Information Load: A Test of an Inverted-U Hypothesis with Hourly and Salaried Employees

information Load: A Test of an Inverted4 Hypothesis with Hourly and Salaried Employees

RODGER W. GRIFFETH’ Department of Management

George Mason University

KERRY D. CARSON Department of Management

The University of Southwestern Louisiana

DANIEL B. MARIN Department of Management Louisiana State University

We surveyed 714 hourly and 516 salaried employees of a midwestern telephone company to test the effects of information load on work-related outcomes. Using curvilinear regression analyses, we found support for our hypothesis that employees are less satisfied with outcomes as the load of information deviates positively or negatively from some level. We also predicted, and found, that this quadratic function was more prevalent in the hourlygroup than in the salaried group. Implications, future research directions, and limitations of the present study are discussed.

Weick (1987) suggests that communication is the essence of organization. However, the processing of information in organizations has become increasingly difficult with the continuing explosion in communication technology (Fulk & Boyd, 1991). For more than a decade, popular textbooks of organizational behavior have recognized information overload as a barrier to effective communication in organizations (Arnold & Feldman, 1986; Ivancevich & Matteson, 1987; Katz & Kahn, 1978; Lawler & Rhode, 1976; Rogers & Agarwala-Rogers, 1976; Steers, 1984; Vecchio, 1989). Unfortunately, little empirical research conducted within organizational settings examine effects of information load on individuals’ personal and work outcomes (O’Reilly, 1980). This study will examine outcomes important to employee effectiveness in organizations. These outcome variables include job satisfaction, communication satisfaction, job pressure, and accuracy of information as well as perceptions of teamwork and supervisor support.

‘Correspondence concerning this article should be addressed to Rodger W. Griffeth, Depart- ment of Management, George Mason University, Fairfax, VA 22030.

763

Journal of Applied Social Psychology, 1992,22, 10, pp. 763-779. Copyright @ 1992 by V. H. Winston &Son, Inc. All rights reserved.

764 GRIFFETH, CARSON, AND MARlN

Empirical Research on Information Load

Early studies of information load focused on its relation to individual job performance. Some posited a positive linear relationship between amount of information and performance (e.g., Manis, Fichman, & Platt, 1978) whereas others argued for more complex relationships such as a curvilinear relationship between information load and performance (e.g., Connolly, 1988). These researchers reported that a reduction in the information load increased decision making quality but decreased confidence level and satisfaction (Cher- vany & Dickson, 1974; Jacoby, Speller, & Kohn, 1974), suggesting an opti- mum load of information vis-a-vis performance. Huber, O’Connell, and Cummings (1975) hypothesized a curvilinear relationship between perceived uncertainty and information load but the relationship was not supported in this laboratory study.

Muchinsky’s (1977) exploratory study found information overload unrelated to organizational climate. Others, however, argue that information load is significantly related to work outcomes. For example, O’Reilly (1980) hypothesized a complex relationship: “. . . past some optimal point too much information can lead to decreased decision making performance but increased feelings of satisfaction” (p. 692). He found that the relationship between information load and job satisfaction is linear while that between information load and performance was curvilinear. Employees were unable to judge the optimal amount of information they could process and were more satisfied with some information overload. Boynton (1988) and Penley (1982) found both linear (positive) and curvilinear (inverted-U) relationships between information load constructs and organizational outcome variables.

Penley (1982) and Alexander, Helms, and Curran (1987) examined a construct analogous to load: information adequacy. Defined as the discrepancy between information requirements and processing capacity, information received is adequate if it meets the information needed. Information is inadequate if needs exceed that received; abundant if that received exceeds needs. Generally, inadequate information was associated with lower organizational identification, commitment, task scope, and lower organizational levels. Abundant information was associated with higher responses on these variables. Alexander et al.’s (1987) results led them to conclude that more information is not always better.

Although information adequacy was positively related to the outcome variables, both Penley (1982) and Alexander et al. (1987) used discrepancy scores with their inherent problems (see Johns, 1981, for a review and critique), and neither directly tested for nonlinear relationships using power

INFORMATION LOAD 765

polynomials, a requirement needed to detect significant curvilinear relationships (Cohen & Cohen, 1983; Pedhazur, 1982). One purpose of the present study is to use curvilinear regression to investigate the nature of the relationship between information load and organizational outcomes.

