Effects of Communication Direction on Job Performance and Satisfaction: A Moderated Regression Analysis.

High levels of job performance and job satisfaction occur when congruence of individual needs (growth need strength) and job characteristics (job scope) exists. Downward communication received some statistical support as both moderator and predictor of job performance and job satisfaction in low individual-job congruence situations. Upward and lateral communication had some support as predictors but lacked support as moderators of job performance and job satisfaction in high individual-job congruence situations. These conclusions are derived from this research that examines the moderating effects of communication direction on individual-job congruence and work outcomes (performance/satisfaction). Data from 302 employees who reported job scope, growth need strength, job performance, job satisfaction, and communication direction (upward, lateral, downward) were used for the study. Moderated regression analysis was the statistical technique applied.

Keywords: Direction of Communication, Job Characteristics Model, Job Performance. Job Satisfaction, Organizational Communication

The Job Characteristics Model (JCM) by Hackman and Lawler (1971) and Hackman and Oldham (1976, 1980) proposes that the relationship between job characteristics and work outcomes is moderated by employees' growth need strength (GNS), a measure of employees' desire to obtain growth satisfaction from their work. The JCM implies that when individuals' GNS is matched with job scope or job characteristics, higher performance and satisfaction will result (Hackman & Oldham, 1976; Pierce, Dunham, & Blackburn, 1979; Fried & Ferris, 1987; Bhuian, Al-Shammari, & Jefri, 1996). Although the individual-job congruence proposition or JCM model has received some support (Spector, 1985; Graen, Scandura, & Graen, 1986; Fried & Ferris, 1987), the evidence is inconsistent (Graen, Scandura, & Graen, 1986; Tiegs, Tetrick, & Fried, 1992; Johns, Xie, & Fang, 1992).

Such inconsistency has prompted some researchers (Oldham, Hackman, & Pearce, 1976; Clayton, 1981; Van der Vegt, Emans, & Van de Bliert, 1998) to investigate the moderating effects of other organizational variables on the JCM. One research effort (Pettit, Goris, & Vaught, 1997) investigated organizational communication as a moderator of the association between job performance and job satisfaction. But no study examines organizational communication as a moderator in the JCM. Thus, the purpose of this study is to investigate the moderating impact of communication direction on the Job Characteristics Model.

Rationale for Study

Desired work outcomes such as performance and satisfaction result, according to the JCM, when individuals experience three psychological states: (a) experienced meaningfulness--a job perceived as being important, valuable, and worthwhile; (b) experienced responsibility--a job perceived as providing autonomy; and (c) knowledge of results--a job perceived as providing feedback about how effectively the work is being performed (Robbins, 2001). Experienced meaningfulness, experienced responsibility, and knowledge of results are fostered by certain core characteristics of the job. The JOM proposes that experienced meaningfulness emerges from the following three job characteristics: task significance, task identity, and skill variety. On the other hand, job autonomy and job feedback are the job characteristics expected to produce the psychological states of experienced responsibility and knowledge of results, respectively.

Thus, task significance, task identity, skill variety, autonomy, and task feedback are expected to lead to the psychological states of experienced meaningfulness, experienced responsibility, and knowledge. These states hypothetically will lead to high levels of performance and satisfaction when they are matched with appropriate levels of growth need strength (Tiegs, Tetrick, & Fried, 1992). Using meta-analysis, a comprehensive review of nearly 200 studies of the JCM found support for the proposed relationships between job characteristics and both psychological and behavioral outcomes (Fried & Ferris, 1987).

Other studies, however, have shown results that are inconsistent with the JCM. For instance, a literature review conducted by Graen, Scandura, and Graen (1986) revealed that 10 of 21 tests concerning the association between individual-job congruence and job satisfaction were statistically insignificant. Of 13 studies of the relationship between individual-job congruence and job performance, only three showed significance. More recently, applying univariate and multivariate hierarchical moderated multiple regression analyses, Tiegs, Tetrick, and Fried (1992) found no support for GNS as a moderator of the JCM. This last finding appears consistent with the conclusion made by Graen, Scandura, and Graen (1986).

Because of the reported inconsistent results, new approaches for investigating the JCM are emerging. Adopting a curvilinear approach to the study of the model, Champoux (1992) reported that the results of his investigation pointed "... to adding a curvilinear effect of Job Scope to the basic Job Characteristics Theory" (p. 107). Also, Xie and Johns (1995) detected evidence of a U-shaped curvilinear relationship between job scope and employees' stress. Zeffane (1994) explored the relative effects of both task characteristics and demographics, sub-unit structure, technology, uncertainty, and work-group performance on job satisfaction and found that not only task-related factors but also other organizational constructs might affect job satisfaction.

