The optimum level of argumentativeness for employed women.

Discussion of substantive issues among coworkers is vital to group effectiveness (see, for example, Burnett, 1991), and workplace discussion is rightfully encouraged by many business communication textbooks, which warn of the pitfalls of "groupthink." However, when an employee disagrees with the

boss or a colleague, the employee must carefully consider the merits of pursuing the issue to the point of an argument. Although some recent business communication textbooks explain how to argue (e.g., Ewald & Burnett, 1997; Locker, 1997), there is little discussion in the business communication literature of the potential benefits and risks of argument, or of the personality predisposition known as argumentativeness that leads some individuals to argue more than others. This predisposition is of interest to business communication practitioners and teachers for two reasons: (a) argumentativeness can be influenced by training, and (b) highly argumentative individuals have been shown to have characteristics linked with workplace success.

Achievement of supervisory authority is a widely recognized indicator of workplace success. Since supervisors must both present departmental results to upper management and convey executive decisions to their own subordinates, the possibilities for argumentative discussion are numerous, and the potential importance of those arguments is great. The supervisor's position is held at the pleasure of management, which must have confidence in the supervisor's ability to "communicate about corporate issues" (Gorden, Infante, & Izzo, 1988, p. 20). The importance of this and other management skills increases at higher supervisory levels.

The value of argumentativeness for business people seems clear, yet, for female employees, a contradiction emerges regarding what Infante calls a "desirable communication trait" (1982, p. 146) in the workplace. Recent assertiveness research shows that males may react negatively to females' use of powerful communication (Carli, 1990; Wiley & Eskilson, 1985), defined as direct, assertive communication, in contrast to a traditional, tentative communicative manner (Wiley & Eskilson, 1985, p. 996). Yet argumentativeness is an assertive trait (Rancer, Baukus, & Infante, 1985) which claims numerous positive assertions over the last 15 years of study (Infante & Rancer, 1996, p. 321). The contradiction between the argumentativeness and the assertive research is striking. For ambitious women who have been encouraged to be assertive, or who are predisposed to argumentativeness, these findings pose a dilemma which this paper attempts to resolve.

In this paper, I examine the relationship between argumentativeness and women's success in organizations, as operationalized by the women's supervisory levels. I begin by defining argumentativeness, including descriptions of how it relates to assertiveness and aggression, and how it is quantified. Then, I review the literature related to the following issues: the benefits of argumentativeness, relevant cultural perceptions/misperceptions, current workplace demographics, and the contradictory research regarding the value of assertiveness for women. Finally, I present the results of my own study which indicates that a moderate level of argumentativeness is optimum for women seeking supervisory advancement and explore the implications of this new information in research, in the classroom, and in the workplace.

Literature Review

The distinctions among aggressiveness, assertiveness, and argumentativeness are important to understand since the academic definitions, while validated by research, may differ from popular connotations and, therefore, workplace perceptions. Argumentativeness and assertiveness are academically defined as distinct, constructive aspects of aggressive behaviors (Infante & Rancer, 1982).

Argumentativeness is One Type of Assertive (and, Therefore, Aggressive) Behavior.

Aggressiveness is defined as applying force "physically and/or symbolically . . . to dominate . . . or . . . defeat and perhaps destroy the locus of attack" (Infante, 1987a, p. 158). The locus of attack, Infante notes, can be "a person's body, material possessions, self-concept, position on topics of communication, or behavior" (p. 158). "Symbolically" applying force would include verbal expressions of aggression, such as teasing. Aggressive behavior is typically perceived as male behavior in the United States (Infante, 1987a; 1989), although both males and females may exhibit aggressive behaviors. Examples of aggressive behaviors range from competitive sports (a constructive use of aggression) to insults to an opponent's self-esteem (a destructive use of aggression; Infante, Rancer & Womack, 1993). Aggressiveness is comprised of both positive (assertiveness) and negative (hostility) components. This paper deals only with the positive assertive components.

Assertiveness is defined as a "person's general tendency to be interpersonally dominant, ascendant, and forceful" (Infante, 1987a, p. 165). Assertiveness, although aggressive, is generally considered positive behavior in the United States for both men and women (Infante & Rancer, 1996; Varner & Beamer, 1995). U.S. women have been encouraged to communicate assertively for at least three decades (Rich & Schroeder, 1976). Argumentativeness is one type of assertive communication, although not all assertive individuals are argumentative.

Argumentativeness is defined as a "generally stable trait which predisposes [a person] to advocate positions on controversial issues and to attack verbally the positions" of others (Infante & Rancer, 1982, p. 72). Argumentativeness is a type of aggressive behavior that is conceptualized as constructively assertive rather than hostile. Thus, the argumentative person is issue-oriented in disagreements, as distinct from the verbally aggressive person, who is more likely to make destructive personal attacks on an opponent. An argumentative person may argue strongly and frequently, but his or her comments focus on the issue under discussion.

A highly argumentative individual not only provides reasons when asked and approaches situations logically but also takes opportunities to argue that others might ignore. Such an individual enjoys arguing and believes it brings practical benefits (Rancer, Kosberg, & Baukus, 1992). In contrast, low argumentatives perceive argument as an act of hostility to be avoided (Rancer & Baukus, 1987). For both high and low argumentatives, pursuing an argument is influenced by situational variables such as topic, importance of the issue, perceived likelihood of success, ego-involvement of participants, and relationship with the opponent (Infante & Rancer, 1987b). Research has not identified the tone of an argumentative individual's remarks, but superiors find an affirming style, that is, a "friendly, relaxed, and attentive" manner more "appropriate" in the workplace (Infante & Gorden, 1989, p. 87). Both one's attitude toward audience and beliefs regarding argument indicate the level of one's argumentativeness.