Related Constructs in Organizational Theory

The inverted-U relationship can be postulated from complex constructs similar to information load from theories about how managers and organizations process information. For example, Weick (1979) developed the concept of information equivocality to explain how managers interpret ambiguous cues from the environment. When confronted with ambiguity, managers seek consensual validation through discussion to arrive at a common interpretation and frame of reference. This concept of information equivocality relates to information load in that too much ambiguous information can. lead to uncertainty and inaction (Daft & Lengel, 1984). Thus, as information equivocality deviates positively or negatively from a particular level, organizational uncertainty increases and performance decreases.

Also related to the issue of information load is the concept of richness, or “the potential information-carrying capacity of data”( Daft & Lengel, 1984, p. 196). Daft and Lengel(l984, 1986) proposed a five-step continuum of communication media ranging from face-to-face, highest in information richness, to numeric, formal (i.e., computer output), lowest in richness. Their central thesis is that richer information media are preferred for equivocality reduction. The more complex the situation, the richer the media required to reduce equivocality. However, just as information load may exceed needs, so can information richness be too great for the level of organizational complexity. Simple organization phenomena usually require information that is not highly rich. Information richness exceeding the corresponding level of organizational complexity can be detrimental, causing lowered personal and work outcomes.

Hypotheses Development

The preceding review of empirical research and theory raises the possibility that too much information, among other things (e.g., message structure, topic, etc.), can produce uncertainty at both the individual and organizational levels. This means that past some point, more information increases, rather than reduces uncertainty. Further, this suggests that the relationships between information load and individual and organizational outcomes are curvilinear rather than linear. However, none of the organizational research cited above


has empirically tested for the presence of a curvilinear relationship using the appropriate methodology: power polynomials (Cohen & Cohen, 1983; Ped- hazur, 1982). Our primary objective is to use this methodology to examine this notion. Thus,

Hypothesis I : As information load deviates positively or negatively from some range (Alexander et al., 1987), employees will be less satisfied with their jobs, communication in general, and will report higher job pressure with fewer favorable perceptions of supervisor support, teamwork, and accuracy of information.

Most of the previous field studies on communication used small samples of salaried employees-managers and professionals-thereby excluding the bulk of the labor force-hourly employees (Porter & Roberts, 1976). We investigated the presence or absence of quadratic relationships in both of these groups. We suggest that quadratic relationships will be more prevalent in the hourly sample, while linear relationships more prevalent in the salaried group. Our rationale for this distinction originates from two sources. First, from suggestions made by Goldhaber, Dennis, Richetto, and Who (1984) that whereas employees generally were less satisfied receiving than sending information, hourly employees have less opportunity than salaried employees to send information. Also, hourly employees are less accustomed than salaried employees to processing large quantities of information. Hourly employees are less often required to make decisions affecting more than the immediate area of their jobs and might, therefore, feel less in control of their work environments. Indeed, Spector (1982) suggested that it was precisely an internal locus of control that suited managers and professionals (i.e., salaried workers) to their jobs and that members of this group characteristically sought out and used facts for complex information processing in order to control their environments. Thus, we expect the results of our test for quadratic relationships to be more prevalent with the hourly than the salaried group.

Second, this prediction is consistent with current thinking on managerial information processing. Managers, and others with diverse tasks, use a variety of inputs to resolve the uncertainty in their envirorunents (Daft & Wiginton, 1979). The higher in the hierarchy, the more likely it is that the manager encounters equivocal, subjective information (Daft & Lengel, 1984; Daft & Weick, 1984). Similarly, other salaried employees with high task variety may require large amounts of information. Therefore, we expect our salaried group, needing more information, to less frequently experience dissatisfaction as a result of abundant information (Daft & Macintosh, 1981). In contrast, employees involved in tasks with little variety need less information. At their


organizational levels, rules, regulations, and clear-cut goals and objectives define the jobs and require less interpretation (Daft & Lengel, 1984). There- fore, compared to salaried employees, we expect hourly employees, generally needing less information, to become dissatisfied with the job and communications in general, and perceive a less favorable working environment when they received more information than they need. Two alternative hypotheses are possible for the hourly and salaried employees: First,

Hypothesis 2: Quadratic relationships between information load and personal and work outcomes will be more prevalent with hourly than salaried employees.