Some authors (e.g., Porter, Lawler, & Hackman, 1975; Clayton, 1981; Brousseau, 1983) have suggested that an individual's reaction to his or her job may be influenced not only by the properties of the job and his or her needs but also by the work context or the organizational "milieu" surrounding the job. Van der Vegt, Emans, and Van de Bliert (1998) report that task and outcome interdependence of team members may influence the way the JOM functions. Other variables such as interpersonal relations, security, pay, and need for independence also have been shown to moderate the relationship between individual-job congruence and performance and satisfaction (Oldham, Hackman, & Pearce, 1976; Clayton, 1981). Consequently, research (Zedeck, 1971) has examined in what manner organizational and individual facets exert their influence and alter the relationship among target variables in the JCM.

The individual-job design-organizational structure congruence model theorized by Nemiroff and Ford (1976) and Porter, Lawler, and Hackman (1975) provides a rationale for expecting direction of communication--upward, downward, and lateral communication--to moderate the JCM. This model predicts high performance and high satisfaction when the individual has high growth needs, the job is enriched, and the organizational design is organic. Alternatively, it predicts that a congruent condition will be present when the individual has low growth needs, the job is simple, and the organizational design is mechanistic.

In mechanistic systems, according to Burns and Stalker (1961), the

interaction within management tends to be vertical. . . . Management, often visualized as the complex hierarchy familiar in organization charts, operates a simple control system, with information flowing up through a succession of filters, and decisions and instructions flowing downwards through a succession of amplifiers. (p. 7)

In organic structure, according to Burns and Stalker (1961), "Inter-communication between people of different ranks tends to resemble lateral consultation rather than vertical command" (p. 6). Information flow in mechanistic and organic systems has also been examined by other organizational theorists (e.g., Schuler, 1977; Huseman & Alexander, 1979). Specific dimensions of information flow found to characterize mechanistic and organic systems have been shown to correlate with job characteristics (Burns & Stalker, 1961; Woodward, 1965; Penley & Alexander, 1979; Pierce, Dunham, & Blackburn, 1979; James, 1981).

Burns and Stalker (1961) and Woodward (1965) found that written, vertical communication is more effective in mechanistic structures with unchanging, simple tasks, while both horizontal and vertical flows of communication are more effective in organic structures with changing, complex tasks. Penley and Alexander (1979) found that "as work group technology becomes less routine, emphasis shifts from downward to upward communication and finally to lateral communication" (p. 332). In a laboratory study, James (1981) investigated the effects of vertical and lateral task-related information flow on the JCM. He reported strong evidence to support the main effects of growth need strength and communication flow on measures of job satisfaction. Pierce, Dunham, and Blackburn's 1979 study of 397 employees of an insurance company found the highest levels of performance and satisfaction among individuals with strong growth needs who performed complex jobs within organic (i.e., open communication) organizational units. Thes e researchers concluded that "full effects of job design cannot be understood without knowledge of both the worker and the organization" (p. 239).

The organizational theory and findings described above correlate downward communication with mechanistic organizations and lateral communication with organic systems. Mechanistic organizations are in turn associated with low individual-job congruence situations while organic structures are associated with high individual-job congruence conditions, as posited by the JCM. Thus, it appears reasonable to expect direction of communication-upward, downward, and lateral--to moderate the JCM. A moderator variable is one that interacts with another variable to enhance the predictability of a second variable (Cohen & Cohen, 1975). This research proposes that direction of communication interacts with both job scope and growth need strength to predict job performance and job satisfaction, as shown in Figure 1. This investigation tests three hypotheses:

H1: Individuals' perception of job satisfaction and rated level of job performance will be higher under conditions of congruence (high growth need strength individuals in high scope jobs and low growth need strength individuals in low scope jobs) than under conditions of incongruence.

H2: Lateral and upward communication will moderate the relationship between the high individual-job congruence condition (high growth need strength individuals in high scope jobs) and both job performance and job satisfaction.

H3: Downward communication will moderate the relationship between the low individual-job congruence condition (low growth need strength individuals in low scope jobs) and both job performance and job satisfaction.

Methodology

Six hundred twenty-nine (629) employees from two companies, X and Y, constituted the original sample. Seventeen employees were excluded because of business trips, sickness, or vacations. Thus, 612 individuals (102 managers and 510 non-managers) constituted the sample of this research. Companies X and Y are in a large metropolitan area in the southwestern United States. Both firms are in the industrial sector of the economy and distribute products globally. Company X with 27 employees is the headquarters of a manufacturing firm that produces a comprehensive line of heavy duty premium quality lubricants. Company Y with 602 employees is a division of a multinational firm that deals with the research, design, engineering, and manufacturing of petroleum drilling products.

