Schriesheim and Hinkin (1990) was the most reliable of the three (? = .79). The subscale developed by Yukl and Falbe (1990) also achieved an acceptable reliability level (? = .75). The measure developed by Kipnis et al. (1980) earned a coefficient alpha in this meta-analysis below the desired reliability level (? = .64). This raises questions as to reliability of this subscale for research use.

For upward appeals, the measure developed by Schriesheim and Hinkin (1990) was the most reliable of the three (? = .77). The subscale developed by Kipnis et al. (1980) achieved an acceptable reliability level (? = .71). However, the subscale developed by Yukl and Falbe (1990) demonstrated the lowest reliability, albeit with a limited population sample (? = .55).

For assertiveness, the measure developed by Schriesheim and Hinkin (1990) performed similarly well to the measure developed by Kipnis et al. (1980) (? = .71 and ? = .70, respectively). Both subscales achieved the recommended reliability level (greater than .70). Yukl and Falbe (1990) did not measure assertiveness, so they were not included in this comparison. The subscale closest to assertiveness measured by Yukl and Falbe (1990) was pressure. We did not include pressure in this comparison because its definition is more aggressive and intimidating than that of assertiveness, as described by other researchers (see Kipnis, et al. 1980; Schriesheim & Hinkin, 1990).

Examination of Independent Variable Measures

The seven independent variables were assessed for their overall reliabilities. Impression management was examined in five independent samples, with a cumulative sample size of 1,131 and a strong sample-weighted alpha of .82. Machiavellianism was measured in 13 independent samples, with a cumulative sample size of 1,195 and a sample-weighted alpha of .67. All these studies used a version of Christie and Geis's (1970) measure of Machiavellianism, therefore raising concern about its internal reliability. Self-monitoring was measured in 13 independent samples, with a cumulative sample of 2,923 and a sample-weighted alpha of .75. Locus of control was measured in nine independent samples, with a cumulative sample of 1,799 and a sample-weighted alpha of .82. Social/referent identification was measured in ten independent samples, with a cumulative sample size of 2,372 and a sample-weighted alpha of .76. Intrinsic/internal motivation was measured in five independent samples, with a cumulative sample of 853 and a sample-weighted alpha of .68. Extrinsic/external motivation was measured in 12 independent samples, with a cumulative sample size of 1,990 and a sample-weighted coefficient alpha of .69.

Examination of Person-Oriented Antecedents of Influence Tactics Used

Impression Management and Influence Tactics

Five studies examined the relationship between impression management and influence tactics (see table 3). Results of the sampleweighted analyses showed a relationship between agents' impression management and their use of upward appeals (n = 749; r = .24; p = .27). Other relationships were found between impression management and agents' use of coalitions (n = 749; r = .19; p = .21), assertiveness (n = 749; r = .18; p = .23), and ingratiation (n = 897; r = .15; p = .19). The remaining tactics (rationality and exchange) demonstrated little correlation across four studies. While some relationships appear to exist between impression management and influence tactics used, the cumulative sample sizes achieved in the meta-analysis were limiting.

Machiavellianism and Influence Tactics

Eleven studies examined the relationship between Machiavellianism and influence tactics used by agents (see table 4). Results of the sample-weighted analyses demonstrated several relationships. The strongest positive relationship was between agents' Machiavellianism and their use of exchange tactics (n = 689; r = .23; p = .33) and assertiveness (n = 653; r = .23; n = .33). Other relationships were found between agents' Machiavellianism and ingratiation (n = 2,244; r = .18; p = .26) and upward appeals (n = 689; r = .16; p = .23). Coalition (n = 785; r = .13; p = .17) also had a relationship with Machiavellianism. While several relationships emerged between the Machiavellian personality and a number of influence tactics used, the cumulative sample sizes achieved from the meta-analysis continue to represent opportunities for future research.