Researchers identify a person's argumentativeness with the argumentativeness scale developed in 1982 by Infante and Rancer. Their Likert-style scale is based on twenty questions: Ten of these questions assess the tendency to approach argument, and ten assess the tendency to avoid argument. For example, the scale questions whether respondents "enjoy a good argument over a controversial issue" or are "happy when they keep an argument from happening" (Infante & Rancer, 1982, p. 76). The argumentativeness scale measures attitudes toward or beliefs about argument, rather than actual skill in arguing; only two of the twenty questions include a self-perception of ability. A zero score (the mid-point of the scale), indicates a person is moderately argumentative and more likely to be influenced by situational considerations (p. 73). Women tend to have a mean score of zero; males' mean argumentativeness is slightly higher. Since the scale's validity was first established in 1982, several studies have supported it (see Infante & Rancer, 1996, for a complete literature review).

Argumentativeness Produces Positive Results.

A substantial body of research since the early 1980s has linked argumentativeness to several positive outcomes relevant to workplace success: better decision-making and problem-solving skills (Infante, 1989), increased competence in communication (Onyekwere, Rubin, & Infante, 1991), and enhanced credibility (Infante, 1985). These qualities may explain why researchers have found that high argumentatives are more often selected as group leaders (Schultz, 1982), have a greater likelihood of succeeding as supervisors (Infante, Rancer, & Womack, 1993), and are perceived as more effective in their upward communication by organizational subordinates (Infante & Gorden, 1985b). Further, in a nine-variable canonical correlation study examining outcomes for employed supervisors (94 men and 38 women), argumentativeness was "slightly related" to earning higher salaries and substantially related to subjective self-perceptions of career satisfaction and upward effectiveness, although there was no relationship to more objective variables: promotions, management level, and number of subordinates (Infante & Gorden, 1985a).

In recognition of the multiple positive outcomes associated with argumentativeness, ambitious employees, both male and female, might conclude that increased argumentativeness would benefit them. Both women's and men's argumentativeness can be increased through training (Anderson, Schultz, & Staley, 1987; Infante, 1985), and if argumentativeness is linked to business success, we may wish to cultivate it in the classroom. However, the business communication literature identifies several diverse perspectives that should be taken into consideration.

Whether a Behavior Is "Aggressive" Depends on the Audience.

Cultural perspectives concerning the components of aggressive behavior differ markedly and can lead to misunderstandings even in a single country (Shuter & Turner, 1997). For example, researchers have found that a style of aggressiveness which includes personal attacks and verbal aggressiveness is more prevalent among African Americans (Nicotera, Rancer, & Sullivan, 1991, as cited in Infante & Rancer, 1996). In observing such interchanges, U.S. whites tend to believe that the interchange will lead to violence (Kochman, 1981, cited in Locker, 1997, p. 339). In contrast, one might choose to describe African-American speech as a tendency to speak "powerfully and confidently" (Kenton & Valentine, 1997, p. 63). African-American and European-American corporate women use different strategies "to reduce conflict" in the workplace (Shuter & Turner, 1997, p. 89). Both groups see those of their own race as reducing conflict: European Americans see African Americans' behavior as "maintaining the conflict" (p. 89), while African-American women tend to view European-American women as more conflict-avoidant and believe that their own direct approach reduces conflict. The direct approach, if focused on the issues, would fit the definition of argumentativeness.

However, although differences in verbal aggressiveness have been noted between racial groups, Infante and Rancer's scale shows no differences in argumentativeness within the United States between European Americans and other ethnic groups or between regional groups of Americans. The groups tested include European Americans and African Americans, as well as Asian Americans and Hispanics (Nicotera & Sullivan, 1991, as cited in Infante & Rancer, 1996). A regional comparison showed that argumentativeness differences only approached significance between northerners and southerners (p = 0.057; Geddes, 1992, as cited in Infante & Rancer, 1996, p. 335). However, there is a significant difference in argumentativeness scores by gender (Nicotera & Rancer, 1994), which may lead to perceptual differences. Further research on such perceptual differences is essential (Fine, 1996), since some audiences may be more likely than others to see argumentativeness as a desirable behavior. When that audience makes promotion decisions, the impact of perceptual differences may be crucial.

Despite women's recent gains in mid-level management (Chaney & Martin, 1995, p. 45), men continue to dominate the executive ranks. In the mid-1990s, only 5% of upper management positions in the United States were held by women (Hermon, 1996; Glass Ceiling Commission, 1995). Females comprise only 2.6% of the U.S. corporate officers (vice president or above) in Fortune 500 companies (Haglund, 1996). Therefore, men make most upper-level promotion decisions. Males' perceptions of females' behaviors are clearly relevant for females' success in attaining executive-level promotions.

Is Argumentativeness Advantageous for Women?

Both males and females expect males to be more argumentative and more verbally aggressive than females (Nicotera & Rancer, 1994). These stereotypes emerged in a multi-national cross-cultural study, which found females consistently described with traditional adjectives such as submissive, timid, changeable, and weak, while men were termed dominant, forceful, rational, and strong (Williams & Best, 1990, cited in Nicotera & Rancer, 1994, p. 290). Such stereotypes have a substantial history, as we can see by reviewing the results of a few studies. In the early 1980s, females were still described as "antisocial" for using malelike "aggressive persuasive strategies" (Burgoon, Dillard, & Doran, 1983, p. 292) instead of abiding by 1950s style traditional norms. Two contradictory studies in 1985 sent mixed signals about gender expectations. In one study, male students rated female high argumentatives as more credible (Infante, 1985); in another, male students predicted female applicants using less powerful speech would be better liked by colleagues and subordinates (Wiley & Eskilson, 1985, p. 1000). More recently, male students awarded greater influence and trust to tentative female speakers (Carli, 1990). In contrast, females in all the studies reacted positively to the female speaker's powerful communication. More recent studies of female assertiveness have yielded similarly "complex and often contradictory [research] findings" (Gardner, Peluchette, & Clinebell, 1994, p. 133). Perceptions of the components of assertiveness, such as argumentativeness, may help explain this contradiction. Researchers recently have shown that the use of masculine behaviors can increase perceptions of women as "male" with a resulting negative overall perception (Nicotera & Rancer, 1994).