Three hundred and two workers (49.35 percent of the sample) completed and returned the research instruments of the study. There were 249 male respondents and 53 female respondents. Twenty (20) percent were less than 30 years old, 54 percent between 80 and 39, and 26 percent were over 39.

Measures

A subscale of the Job Diagnostic Survey as developed by Hackman and Oldham (1975, 1980) was used to determine the Motivational Potential Score (MPS) or job scope level as perceived by participants. The MPS reflects employees' perceptions of jobs in terms of variety, identity, significance, autonomy, and feedback. This is the most widely used perceptual measure of job scope. A seven-point response scale is used to determine job scope (1 = low, 7 = high).

This research used the job choice subscale of the Job Diagnostic Survey to measure growth need strength (Hackman & Oldham, 1975, 1980). The subscale provides an index of strength of higher order needs relative to lower order needs. It reflects employees' desire to obtain growth satisfactions from their work. This desire is determined by asking respondents to indicate their relative preference for pairs of hypothetical jobs, like, for example "A job where you are often required to make important decisions vs. A job with many pleasant people to work with" (Hackman & Oldham, 1975, p. 161). Internal consistency reliabilities for all Job Diagnostic Survey sections are reported to range from a high of .88 to a low of .56 (Hackman & Oldham, 1975).

This study obtained three measures of job performance: (a) quality of performance, (b) quantity of performance, and (c) overall job performance. They were obtained by having employees rate their performance and by having their supervisors rate them. Both sources used a seven-point scale with 7 being the highest and 1 the lowest. Specifically, participants were requested to respond to the following two questions:

1. How would you rate the quality of your own performance in your job?

2. How do you think your supervisor would rate the quality of your performance?

Two other questions asked about the quantity of performance. Supervisors were requested to assess the performance of their subordinates in terms of quality and quantity, independently.

Employees' self-evaluations were averaged with their evaluations from their supervisors to derive the specific measures of quality of performance and quantity of performance. Combining self-report and supervisor's scores may provide more objective performance measures (Wanous, 1974). The mean score of the quality and quantity of performance was calculated and used as a measure of overall job performance to provide a global measure of performance and simplify the analysis. Thus, the final measures of quality of performance, quantity of performance, and overall job performance provided specificity in this study for analyzing the data and interpreting relationships with other variables.

The performance data collected yielded a significant (p [less then] .01) relationship between the self rating of quality of performance and the supervisory rating of quality of performance (r = .32). A weaker but still significant (p [less than].01) relationship also was found between the two measures of quantity of performance (r = .19). After the integration took place, the correlation between the performance dimensions of quality and quantity (r = .62) was significant (p [less than] .01).

The Job Descriptive Index (JDI) provided the primary means for measuring job satisfaction (Smith, Kendall, & Hulin, 1969). The measure assesses satisfaction with five dimensions of a job: work, pay, promotions, supervision, and co-workers. The JDI is recognized as one of the most valid scales for measuring job satisfaction (O'Connor, Peters, & Cordon, 1978). An overall measure of job satisfaction was attained by summing up the Likert-type responses to two global, direct questions:

1. Generally speaking, how satisfied are you with your job?

2. How characteristic is this statement of you? Taking everything into account, I am very satisfied with my job.

Wanous (1974) reported a correlation of .76 between the two items.

The Roberts and O'Reilly's (1974, 1979) organizational communication instrument was used to measure direction of communication--upward, downward, and lateral. Roberts and O'Reilly found that their instrument had desirable psychometric properties: a test-retest reliability coefficient of .82 and a Cronbach alpha coefficient of .71. The instrument measures 15 different facets of communication. Roberts and O'Reilly (1974) suggest, however, that researchers select specific scales that meet the unique designs of their studies. This study uses the direction of communication, which includes nine questions (e.g., "While working, what percentage of the time do you spend in contact with others at the same job level?") to determine the amount of upward, downward, and lateral communication.

Data Collection

Data were collected with a pre-tested questionnaire incorporating the foregoing research measures. The researchers secured lists of the participants from the organizations and then distributed a numbered copy of the questionnaire to each employee listed. Individuals were told that the number on their questionnaires would be used for matching statistical data only and that no one else would see the number except the researchers. This number was important for later identifying each respondent's supervisor.