Self-Monitoring and Influence Tactics

Eight studies examined the relationships between self-monitoring and influence tactics used by agents (see table 5). Results of the sample-weighted analyses demonstrated a relationship between self-monitoring and ingratiation tactics (n = 2,663; r = .24; p = .32). Additionally, relationships were found between self-monitoring and coalitions (n = 1,354; r = .17; p = .22), upward appeals (n = 1,350; r = .16; p = .22), exchanges (n= 1,350; r = .16; p = .22), and assertiveness (n = 1,350; r = .10; p = .15). A relationship was also found between self-monitoring and rationality (n= 1,610; r = .09; p = .12). Sample sizes in each examined relationship between self-monitoring and influence tactics were acceptable to facilitate interpretation of results.

Locus of Control and Influence Tactics

Eight studies examined the relationship between external locus of control and influence tactics used (see table 6). Results of the sampleweighted analyses demonstrated a negative relationship with rationality (n = 796; r = -.14; p = -.19). A positive relationship was found with exchanges (n = 797; r = .15; p = .17), upward appeals (n = 795; r = .13; p = .17), and ingratiation (n = 1,821; r = .14; p_ = .18). The remaining tactics (coalitions and assertiveness) showed no significant relationship with locus of control. Taken together, external locus of control shared some relationships with specific influence tactics used.

Social/Referent Identification and Influence Tactics

Five studies examined the relationship between social/referent identity and influence tactics used (see table 7). Results of the sample-weighted analyses revealed a negative relationship between social/referent identity and assertiveness (n = 890; r = -.27; p = -.34). A negative relationship also was found between social/referent identity and both upward appeals (n = 749; r = -.23; p = -.28) and coalitions (n = 749; r = -.17; p_ = -.20). The remaining three tactics (ingratiation, rationality, exchanges) were not related to agents' social/referent identification. These results demonstrate a negative relationship between social/referent identification and both assertiveness and upward appeals.

Intrinsic/Internal Motivation and Influence Tactics

Seven studies examined the relationship between intrinsic/internal motivation and influence tactics. Results of the sample-weighted analyses revealed relationships between intrinsic motivation and both ingratiation (n = 1,038; r = .08; p = .11) and rationality (n = 1,175; r= .12; p = .17). Exchanges and coalitions were not related to intrinsic motivation. Taken together, intrinsic/internal motives share little variance with influence tactics used.

Extrinsic/External Motivation and Influence Tactics

Ten studies examined the relationship between extrinsic motivation and influence tactics used (see table 9). Results of the sample-weighted analyses revealed several relationships. Significant relationships were found between extrinsic motivation and assertiveness (n = 1,427; r= .13; p = .18), ingratiation (n = 2,213; r = .18; p = .21), upward appeals (n = 823; r = .09; p = .13), coalitions (n = 1,427; r = .11; r = .13), and rationality (n = 1,105; r = .09; p = .13). With the exceptions of assertiveness and ingratiation tactics, extrinsic/external motivation explained little variance in influence tactics used.

Summary of Findings

Dispositional antecedents demonstrated some fairly consistent correlations with influence tactics used. The power of these relationships was acceptable, suggesting that the sampling effect sizes were adequate. The strongest relationships found were between impression management and upward appeals; between Machiavellianism and both exchange and assertiveness, self-monitoring, and ingratiation; between locus of control and rationality (-); between social/referent identity and both assertiveness (-) and upward appeals (-); and between extrinsic motivation and both exchanges and ingratiation. Intrinsic motivation also had significant relationships with influence tactics used. It appears as though personoriented antecedents of influence tactics have been fairly well established in the literature to date, and, judging by the results of these analyses, offer some insight for understanding influence behavior.

Discussion

The purpose of this study was to examine the dispositional effects of influence tactics used. Several dispositional trends stemmed from the meta-analysis. Impression management seems to share a relationship with the degree to which agents make use of upward appeals, indicating that individuals engaged in self-presentation or image enhancement are prone to seeking support from those in higher positions of authority when attempting to influence others. A person with high degrees of impression management may therefore go over the heads of supervisors to gain desired influence.