Women have been hailed as excellent managers since they exhibit traditionally "male" management attributes (e.g., resourcefulness, decisiveness) as well as traditionally "female" qualities (e.g., communication, feedback, and empowerment; Collingwood, 1995). Yet the successful women in a 1995 survey cited "developing styles that don't threaten male managers" as a factor in their career success (Haglund, 1995). These stellar performers stated that negative stereotyping by (mostly male) bosses was a major hurdle in attaining career promotions. Such widespread viewpoints might also be expected at executive levels, where organizational norms have been well inculcated. While it would be desirable for organizations (and humanity in general) to discard stereotypical norms (Sauer, 1994), research continues to show that these behavioral expectations endure.

Instead of short-circuiting the stereotypes, some well-intentioned business communication textbooks may be perpetuating these generalizations. Pomerenke, Varner, and Millar (1996) found that 12 randomly selected textbooks included pictures of both male and female "managers," but males were "serious" 86% of the time, while females were shown as "serious" in only 64% of the pictures. The remaining females were shown smiling. Such pictures appear "to reinforce generally held cultural role expectations" of women as "friendly, nurturing and happy" (p. 42) and strengthen the image of male behaviors as correct for business management. Kenton and Valentine (1997) point out, "Similarity is a criterion for perception of credibility. . . . men tend to be more credible to other men. Women may be more credible to some other women" (p. 34).

Adopting Male Behaviors is Problematic for Females.

Women who attempt to gain credibility by adopting male behaviors (such as argumentativeness) may be misunderstood by both men and women. Nicotera and Rancer (1994) discovered that both men and women associate higher argumentativeness with higher verbal aggressiveness when (but only when) the source is of the opposite sex (1994). That is, men tend to stereotype women as being verbally aggressive, and vice versa. The research showing males' negative responses to females' "powerful" communication (Carli, 1990; Wiley & Eskilson, 1985) may explain males' reluctance to promote highly argumentative women. The result could be that the careers of highly argumentative women are jeopardized, in contrast to the positive outcomes found for men.

Low-argumentative females tend to perceive arguing as conflict. However, argumentativeness is not necessarily among the elements in academic definitions of conflict: "differences, expressed struggle, incompatib[ility], frustration, interference, perception, and interdependence" (Borisoff & Victor, 1989, p. 2; emphasis in original). Instead, conflict, by definition, must be perceived by both parties (Schultz & Anderson, 1984; Donohue, Weider-Hatfield, Hamilton, & Diez, 1985). High argumentatives' preference to engage in argument as a learning experience (Rancer, Baukus, & Infante, 1985; Kenton & Valentine, 1997) may be perceived by some interactants as conflict, although academics do not define it as such.

Research Question

Since women are a minority among those in power, it is important to study those women who do rise to the top to discover patterns. Past research showing positive outcomes for argumentativeness for business people contrasts with other research on women's assertive behaviors, resulting in inconclusive evidence as to whether females' argumentativeness helps their careers. As recently as a decade ago, researchers may have found it difficult to collect sufficient data because of the limited number of women in management. Today, with women holding 48% of the managerial (but not executive) positions (U.S. Dept. of Labor, Women's Bureau, 1996; Haglund, 1996), it should be possible to obtain the data necessary to understand the relationship of argumentativeness to women's success. A study which focuses on women and their success in relation to argumentativeness would provide additional insight.

Rather than asking whether argumentative behaviors should gain male or supervisory approval, this article studies actual results obtained by women. There are more women in today's workforce than ever before, but the Glass Ceiling remains. If women's high argumentativeness is not yet culturally acceptable, women need that information to choose whether to modify their actions. Argumentativeness is a communication characteristic that may have direct bearing on a woman's promotional chances. The research question guiding this study is: What is the relationship between argumentativeness and women's supervisory levels in organizations?

Methods

The sample of nearly 400 women employed full-time was obtained with the assistance of students at Eastern Michigan University and Wayne State University, two large midwestern state universities with diverse populations. In an effort to gain a broad and diverse but representative sample, students were asked to voluntarily distribute the surveys to women they knew who were employed full-time in any profession, at any supervisory level. About 80% of the students are employed and so would have access to employed women through their work, as well as through their families and friends. There is obviously some self-selection in students who volunteered to distribute the surveys; the women in the sample may have economic or social backgrounds similar to the students'. However, there is no reason to suspect that this procedure would result in a sample biased with respect to argumentativeness.

The sample is at least grossly representative of the population of working women. The validity of my sample is supported by its normality in regard to argumentativeness, experience, and age. One would not expect supervisory levels to be normally distributed, as they are not "normal" in the population.

The strength in choosing the sample in this way is that it provided a sample size large enough for generalizability. Gaining access to women was more efficient by personal contact than it would have been through the mail, which typically has very low return rates. The sample I obtained is more than twice the size required to assess statistical significance for the Detroit metropolitan area population of working women (following Fowler, 1993).

All data are anonymous and self-reported. The women returned the surveys directly to the researcher through the mail.