The internal mail system of each of the participating organizations was used to deliver the research instrument and introductory letter. Completed questionnaires were mailed to the Department of Management of an area university. Three hundred and sixteen (316) employees completed and returned questionnaires. Of this total, there were seven whose identification numbers had been cut or erased, and two others were improperly answered. Consequently, 307 questionnaires were usable (316 - 9). Supervisors then were asked to rate each respondent's performance.

Seventy-nine supervisors were identified using the number assigned to the returned questionnaires. Again, the internal mail system was used to deliver the performance scales and related materials to the supervisors. Supervisors returned their evaluations by mail directly to the university. During the two weeks taken by supervisors to assess employees' performance, one respondent resigned, while another was dismissed. Additionally, one supervisor for three individuals did not respond. Therefore, five additional cases (2 + 3) were dropped from the list of actual respondents, leaving a total of 302 participants, a response rate of 49.4 percent.

With a span of control ranging from two to eight subordinates, none of the 78 supervisors who participated in the assessment of respondents' performance accounted for a disproportionate amount of the data collected. No events--projects or activities--were detected that might have influenced participants' performance during the period between the time respondents completed the research instrument and the time supervisors conducted the performance assessment.

Statistical Techniques

For statistical analysis purposes, all research variables were reduced to 1-7 scales (1 = low; 7 = high). In order to assess Hypothesis 1, respondents were dichotomized into high and low groups based on their median scores on job scope and growth need strength. Individuals whose scores were equal to or greater than the medians of job scope and growth need strength, respectively, constituted the high-congruent group. The low-congruent group was formed by those whose scores were less than the medians. After the dichotomization took place, job performance and job satisfaction means were calculated for congruent and incongruent groups and their significance tested.

Hypotheses 2 and 3, which concerned the expected moderating influence of the direction of communication on the JCM, were assessed using moderated regression analyses. Following Zedeck (1971) and Cohen and Cohen (1975), moderated regression analyses were conducted by regressing the work outcomes, job performance and job satisfaction, on a linear combination of predictors, moderator, and predictors, moderator interactions. Three regression equations were developed for the test:

1. Y = a + bX

2. Y = a + bX + cZ

3. Y = a + bX + cZ + dXZ

where Y is the predicted value of the criterion; a is the least squares estimate of the intercept of the best fit equation; b is the least squares estimate of the population regression coefficient for X, the main predictor; c is the least squares estimate of the population regression coefficient for Z, the proposed moderator; and d is the least squares estimate of the population regression coefficient for the product term which carries information about the interaction between X and Z.

In the first equation, the dependent variable was regressed on the two predictors: growth need strength and job scope. In the second equation, communication direction was added. Finally, in the last equation, the dependent construct was regressed on the predictors, proposed moderator, and predictors-moderator interaction. Any moderating effects would be detected if the predictors-moderator interaction accounted for a significant amount of criterion variance after both predictors and moderator variables entered the regression equations. Zedeck (1971) states that "if Equations 2 and 3 are significantly different from Equation 1, but not from each other, then the variable [the suggested moderator, direction of communication] is an independent predictor and not a moderator variable" (p. 304). An F test was performed to determine the significance of variance in R2.

The nature of moderating and predicting effects detected was explained using differential predictability (Zedeck, 1971). Because this technique requires the dichotomization of respondents, they were divided into high and low groups based on their median scores on job scope and growth need strength. The significance and direction of the correlation of lateral and upward communication with job performance and job satisfaction for individuals in the high congruence group were assessed. Also, the significance and the direction of the correlation of downward communication with performance and satisfaction for individuals in the low congruence group were examined.

Measuring organizational and individual constructs by aggregating self reported data from research participants, as was done in this study, may raise questions. For several reasons, however, the researchers are confident that the measures of job performance, job satisfaction, and direction of communication are meaningful. First, as shown in the intercorrelations of the variables included in this study in Table 1, job performance/job satisfaction correlations ranged from .05 to .26 with the median and mean of these correlations both equal to .15. These results mirror previously reported findings (Brayfield & Crockett, 1955; Vroom, 1964; Srivastva, 1975; Fisher, 1980; Petty, McGee, & Cavender, 1984), suggesting that the variables assessed in this study are not just fallacious or spurious averages. Second, a graphical analysis of residuals yielded properties that suggest the residuals were independent, had zero mean, had a common variance, and followed a normal distribution. Finally, a review of the means, stan dard deviations, and medians, presented in Table 2, shows a variability in the way participants responded to the different scales. This variability also points to the normal distribution of the data collected and analyzed. Thus, these three reasons support the belief that the data represent correctly the constructs in this investigation.

Results and Discussion

Hypothesis 1 posits that congruence between growth need strength and job scope will increase job performance and job satisfaction. The extent of congruence should forecast the need for examining communication direction as a moderating variable in high and low congruence situations, the subjects of Hypotheses 2 and 3. All three hypotheses contain the work outcome variables of performance and satisfaction.