Machiavellianism was related to both exchanges and assertiveness, indicating that individuals willing to achieve goals by any means are more likely to use exchanges and assertiveness to influence others. Self-monitoring and ingratiation tactics were related, indicating that individuals capable of adapting their behaviors and attitudes to the situation are more likely to use flattery or compliments when attempting to influence others.

Locus of control was negatively related to the use of rationality. This seemed fitting since a high score on locus of control represents external locus of control - so individuals high in this disposition believe they are not capable of persuading others. Similarly, these results suggest that individuals who lean toward internal locus of control are more likely to use rationality when attempting to influence others. Social identity was negatively related to assertiveness and upward appeals, suggesting that the more connected or affiliated individuals are to an organization or group, the less apt they will be to using assertiveness or upward appeals. Conversely, lower social identification may lead to increased use of assertive and upward appeals. Extrinsic motivation was positively related to the use of ingratiation and exchange tactics. This finding indicates that individuals who are motivated externally are more likely to flatter others and are more likely to offer tangible or social exchanges. Individuals low in extrinsic motivation are less likely to use flattery and are less likely to offer tangible or social exchanges. This may be somewhat consistent with Kegan's (1982) developmental lens perspective, where our own motives reflect perceptions of others' motives. An extrinsic motivation lens may lead to increased use of extrinsic reward use when attempting to influence others. Taken together, these findings offer some support for dispositional effects in the influence process, as several significant correlates were found. However, the true score correlations ranged from -.34 to .33, which approximates 10% of variance explained in these relationships. Other dispositions not studied in this analysis may also contribute to explained variance in influence tactic selection. Meta-analyses of other dispositional effects will provide further evidence of its substantive role in organizational behavior.

Reconsidering Measures of Influence Tactics

The three measures of influence tactics achieved the recommended reliability level of .70 for most of its subscales. However, there may be opportunities to replace less reliable subscales with corresponding subscales from other measures. This approach would more reliably capture the domains of interest, improve subscale reliabilities and decrease the risk of measurement error. Admittedly, individual item analysis would provide a more robust test of the constructs; however, given the limitations of validity generalization procedures, the raw data from all research studies included in the metaanalysis were not obtainable.

The most reliable measures for each subscale have been identified (see Table 1), alluding to a more holistic approach to instrument selection. An integrative measure would draw from the Kipnis et al.(1980) for ingratiation and rationality; Schriesheim and Hinkin (1990) for coalitions, upward appeals, and assertiveness; and Yukl and Falbe (1990) for exchange.

A Call for More Research Using Influence Tactics

This research effort is an important step for testing dispositional antecedents of intraorganizational influence tactics. Generally, cumulative samples in excess of 800 were included in the meta-analysis (Rosenthal & Rosnow, 1991). Covariates that did not have large enough samples in the aggregate included role clarity, role ambiguity, gender, age, education, cognitive styles/types, and bases of social power. These variables require more research before validity generalization may ascertain their general effects.

Given these sample size requirements, more research is needed to generalize the relationships between impression management and ingratiation; Machiavellianism and both coalitions and upward-appeals; self-monitoring and coalitions; exchanges and upward appeals; locus of control and ingratiation tactics, social identity and coalitions; and intrinsic motivation and both ingratiation and rationality; and more research is needed to generalize the relationships between extrinsic motivation and both rationality and coalitions - as each of these relationships did not achieve the recommended cumulative sample size (800) (see Table 11).

A summary of findings in the meta-analysis are summarized to show direction of relationships, significance (or no significance), sample size requirements met, and to identify those relationships that were sought but not analyzed due to sample size limitations (Table 11). This summary provides directions for future research in several areas.

Several relationships between person-centered variables (such as impression management, Machiavellianism, self-monitoring, locus of control, social identity, intrinsic motivation, and extrinsic motivation) and influence tactics used were revealed in this study. Using validity generalization procedures, the meta-effects were assessed providing some conclusions for these relationships. With additional research, the antecedents of influence tactics can be better understood and empirically derived. It is our hope that the field's understanding of the influence process will continue to evolve in the literature.