Key Variables in the Survey

The women were surveyed for several variables relevant to their workplace success: their argumentativeness, supervisory level, number of years of relevant experience, and other related and demographic information.

I measured the women's argumentativeness using Infante and Rancer's argumentativeness scale, which has demonstrated reliability and validity (1982).

The women were asked to identify their supervisory level, operationalized as the number of levels of supervisory authority. On the questionnaire, a supervisor was defined as someone who makes hire-and-fire recommendations and evaluates employee performance. For example, a woman who did not meet the definitional criteria categorized herself as supervisory level zero (a non-supervisor), in spite of any other managerial responsibilities. A woman whose authority was limited only to non-supervisors was a level-one supervisor. A woman who had level-one supervisors reporting to her classified herself as a level-two supervisor, and so forth. Supervisory level defined in this way is a much more valid and reliable measure of workplace success than job titles such as manager, director, coordinator, or supervisor. Increasing levels of hire-and-fire authority undoubtedly represent increasing levels of status, authority, and responsibility in an organization and should show less variation across professions or industries than would, for example, salary, number of subordinates, or job title. Use of such a stringent, conservative measure for the workplace success construct enhances the meaningfulness of any correlations with the variable that are found. Supervisory level is independent of experience, which was separately reported.

The number of years of relevant experience was defined as any relevant experience that a woman would claim (e.g., during an employment interview) had helped her prepare for a promotion. Respondents determined what was "relevant." Typically, experience is a prime consideration for advancement to a supervisory role and therefore must be considered as a possible confounding variable.

Sample

Three hundred eighty-six women employed across numerous industries responded to the survey, a 48% return rate. A subset database was established which contained only those respondents who were full-time employees and completely answered, at minimum, questions regarding supervisory level and relevant experience, and the argumentativeness questionnaire. The resulting subset of 333 respondents includes 204 non-supervisors (level-zeroes) and 129 supervisors. Numerous professions - from office worker to chief financial officer - were represented. The respondents were employed in the varied industrial categories of manufacturing, health care, finance/insurance/real estate, retail/wholesale trade, education, public administration, and legal services/engineering. The women's ages ranged from 20 to 61 years, with educational preparation from high school to JD, PhD, and MD degrees (Schullery, 1997).

Table 1 shows the sample distribution by supervisory level. The distribution is skewed to the low end, as is true of the population. Sixty-one percent of the respondents indicated that they did not hold supervisory responsibilities (n = 204). Supervisors were distributed over the first, second, third, and fourth supervisory levels (n = 129) at 26%, 7%, 4%, and 1.5%, respectively. Two level-four respondents noted that they were "second-in-command."

[TABULAR DATA FOR TABLE 1 OMITTED]

Table 1 also shows the sample's distribution of argumentativeness, both overall and by supervisory level. The sample was approximately normally distributed with respect to argumentativeness. Argumentativeness ranged from -35 to +30 (M = -0.13; s = 12.1). No differences among the mean argumentativeness scores by supervisory level achieved the 0.05 level of significance.

The sample was approximately normally distributed with respect to relevant experience. The women's experience level ranged from one-half year to 36 years, with nearly 50% of the women having had at least 10 years of experience and 15% having had 15 years or more. The extent of the women's experience is, as expected, highly correlated with supervisory level (r = 0.31; p = 0.000) for all respondents (N = 333). There were no significant differences in mean argumentativeness among the respondents' industries at any supervisory level.

Statistical Methods

All statistical analysis was done using the SPSS software package. Two main types of tests were used to examine the data. The average argumentativeness scores at each supervisory level were tested by the independent samples t-test for significant differences between any two levels. The relationships between the primary independent variables (argumentativeness and relevant experience) and the dependent variable (supervisory level) were examined by bivariate and standard multiple regression analyses. Regression analysis was used rather than ANOVA to take advantage of all of the information contained in the continuous variables used in this study. Using continuous variables, rather than, for example, dividing samples by median splits increases the power of the statistical test and allows patterns to emerge that might otherwise be obscured (Aiken & West, 1991).

I used two levels of regression analysis. First, a bivariate regression investigated the effect of each independent variable (argumentativeness and relevant experience) separately on women's advancement. Then, I combined the two predictor variables in a multiple regression to test for any unique or confounded effect of each variable while controlling for the other. Following this analysis, I repeated the bivariate and multiple regression procedures to examine an unexpected apparent relationship between moderation in argumentativeness and supervisory level.

Results

The bivariate regression of argumentativeness and supervisory level (SL) shows that argumentativeness alone has no net linear relationship with supervisory level (r = 0.047; p = 0.19). A multiple regression including the relevant experience variable also failed to show a net relationship between supervisory level and argumentativeness.

However, despite the lack of a net, or overall, linear relationship, the possibility of some more complex relationship among the variables still exists. The pyramidal distribution of argumentativeness evident in Figure 1 suggests a possible dependence of supervisory level on distance from the center of the argumentativeness scale, with women at higher supervisory levels showing a pattern of scores closer to the zero midpoint of the scale. That is, higher supervisory level appears to be favored by moderation in argumentativeness. This finding led to a new hypothesis: Moderation in argumentativeness (nearness to midpoint) increases with supervisory level. I tested this hypothesis with a bivariate regression of supervisory level on a new variable, the absolute value of argumentativeness.

The absolute value of argumentativeness is the unsigned distance between a respondent's argumentativeness score and the zero midpoint of the argumentativeness scale. Thus, the scale for the absolute value of argumentativeness ranges from zero to 40, with zero corresponding to maximum moderation in argumentativeness, and 40 corresponding to the absence of moderation, due to either extreme low (-40) or extreme high (+40) argumentativeness. Table 2 shows the mean absolute value of argumentativeness, along with the range and standard error, for each supervisory level. A decreasing trend from supervisory levels one through four is apparent for all these measures of variability.