Relationship Between Congruence and Job Performance and Satisfaction

As Table 3 shows, the direction of the relationship in Hypothesis 1 is supported: congruence between growth need strength and job scope produced increased job performance and job satisfaction. (If there had been negative differences, the incongruent situation would have increased performance and satisfaction. Such is not the case.) Not all of the directional relationships tested significantly, however.

Two of the three mean differences of job performance proved to be significant at the .05 level (T = 2.25, T = 1.86). In addition, three of the six mean differences of job satisfaction showed significance--two at the .001 level (T = 2.73, T = 3.08) and one at .05 (T 2.27). Thus, job quality and overall performance combined with satisfaction with pay, promotion, and overall satisfaction are higher in congruence situations than in incongruent conditions.

These findings enhance the validity of the individual-job congruence approach of the Job Characteristics Model as reported in the literature (Hackman & Oldham, 1976, 1980; Spector, 1985; Graen, Scandura, & Graen, 1986; Fried & Ferris, 1987). Accordingly, managers would do well to match growth needs with job scope as a way to produce high levels of both performance and satisfaction.

The Effect of Upward and Lateral Communication

Hypothesis 2 was not supported. That is, upward and lateral communication did not receive support as moderators of performance and satisfaction in high congruent situations.

Models 1, 2, and 3 in Table 4 provide data for the three regression equations explained previously in the methodology section. It is best to read the figures in the table horizontally as each predictor dimension is added successively to the equations in different ways. In all three equations, the various aspects of job performance or job satisfaction, respectively, are the dependent variables. In Model 1, growth need strength and job scope are used as predictors of the dependent variable. In Model 2, the proposed communication moderator is added to the equation of Model 1. Finally, in Model 3, the interaction or product of the proposed communication moderator and the predictors, growth need strength and job scope, is added to the equation of Model 2. An F test was performed to determine the significance of variance ([delta]) in [R.sup.2]. As Sedeck (1971) states,

if Equations 2 and 3 are significantly different from Equation 1, but not from each other, then the variable [the suggested moderator, direction of communication] is an independent predictor and not a moderator variable (p. 304).

Sections 1 and 2 of Table 4 show the results regarding Hypothesis 2, which proposes that lateral and upward communication will moderate the relationship between the high individual-job congruence condition and both job performance and job satisfaction. When changes in [R.sup.2]. scores were compared as previously described, both lateral and upward communication were left without support as moderators (Model 3). However, they receive some support as main predictors (Model 2). Specifically, upward communication was found to be a main predictor of both satisfaction with pay ([delta][R.sup.2] = .102) and promotion ([delta][R.sup.2] = .053), while lateral communication was shown to be a predictor of both job quantity ([delta][R.sup.2] = .035) and overall job performance ([delta][R.sup.2] = .034).

The fact that upward and lateral communication rather weakly, if at all, enhance performance and satisfaction runs counter to ideas conceptualizing relationships about communication and work outcomes (Memeroff & Ford, 1976; Porter, Lawler, & Hackman, 1975; Burns & Stalker, 1961; Penley & Alexander, 1979; James, 1981). Put into a practical context, managers who use upward and lateral communication to attain performance and satisfaction work goals of employees matched in high growth need strength and job scope will likely not be successful. They should use other inducements such as pay, security, and interpersonal relations, which have been proven to enhance job outcomes in conditions of high congruence (Oldham, Hackman, & Pearce, 1976; Clayton, 1981), to motivate their employees instead.

The Effect of Downward Communication

Two of the three measures of job performance and one of the six measures of job satisfaction supported Hypothesis 3. (See Section 3 of Table 4.) Hypothesis 3 predicted that downward communication would moderate the relationship between the low individual-job congruence condition and both job performance and job satisfaction. Calculated changes of [R.sup.2] show downward communication acting as a moderator on the prediction of job quantity ([delta][R.sup.2] = .051), overall performance ([delta][R.sup.2] = .052), and overall satisfaction ([delta][R.sup.2] = .051). Downward communication also received support as a main predictor of satisfaction with work ([delta][R.sup.2] = .036), satisfaction with pay ([delta][R.sup.2] = .033), and satisfaction with co- workers ([delta][R.sup.2] = .083) in low individual-job congruence conditions.

Put into an organizational situation, managers my indeed influence employees' performance and satisfaction with communication, provided it is downward communication and workers are matched in appropriate job scope and growth need strength. Behavior of low growth need strength employees may be influenced by specific and concrete instructions provided by their supervisors when they are performing mechanistic, routine type of tasks.