Alternatively, salaried employees require and receive more information but, perhaps, are better able to process it; or they possess more relevant training about processing information. Hourly employees require and receive less information but may have more limited processing capabilities, or less training opportunities. Driver and Streufert (1969) suggest that there is not just one curvilinear pattern with information load, but a “whole family of such curves” (p. 276). These researchers found that both subjects with higher conceptual structure and subjects with lower conceptual structure displayed an inverted U-shaped curve. However, mean score outcomes for the higher conceptual group were higher than mean score outcomes for the lower conceptual group (Streufert & Driver, 1965). Thus,

Hypothesis 3: Both groups have quadratic relationships between information load and the personal and work outcomes, but the mean load will be higher for the salaried group than for the hourly group.

Method

Sample and Procedures

Surveys were administered to small groups (5-10) of employees on company time at 10 statewide locations of a large decentralized midwestern telephone company. The respondents noted whether they were classified as hourly or salaried employees. They also indicated in which department they worked (support services, central office, marketing, engineering and construc- tion, PBX/ I&R, commercial, accounting traffic systems, data processing, and data/ support services). Participation was anonymous and voluntary, and confidentiality of individual responses was guaranteed. Employees placed their completed surveys in envelopes, which were then mailed directly to the


first author’s business address. Approximately 1300 surveys were administered. Employees returned 1259, of which 1230 presented usable data, yielding a final response rate of 95%. Fifty-eight percent of the sample classified themselves as hourly (n = 714) and 42% as salaried (n = 516). Fifty-three percent of the sample was male.

Measures

Job satisfaction. We used the four-item Hoppock (1935) measure of general job satisfaction. McNichols, Stahl, and Manley (1978) found that this measure performs well in terms of construct, convergent, and concurrent validities and reliability.

Perceptions of the work environment. We used a 5-point ( 1 = strongly disagree, 5 = strongly agree) Likert scale to assess perceptions of the work environment containing 13 items adapted from Litwin and Stringer’s (1968) organizational climate measure (Form B). Because others (Sims & LaFollette, 1975) reported low reliability and validity with this measure, the 13 items were analyzed using principal component factor analysis with varimax rotation. Examination of the eigenvalue pattern and scree test supported three factors, which together accounted for 65% of the explained variance. Factor I , defined by seven items (e.g., “My boss encourages me to develop my ideas”) with substantial loadings (> .59), explained 44.3% of the variance and we labeled it “Supervisor Support.” Factor 2, accounting for 12.5% of the extracted variance, had five items (e.g., “There is a feeling of teamwork at this company”) with high loadings (> .59). We designated this factor “Teamwork.” Factor 3, accounting for 7.9% of the explained variance, consisted of only one item (“The work is a series of deadlines and tight schedules”) which loaded .95 on the third factor. We labeled it “Pressure from the Job.”(Note: This variable was reversed scored to be consistent with the inverted-U hypothesis.)

Satisfaction with communication. We measured satisfaction with communication with two 7-point items developed by Roberts and O’Reilly (1974b). One item measured the degree of satisfaction with communication in the organization ( 1 = not at all satisfied; 7 =very satisfied). The second measured the reception of information on a timely basis (1 = almost never; 7 = almost always). The two items were correlated (r = .62,p < .05) and were averaged.

Accuracy of information. We used a four-item, 7-point measure adapted from Roberts and O’Reilly (1974b) assessing the accuracy of information received from the department head, supervisor, subordinates, and peers ( 1 = completely inaccurate, 7 = completely accurate). The four items were averaged to represent a composite measure of this construct.