Regression analysis indicates a significant negative correlation between the absolute value of argumentativeness and supervisory level once women attain supervisory level one (r = -0.1772; p = 0.022). The absolute value of argumentativeness explains 3% of the variance in supervisory level ([r.sup.2] = 0.031) after women reach supervisory level one (Schullery, 1997). This supports the hypothesis' prediction that higher level supervisors are more moderate in their argumentativeness. There is no significant relationship for the entire group of respondents (r = -0.066; p = 0.11), indicating that moderation in argumentativeness is not significant with respect to women's promotion to a first supervisory level.

Table 2

Variability in Argumentativeness for Each Supervisory Level

Supervisory    Standard error of    Range    Mean absolute value of
level          argumentativeness               argumentativeness

0                    12.10            58              9.97
1                    12.83            61             10.58
2                    11.06            40              9.57
3                    10.94            40              8.00
4                     6.76            16              5.20

The actual importance of this seemingly small effect was assessed by comparison with the effect of relevant experience, a variable that is undoubtedly important in promotion decisions. Of all the variables measured in this study, experience has the highest bivariate correlation with supervisory level (r = 0.3141; p = 0.000). A simple comparison of the squared correlation coefficients suggests that the absolute value of argumentativeness ([r.sup.2] = 0.031) is about one-third as important as relevant experience ([r.sup.2] = 0.099) in explaining the variability of supervisory level. However, the above comparison is strictly valid only in cases where the predictor variables are totally unrelated to each other, which is not the case here. Although the absolute value of argumentativeness and relevant experience are not significantly correlated (r = -0.094; p = 0.15 [one-tailed]), neither is their correlation zero.

To obtain a more reliable comparison of the impact of these two variables on supervisory level, a multiple regression was performed. The results are presented in Table 3. When relevant experience is included with the absolute value of argumentativeness in a multiple regression analysis, the two account for a total of 9.2% of the variance in supervisory level ([R.sup.2] = 0.092; p = 0.001). This procedure provides two statistics that may be used to assess the relevant importance of the variables: beta weights and squared semipartial correlation coefficients. Each of these will be discussed below, but it is important to recognize that there is no certain way to identify the relative importance of intercorrelated predictor variables, even though, in the present case, both the intercorrelation of -0.094 and the tolerance of 0.991 indicate a relatively small 1% shared variance.

As Kachigan (1982) notes, the relative importance of two predictor variables may be estimated by calculating the ratio of their squared beta weights. For my sample, this calculation suggests that the absolute value of argumentativeness is 39% as important as relevant experience in explaining the variability of supervisory level. Tabachnick & Fidell (1996) prefer comparison of the squared semipartial correlation coefficients, which correspond to the fractions of uniquely explained variability. The ratio of the squared semipartial correlation coefficients suggests the same 39% relative importance of the absolute value of argumentativeness compared to experience as inferred from the beta weights.

[TABULAR DATA FOR TABLE 3 OMITTED]

Discussion

Although previous research found positive workplace-related outcomes for high argumentativeness and implied that increasing argumentativeness may be increasingly beneficial, my study demonstrates that there is no such simple relationship between supervisory level and argumentativeness for women. My results suggest that it is valuable for women in supervisory roles to be argumentative, but only up to a point. In contrast to research that identified high argumentatives as competent communicators in leadership roles, these results demonstrate that proximity to the midpoint of argumentativeness is more predictive of women's promotional success than is the direction (high or low) of their argumentativeness. In fact, the data indicate a linear relationship between supervisory level and moderation in argumentativeness (proximity to mean) for women in advanced supervisory positions. The finding that high argumentativeness may jeopardize supervisory advancement has not previously been demonstrated. Therefore, argumentativeness is seen to be related to women's success, but in a more complex manner than previously identified.

Moderation in argumentativeness accounts for an apparently small 3% of the variance in supervisory level. However, considering all the various qualifications and attributes that are relevant to promotion, the finding of a single communication characteristic having as much as a 3% influence toward women's advanced promotion(s) is remarkable. The practical significance of these results for female supervisors is further emphasized by the beta weight and squared semipartial correlation coefficient comparisons, which suggest that argumentative moderation is approximately one-third as influential as is relevant experience, a variable that is certainly important for supervisory promotions.

Theory Supports My Results.

The finding of moderation in argumentativeness as valuable for supervisors is consistent with current organizational theory. Moderation in argumentativeness might be essential to the process of advocacy, which requires "championing ideas, proposals, actions or people" to superiors (Eisenberg & Goodall, 1993, p. 230). In these days of limited resources, supervisors must be able to argue strongly for controversial positions and counter those who would attack their positions. Supervisors must become advocates for their departments, developing persuasive arguments to accomplish departmental goals. To the extent that argumentativeness and advocacy work together to facilitate productive organizational participation, both are valuable. At very low levels of argumentativeness, the advocacy process might never begin to discuss substantive issues (Burnett, 1991); at very high levels, the process might turn from pursuing organizational goals to the pursuit of an individual argumentative's "win" (Infante, 1981). Either extreme of argumentativeness, therefore, could impair teamwork and organizational goals, and both extremes appear to be outside upper management's behavioral expectations for higher-level supervisors, at least for women.

Are These Results Unique to One Industry or Profession?

The sample consisted of full-time employed women (non-owners) in the metropolitan Midwest. To encourage candor in replies, the women were assured of anonymity, which makes it impossible to identify women from the same company, or to determine company characteristics such as age, culture, climate, or the gender of executives.