Statistically insignificant changes in [R.sup.2] as well as significant changes regarding Hypotheses 2 and 3 are reported in Table 4 to insure the integrity of the research and to provide completeness and objectivity.

Aguinis and Pierce (1999) present a way to interpret the lack of definite support for direction of communication as a moderator, shown in this research. They suggest that, despite the "pervasive use of moderated regression analysis in articles published in three of the most influential journals in organizational science," its low statistical power may be "a possible explanation for a lack of support for hypothesized moderating effects" (p. 4). Aguinis (1955). defines low statistical power as "the probability of rejecting a false null hypothesis of no moderating effects. If power is low, Type II statistical error rates are high and, thus, researchers may erroneously dismiss theoretical models that include moderating effects" (p. 1142). More specifically, Aguinis and Pierce (1999) state that "low power affects substantive research conclusions in that a researcher may incorrectly conclude that the data in hand do not support a hypothesized moderating effect" (p. 2). Thus, the lack of definite support for direct ion of communication as a moderator, as analyzed in this study, may not be conclusive. Aguinis (1995) and Aguinis and Pierce (1999) also discuss some alternatives to deal with the low statistical power of moderated regression analysis that future similar research efforts may take into consideration.

Another way to provide insight into the analyses of the results regarding Hypotheses 2 and is to look at the correlation ratios of communication direction with job performance and job satisfaction, as shown in Table 5. Twenty-three (23) of the 27 ratios are negative while only 4 are positive. In addition, 5 of the ratios show degrees of significance, but all of the 5 relate to the interaction of upward and downward communication with job satisfaction. No significant relationships were found for lateral communication or for upward and downward communication with job performance.

If Table 4 is juxtaposed with Table 5, predictors and moderators can be compared with correlation ratios. Both of the predictors for upward communication, satisfaction with pay and promotions, showed significance in correlation, pay at .01 and promotions at .05. None of the prodictors for lateral communication showed significance in correlation analysis, however. And all three of the predictors for downward communication tested significant in correlation--work at .05, pay at .10, and coworkers at .01. Yet no moderator of the three communication dimensions proved to be a significant correlate. All of the significant correlations are negative and those results need thorough explanation.

The negative correlations in Table 5 mean that high levels of upward and downward communication may restrain job performance and job satisfaction. Certainly, this finding counters the theoretical and empirical work about moderators and work outcomes in the literature (Memeroff & Ford, 1976; Porter, Lawler, & Hackman, 1975; Burns & Stalker, 1961; Penley & Alexander, 1979; James, 1981). However, it might be explained more adequately in terms of a curvilinear effect and decision making under conditions of contingency as suggested by findings reported by Champoux (1992) and Xie and Johns (1995).

If a curvilinear relationship exists, the pattern between the dependent variables (job performance and job satisfaction) and the independent variables (upward, lateral and downward communication) may resemble an inverted U-shaped curve with elements of positive, zero, and negative correlation. That is, the communication direction dimensions may increase work outcomes positively up to a point. After that point, further communication saturation may decrease performance and satisfaction. Such a variable effect can be seen as a situation of management contingency-of isolating key variables and determining the proper mix in specific circumstances.

The possible presence of curvilinear effects may have even more specific practical implications as well. High levels of upward communication might cause employees to believe that their superiors do not know how to perform their jobs. If this perception emerges, employees might think they are being exploited when required to transmit more and more information upward. Or they may feel overly controlled by their bosses. In addition, excessive levels of lateral communication may generate uncomfortable feelings among employees when they have the possibility of comparing themselves with their peers in terms of resource distribution, reward administration, management policy application, quota assignments, and such. Perceived inequities can negatively affect employees' performance and satisfaction.

Likewise, too much downward communication from superiors might create feelings of dependency that employees may resist. Situations such as these demonstrate the practical implications of situations that can occur when direction of communication becomes detrimental at higher levels than necessary. To be sure, communication flow that becomes excessive creates problems among employees. Xie and Johns (1995) suggested that too little or too much of good thing may provoke negative reactions. But only specific research to understand the possible curvilinear impact of communication on performance and satisfaction can provide answers. Such research needs to be done.

Conclusions and Implications

Three conclusions emerge from this study:

1. High levels of both job performance and job satisfaction often result from a match between individual needs for growth (GNS) and job characteristics (job scope).

2. Upward and lateral communication do not usually affect job performance and job satisfaction when high individual-job congruence exists.

3. Downward communication sometimes affects job performance more than job satisfaction when low levels of individual-job congruence exists.