The diversity of the industries and professions represented by the respondents is both a strength and a limitation of this study. The strength is that the findings are more likely to be generally applicable and not limited to a particular industry or profession. I broadly segregated the respondents by industry and found no significant differences. The limitation is that it is not possible to sort the respondents into statistically useful groups based on the nature of their job activities or duties. The women identified their professions with job titles and brief descriptions, and I attempted without success to categorize positions using the U.S. Government Commerce Department job classification scheme. For example, the government classifications distinguish five possible types of "programmer consultant" and twelve overlapping categories for secretary/bookkeeper/receptionist/database manager. Several respondents identified their roles as including more than one of these responsibilities, making classification unreliable. Thus, the possibility remains that the role of argumentativeness may be more or less influential in some particular profession or industry. In a future study, the current findings, obtained across multiple industries, might be more narrowly examined among women within a single industry, while also controlling for profession within the industry. For example, it may make a difference if a woman is in a "pink collar" profession, such as human resources, but works in a male-dominated industry, such as computers or heavy manufacturing. Such a study should involve extensive interviews, observations, and research in specific companies in a common industry. A wide range of relevant situational details under the employers' control as well as the women's approaches to argumentative encounters, including the nonverbals associated with an "affirming" style, should be determined.

Are These Results Unique to Women?

This study examined only two variables among the many that might conceivably be responsible for promotion to a supervisory position. Nonetheless, these two variables appear to have substantial influence. By focusing exclusively on women, I avoided a substantial set of troublesome confounding variables - the gender-related issues. Since current demographics show that fewer than 5% of upper management positions are held by women, I assume that most women in this study were promoted by males. Further work might expand the focus to other measurable communication characteristics to determine the import of each separately, as well as controlling for the gender of supervisors who granted the promotions.

There is reason to suspect that the value of moderation may be unique to women. Although positive outcomes for high argumentativeness have been reported for both sexes (Infante & Rancer, 1996), both males and females perceive women as being less argumentative than men (Nicotera & Rancer, 1994). Males' unfavorable reactions to females' powerful communication may result in a "generalized" perception of other "negative" traits (Nicotera & Rancer, 1994), and lead to a possible lack of executive confidence in highly argumentative women. To be rewarded with high-level supervisory positions, women will have to gain executives' confidence in their ability to "communicate about corporate issues" (Gorden, Infante, & Izzo, 1988, p. 20). High-argumentative women may have difficulty obtaining that confidence (a) due to males' generalized negative perceptions of women's aggressive behaviors (Nicotera & Rancer, 1994); and (b) due to a greater likelihood of gaining credibility from same-gender individuals (Kenton & Valentine, 1997), instead of gaining the necessary credibility with executive male decision-makers. Thus, it may be that women require a disproportionate level of moderation in argumentativeness in order to compensate for a "generalized" negative perception. These concerns call for a similar study of employed males to test whether moderation in argumentativeness is valued for both males and females.

Another issue requiring further study is the lack of significant correlation between moderation of argumentativeness and supervisory level for the entire group. That is, why doesn't the relationship hold for female employees moving from non-supervision (level zero) to level one? One speculation is that promotion to supervisory level one may largely depend on meeting one or more unique, one-time qualifications, such as earning a college degree, acquiring a certain experience level, or satisfying union regulations. Another possibility is that the supervisory level-ones are more likely to have been promoted by a female boss, since women hold 48% of managerial positions (U.S. Dept. of Labor, Women's Bureau, 1996; Haglund, 1996). As the supervisory levels rise, current demographics indicate further promotions are more dependent on executive males. A final possible explanation is that a qualitatively different set of criteria may be operating for the higher-level promotions, perhaps including a greater emphasis on communication skills. That is, organizational or management communication norms may become more stringent among senior executives. Thus, as supervisory level increases, moderation in argumentativeness would become more essential.

An additional issue regards the time sequence for the development of argumentativeness with respect to the time(s) of promotion(s). A longitudinal survey would be required to understand whether successful women come into the organization as moderate argumentatives and are promoted for this quality, or if they become more moderately argumentative as they observe corporate culture, gain experience and training, and earn promotions. The cause versus effect conundrum common to cross-sectional survey-based research remains.

Implications for Pedagogy

Some business communication textbooks contain a section explaining how to argue, and, similarly, considerations of audience are well-prescribed in the business communication literature. However, my finding that female supervisors are most likely to be moderately argumentative suggests that would-be female executives should be trained to argue effectively but also to identify when to argue. Also, the data suggest that providing the same argumentativeness training for all women may be counter-productive. Training directed towards increasing women's argumentativeness should be limited to low argumentatives. Such efforts might well include open discussion of the negative connotations that such individuals commonly associate with arguing. Success at increasing argumentativeness levels has been described (Anderson, Schultz, & Staley, 1987; Infante, 1985). In contrast, my data suggest that female high argumentatives who aspire to high-level supervisory positions should be alerted to the possible negative consequences of pursuing their inclination to argue. They would be well served by training which focused on an analysis of situational variables, such as relationship to the opponent, knowledge of the topic, importance of the issue, and the probability of success (Infante & Rancer, 1993) before yielding to the urge to argue.

Both groups of women, although aware of situational variables, might benefit from classroom simulations which analyze the importance of such variables (e.g., opponent, topic, importance of issue, and probability of success), thereby training them to adapt (either up or down) their natural level of argumentativeness to organizational norms and specific situations. Also, both groups would benefit from understanding that research has shown that argumentativeness should be conveyed in a "friendly, relaxed, and attentive" (affirming) manner (Infante & Gorden, 1989, p. 87). Becoming skilled in adapting one's argumentativeness to simulated norms should help students develop the aptitude for audience analysis urged by business communication textbooks, and better prepare them for successful careers.

Implications for Managerial Practice

Both current executive management and aspirants to such positions need to be aware of the disadvantages of either extremely low or extremely high argumentativeness. In the course of management training programs, instructors might inform candidates that, for women, moderation in argumentativeness appears to be about one-third as important as years of experience when determining upper-level promotions. Further, managers or mentors might incorporate an appreciation for moderation in argumentativeness into in-house training programs. Just as classroom teachers might plan simulation sessions for their classrooms, corporate trainers could develop simulations customized for their industries.

Conclusion

This study demonstrates that female supervisors at higher levels are moderate argumentatives, indicating that the relationship between argumentativeness and success is more complex than previously identified. While earlier studies have noted positive outcomes for high argumentatives in organizations, this study clarifies the outcomes for supervisory women. The data show that argumentativeness per se is not linearly related to supervisory level, but rather, moderation in argumentativeness is characteristic of women who have attained advanced supervisory positions. The lack of significance of these results for promotion from non-supervisor to supervisory level one requires further study. These results confirm that the ability to argue is not only culturally acceptable, but essential for supervisory women, up to a point. The results support the earlier notion that women who are low in argumentativeness would benefit from developing their argumentativeness. However, the data also show that encouraging argumentativeness in all women may be counter-productive. Rather, argumentativeness training to enhance women's workplace success would most profitably focus on low-argumentative females. Further, a promotion-related benefit should not be expected until after promotion to the first supervisory level.

NOTES

This article is based on her doctoral dissertation (Matthew Seeger, advisor). In addition to Dr. Seeger, the author would like to thank the anonymous reviewers, Dr. Stephen Schullery, and the editor for their helpful questions and suggestions.

Earlier drafts of various portions of this manuscript were presented at the 1996 Organizational Behavior and Training Conference, the 1996 Michigan Association of Speech Communication, the 1996 annual convention of the Association for Business Communication, and a Small Group Communication Workshop preceding the 1996 Annual Conference of the Speech Communication Association.

REFERENCES

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage.

Anderson, J., Schultz, B., & Staley, C. C. (1987, Fall). Training in argumentativeness: New hope for nonassertive women. Women's Studies in Communication, 10(2), 58-66.

Borisoff, D., & Victor, D. A. (1989). Conflict management: A communication skills approach. Englewood Cliffs, NJ: Prentice Hall.

Burgoon, M., Dillard, J. P., & Doran, N. E. (1983, Winter). Friendly or unfriendly persuasion: The effects of violations of expectations by males and females. Human Communication Research, 10(2), 283-294.

Burnett, R. E. (1991, November). Substantive conflict in a cooperative context: A way to improve the collaborative planning of workplace documents. Technical Communication, 38(4), 532-539.

Carli, L. L. (1990, November). Gender, language, and influence. Journal of Personality and Social Psychology, 59(5), 941-951.

Chaney, L. H., & Martin, J. S. (1995). Intercultural business communication. Englewood Cliffs, NJ: Prentice Hall Career and Technology.

Collingwood, H. (1995, November). Women as managers: Not just different - better. Working Woman, 14.

Donohue, W. A., Weider-Hatfield, D., Hamilton, M., & Diez, M. E. (1985, Spring). Relational distance in managing conflict. Human Communication Research, 11(3), 387-405.

Eisenberg, E. M., & Goodall, H. L., Jr. (1993). Organizational communication: Balancing creativity and constraint. New York: St. Martin's.

Ewald, H. R., & Burnett, R. E. (1997). Business communication. Upper Saddle River, NJ: Prentice Hall.

Fine, M. G. (1996, October). Cultural diversity in the workplace: The state of the field. The Journal of Business Communication, 33(4), 485-502.

Fowler, F. J. (1993). Survey research methods (2nd ed.). Newbury Park, CA: Sage.

Gardner, W. L., III, Peluchette, J. V. E., & Clinebell, S. K. (1994, November). Valuing women in management: An impression management perspective of gender diversity. Management Communication Quarterly, 8(2), 115-164.

Glass Ceiling Commission (1995, March). Good for business: Making full use of the nation's human capital: The environmental scan: A fact-finding report of the Federal Glass Ceiling Commission. Washington, DC: U.S. Government Printing Office.

Gorden, W. I., Infante, D. A., & Izzo, J. (1988, August). Variations in voice pertaining to dissatisfaction/satisfaction with subordinates. Management Communication Quarterly, 2(1), 6-22.

Haccoun, D. M., Haccoun, R. R., & Sallay, G. (1978, February). Sex differences in the appropriateness of supervisory styles: A non-management view. Journal of Applied Psychology, 63(1), 124-127.

Haglund, R. (1996, November 3). Few women at top of large state firms. Ypsilanti Press, p. B3.

Hermon, M. V. (1996, October). Building a shared understanding and commitment to managing diversity. The Journal of Business Communication, 33(4), 427-442.

Infante, D. A. (1981, Winter). Trait argumentativeness as a predictor of communicative behavior in situations requiring argument. Central States Speech Journal, 32(4), 265-272.

Infante, D. A. (1982, April). The argumentative student in the speech communication classroom: An investigation and implications. Communication Education, 31(2), 141-148.

Infante, D. A. (1985, Spring). Inducing women to be more argumentative: Source credibility effects. Journal of Applied Communication Research, 13(1), 33-44.

Infante, D. A. (1987a). Aggressiveness. In J. C. McCroskey & J. A. Daly (Eds.), Personality and Interpersonal Communication (pp. 157-192). Beverly Hills, CA: Sage.

Infante, D. A. (1987b, Fall). Enhancing the prediction of response to a communication situation from communication traits. Communication Quarterly, 35(4), 308-316.