While these conclusions varied by degree and by circumstance, the findings suggest patterns that might need to be explored further by business communication scientists.

The design of this research used total communication (oral, written, graphic, nonverbal, electronic, etc.) as the way to measure communication. When other topics are included within the research domain of this study, selected implications for further investigations emerge:

1. Would the results have been the same had written communication been singled out? Or oral? Or a combination of written and oral?

2. Would message design strategy (direct/indirect) have produced similar findings? What effect would it have had?

3. Would electronically mediated communication have affected the study? How?

4. How would principles of effect (persuasion) in message design (you viewpoint, positive emphasis, conversational tone, adaptation, and such) affect this study?

5. What length and types of reports (oral or written) should be used in upward, lateral, and downward communication?

Questions such as these for future research are limited only by the imagination; yet they can be focused and honed through researcher interest and appetite. Their expression and scientific investigation, are sorely needed as a new century begins.

The original version of this paper was presented at the Southwest ABC meeting in Dallas, Texas, on March 5, 1998. The paper received the Irwin/McGraw-Hill Distinguished Paper Award in the business communication division of the Southwestern Federation of Administrative Disciplines (SWFAD) and appeared in the Proceedings.

Jose R. Goris is Professor of Management in the School of Business at Andrews University, Michigan. He received his B.A. from Antillean Adventist University, Puerto Rico; his M.A. from Loma Linda University, California; his M.B.A. from the Inter American University of Puerto Rico; and his Ph.D. in organization theory from the University of North Texas.

Bobby C. Vaught is Professor of Management at Southwest Missouri State University. He received his B.S.I.M. from the University of Arkansas and both his M.B.A. and Ph.D. from the University of North Texas.

John D. Pettit, Jr., a contributor to the management and communication literature for over 30 years, is Professor and Interim Head in the Department of Information and Decision Sciences at the University of Texas at El Paso. He earned his undergraduate and master's degrees at University of North Texas and his Ph.D. in management at Louisiana State University.

REFERENCES

Aguinis, H. (1995). Statistical power problems with moderated multiple regression in management research. Journal of Management, 21, 1141-1158.

Aguinis, H., & Pierce, C.A. (Summer, 1999). Improving the power of moderated multiple regression to estimate interaction effects. Research Methods Forum, Vol. 4, Academy of Management, Research Methods Division.

Bhuian, S., Al-Shammari, E., & Jefri, O. (1996). Organizational commitment, job satisfaction and job characteristics: An empirical study of expatriates in Saudi Arabia. International Journal of Commerce & Management, 6, 57-80.

Brayfield, A. H., & Crockett, W. H. (1955). Employee attitudes and employee performance. Psychological Bulletin, 52, 396-424.

Brousseau, K. R. (1983). Toward a dynamic model of job-person relationships: Findings, research questions, and implications for work system design. Academy of Management Review, 8, 33-45.

Burns, T., & Stalker, G. M. (1961). The management of innovation. London: Tavistock.

Clayton, S. H. (1981). Moderators of the relationship between individual-task-structure congruencies and job satisfaction and performance. Unpublished Doctoral Dissertation, The University of Texas at Dallas.

Champoux, J. E. (1992). A multivariate analysis of curvilinear relationships among job scope, work context satisfactions, and affective outcomes. Human Relations, 45, 87-111.

Champoux, J. E., & Peters, W. S. (1987). Form, effect, size and power in moderated regression analysis. Journal of Occupational Psychology, 60, 243-255.

Cohen, J., & Cohen, P. (1975). Applied multiple regression/correlation analysis for the behavioral sciences. New York: John Wiley.

Fisher, C. D. (1980). On the dubious wisdom of expecting job satisfaction to correlate with performance. The Academy of Management Review, 5, 607-612.

Fried, Y., & Ferris, G. R. (1987). The validity of the job characteristics model: A review and meta-analysis. Personnel Psychology, 40, 287-322.

Graen, G. B., Scandura, T. A., & Graen, M. R. (1986). A field experimental test of the moderating effects of growth need strength on productivity. Journal of Applied Psychology, 71, 484-491.

Hackman, J. R., & Lawler, E. E., III (1971). Employee reactions to job characteristics. Journal of Applied Psychology, 55, 259-283.

Hackman, J. R., & Oldham, G. R. (1976). Motivation through the design of work: Test of a theory. Organizational Behavior and Human Decision Processes, 16, 250-279

Hackman, J. R., & Oldham, G. R. (1975). Development of the job diagnostic survey. Journal of Applied Psychology, 60, 159-170.

Hackman, J. R., & Oldham, G. R. (1980). Work redesign. Reading, MA: Addison-Wesley.