Infante, D. A. (1989, Winter). Response to high argumentatives: Message and sex differences. The Southern Communication Journal, 54(2), 159-170.

Infante, D. A., & Gorden, W. I. (1985a, December). Benefits versus bias: An investigation of argumentativeness, gender, and organizational outcomes. Communication Research Reports, 2(1), 196-201.

Infante, D. A., & Gorden, W. I. (1985b, Fall). Superiors' argumentativeness and verbal aggressiveness as predictors of subordinates' satisfaction. Human Communication Research, 12(1), 117-125.

Infante, D. A., & Gorden, W. I. (1989, Spring). Argumentativeness and affirming communicator style as predictors of satisfaction/dissatisfaction with subordinates. Communication Quarterly, 37(2), 81-90.

Infante, D. A., & Rancer, A.S. (1982, February). A conceptualization and measure of argumentativeness. Journal of Personality Assessment, 46(1), 72-80.

Infante, D. A., & Rancer, A. S. (1993, Fall). Relations between argumentative motivation, and advocacy and refutation on controversial issues. Communication Quarterly, 41(4), 415-426.

Infante, D. A., & Rancer, A. S. (1996). Argumentativeness and verbal aggressiveness: A review of recent theory and research. Communication Yearbook, 19, 319351

Infante, D. A., Rancer, A. S., & Womack, D. E (1993). Building communication theory (2nd ed.). Prospect Heights, IL: Waveland.

Kachigan, S. K. (1982). Multivariate statistical analysis: A conceptual introduction. New York: Radius.

Kazoleas, D. (1993, September). The impact of argumentativeness on resistance to persuasion. Human Communication Research, 20(1), 118-137.

Kenton, S. B., & Valentine, D. (1997). CrossTalk: Communicating in a multicultural workplace. Upper Saddle River, NJ: Prentice Hall.

Locker, K. O. (1997). Business and administrative communication (4th ed.). Boston: Irwin/McGraw-Hill.

Pomerenke, P., Varner, I. I., & Mallar, S. (1996, December). The depiction of female and male professionals in business communication textbooks. Business Communication Quarterly, 59(4), 36-46.

Nicotera, A. M., & Rancer, A.S. (1994, Fall). The influence of sex on self-perceptions and social stereotyping of aggressive communication predispositions. Western Journal of Communication, 58(4), 283-307.

Onyekwere, E., Rubin, R. B., & Infante, D. A. (1991, Winter). Interpersonal perception and communication satisfaction as a function of argumentativeness and ego-involvement. Communication Quarterly, 39(1), 35-47.

Rancer, A. S., & Baukus, R. A. (1987). Discriminating males and females on belief structures about arguing. In L. B. Nadler, M. K. Nadler, & W. R. Todd-Mancillas (Eds.), Advances in gender and communication research (pp. 155-173). Lanham, MD: University Press of America.

Rancer, A. S., Baukus, R. A., & Infante, D. A. (1985, January). Relations between argumentativeness and belief structures about arguing. Communication Education, 34(1), 37-47.

Rancer, A. S., & Dierks-Stewart, K. J. (1985, February). The influence of sex and sex-role orientation on trait argumentativeness. Journal of Personality Assessment, 49(1), 69-70.

Rancer, A. S., Kosberg, R. L., & Baukus, R. A. (1992, October). Beliefs about arguing as predictors of trait argumentativeness: Implications for training in argument and conflict management. Communication Education, 41(4), 375-387.

Rancer, A. S., & Infante, D. A. (1985, Summer). Relations between motivation to argue and the argumentativeness of adversaries. Communication Quarterly, 33(3), 209-218.

Rich, A. R., & Schroeder, H. E. (1976, November). Research issues in assertiveness training. Psychological Bulletin, 83(6), 1081-1096.

Sauer, B. A. (1994, Summer). Sexual dynamics of the profession: Articulating the ecriture masculine of science and technology. Technical Communication Quarterly, 3(3), 309-323.

Schullery, N. (1997). Communication behaviors of employed females: A survey focusing on argumentativeness and women's supervisory roles in organizations. Unpublished doctoral dissertation, Wayne State University.

Schultz, B. (1982, Fall). Argumentativeness: Its effect in group decision making, and its role in leadership perception. Communication Quarterly, 30(4), 368375.

Schultz, B., & Anderson, J. (1984, August). Training in the management of conflict: A communication theory perspective. Small Group Behavior, 15(3), 333-348.

Shuter, R., & Turner, L. H. (1997, August). African American and European American women in the workplace: Perceptions of conflict. Management Communication Quarterly, 11(1), 74-96.

Tabachnick, B. G., & Fidell, L. S. (1996). Using multivariate statistics (3rd ed.). New York: HarperCollins.

U.S. Department of Commerce. (1980). Standard industrial occupational manual. Washington, DC: Author.

U.S. Department of Labor. (1996, September). Facts on working women: Twenty facts on women workers (U.S. Department of Labor Women's Bureau Publication No. 96-2). Chicago: Author.

Varner, I., & Beamer, L. (1995). Intercultural communication in the global workplace. Chicago: Irwin.

Wiley, M. G., & Eskilson, A. (1985, May). Speech style, gender stereotypes, and corporate success: What if women talk more like men? Sex Roles, 12(9/10), 993-1007.

Nancy M. Schullery (PhD, Wayne State University, 1997; MBA, Eastern Michigan University, 1982) is an Assistant Professor at the Haworth College of Business, Western Michigan University. She has 13 years of corporate and consulting experience in management, marketing, and public relations. Her research in argumentativeness grew out of her interest in the larger issues of conflict management, power and leadership, and small group communication.