Hartmann, F. H., & Moers, F. (1999). Testing contingency hypotheses in budget. ary research: An evaluation of the use of moderated regression analysis. Accounting Organizations, & Society, 24, 291-315.

Huseman, R. C., & Alexander, E. R. (1979). Communication and the managerial function: A contingency approach. In Huseman, R. C. & A. B. Carroll (Eds.), An integrative approach to organizational behavior. New York: Allyn and Bacon.

James, E. (1981). Communication structure, task complexity and individual need congruence. Unpublished Doctoral Dissertation, The Ohio State University, Ohio.

Johns, G., Xie, J. L., & Fang, Y. (1992). Mediation and moderation effects in job design. Journal of Management 18, 657-676.

Ku, L. (1996). Social and nonsocial uses of electronic messaging systems in organizations. The Journal of Business Communication, 33, 297-32 5.

Nemiroff, P. M., & Ford, D. L. (1976). Task effectiveness and human fulfillment in organizations: A review and development of a conceptual contingency model. The Academy of Management Review, 1, 69-82.

O'Connor, E. J., Peters, L. H., & Gordon, S. M. (1978). The measures of job satisfaction: Current and future considerations. Journal of Management, 4, 17-26.

Oldham, G. R., Hackman, J. R., & Pearce, J. L. (1976). Conditions under which employees respond positively to enriched work. Journal of Applied Psychology, 61, 395-403.

Penley, L. E., & Alexander, E. R. (1979). The communication and structure of organizational work groups: A contingency perspective. Academy of management proceedings. Atlanta, Georgia.

Pettit, J. D., Jr., Goris, J. R., & Vaught, B. C. (1997). An examination of organizational communication as a moderator of the relationship between job performance and job satisfaction. The Journal of Business Communication, 34, 81-98.

Petty; M. M., McGee, G. W., & Cavender, J. W. (1984). A meta-analysis of the relationship between individual job satisfaction and individual performance. The Academy of Management Review, 9, 712-721.

Pierce, J. L., Dunham, R. B., & Blackburn, R. S. (1979). Social systems structure, job design, and growth need strength: A test of a congruency model. The Academy of Management Journal, 22, 223-240.

Porter, L. W., Lawler, E. E., III, & Hackman, J. R. (1975). Behavior in organizations. New York: McGraw-Hill.

Robbins, S. P. (2001). Organizational ehavior. New Jersey: Prentice-Hall.

Roberts, K. H. & O'Reilly, C.,A. (1974). Measuring organizational communication. Journal of Applied Psychology, 59, 321-326.

Roberts, K. H., & O'Reilly, C.,A. (1979). Some correlations of communication roles in organizations. Academy of Management Journal, 22, 42-57.

Schuler, R. S. (77). Role conflict and ambiguity as a function of the task-structure-technology interaction. Organizational Behavior and Human Performance, 20, 66-74.

Smith, P. C., Kendall, L. M., & Hulin, C. L. (1969). The measurement of satisfaction in the work and retirement. Chicago: Rand McNally.

Spector, P. E. (1985). Higher-order need strength as a moderator of the job scope-employee outcome relationship: A meta-analysis. Journal of Occupational Psychology, 58, 119-127.

Srivastva, S. (1975). Job satisfaction and productivity. Cleveland: Department of Organizational Behavior, Case Western Reserve University.

Tiegs, R. B., Tetrick, L. E., & Fried, Y. (1992). Growth need strength and context satisfactions as moderators of the relations of the job characteristics model. Journal of Management, 18, 575-593.

Van der Vegt, G., Emans, B., & Van de Bliert, E. (1998). Motivating effects of task and outcome interdependence in work teams. Group & Organization Management, 23, 124-143.

Vroom, V. H. (1964). Work and motivation. New York: John Wiley.

Wanous, J. P. (1974). A causal-correlational analysis of the job satisfaction and job performance relationship. Journal of Applied Psychology, 59, 139-144.

Woodward, J. (1965). Industrial organizations: Theory and practice. London: Oxford University Press.

Xie, J. L., & Johns, G. (1995). Job scope and stress: Can job scope be too high? Academy of Management Journal, 38, 1288-1309.

Zanzi, A. (1987). How organic is your organization? Determinants of organic/ mechanistic tendencies in a public accounting firm. Journal of Management Studies, 24, 125-142.

Zedeck, S. (1971). Problems with the use of "moderator" variables. Psychological Bulletin, 76, 295-310.

Zeffane, R. M. (1994). Correlates of job satisfaction and their implication for work redesign: A focus on the Australian telecommunications industry. Public Personnel Management, 23, 61-76.