Information load. Information load was a one-item, 7-point measure adapted from Roberts and O’Reilly (1974b) assessing the frequency with


which employees perceive they have more information than they can effi- ciently use (1 = almost never, 7 = almost always). This item, used by other researchers (e.g., Muchinsky, 1977; Penley & Hawkins, 1985), measures the employee’s perception of information abundance. It captures the individual’s reaction to the amount of information received as well as the perceived ability to process this information.

Analysis

To test for the presence of nonlinear relationships (e.g., the inverted U hypothesis) described in the hypotheses, we used hierarchical regression procedures (Cohen & Cohen, 1983; Pedhazur, 1982). First, we regressed each dependent variable on the load variable. Then, we added the quadratic component of the load variable to this two-variable model, thus creating a second degree polynomial (Pedhazur, 1982). Finally, by calculating the F test for the change in R2, we were able to determine whether the quadratic relationship was significant. Since our primary goal was understanding the nature the relationships between load and our dependent variables, we also examined significant curvilinear trends. Hypothesis 2, involving the preval- ence of quadratic relationships in the different employment classifications, was tested by computing a z-score and Pearson X 2 statistic comparing the two proportions resulting from the ratio of the number of significant polynomial relationships to the total number of relationships examined for each employment classification. Hypothesis 3, that both groups have quadratic relationships but the mean load will be higher for the salaried group than for the hourly group, was examined by the test of Hypothesis 2, and by calculating a t-test between the two groups’ level of information load.

Results

Table 1 presents the correlation matrix and coefficient alphas, where possible, for study measures. Inspection of this table shows that the reliability coefficients (in parentheses) are greater than or equal to the range considered acceptable for basic research (Nunnally, 1978). Further inspection shows the independent variable, information load, was significantly, but not highly related to the dependent variables job satisfaction (r = .15, p I .05), supervisor support ( r = .21, p I .05), teamwork (r = .23, p I .05), satisfaction with communication (r = .22, p I .05), and accuracy of information from all sources (r = .15, p I .05). Information load was unrelated to pressure from the job (r = .05, ns). Although the bivariate correlations indicate support for the linear hypothesis that higher loads of information are associated with higher work-related outcomes (e.g., job satisfaction, supervisor support,


Table 1

Intercorrelation Matrix Among Variables

Variables 1 2 3 4 5 6 7

I . Job Satisfaction (34)

2. Pressure from job p.03

3. Supervisor support .37* -.02 (.91)

4. Teamwork .50* .01 .53* (.80)

5. Satisfaction with communications .37* -.04 .44* .52* (.75)

6. Accuracy of information from all sources .25* -.03 .43* .42* .45* (.70)

7. Information load .15* .05 .21* .23* .22* .15* -

Note. Values in parentheses are coefficient alpha reliabilities. *p I .05. n = 1228.

etc.). a rigorous test of the inverted-U hypothesis can only be determined using curvilinear regression. Table 2 presents both the linear and the quadratic regression results f o r each of the two samples of hourly and salaried personnel, testing Hypotheses 1 and 2.

Table 2

Linear and Quadratic Regression Results for Hourly and Salaried Personnel

Hourly Salaried Variables

R2 AR2 FJdf R2 AR2 Fldf

Job satisfaction .025a [.Ol 4.7/688Ib .020a .OO I 1.0/509

Pressure from the job .OOO [.01 8.6/672Ib . O l O .OO 1.9/502

Supervisor support .040a [.01 6.6/690Ib .040a [.01 3.9/509Ib

Teamwork .058a .OO 1.3/690 .036a .OO 2.9/509

Satisfaction with communications .036 [.01 4.9/685Ib .058a .OO 5 1.0/507

Accuracy of information from all sources .026a .OO 5 1.0/678 .014a .OO 1.1/508

aLinear equation is significant (p I .05). significant (p I .05).

bF, AR2 for the quadratic equation is


Hourly Personnel

The first column of this table presents the results of the linear regression analyses conducted on all six of the dependent variables. Consistent with the bivariate relationships, five of the six dependent variables were significantly ( p I .05) predicted by the independent variable, information load. Only pressure from the job was unrelated to information load. Columns two and three of the table present the results of the curvilinear regression analyses on the dependent variables. From these columns of the table it is apparent that four of the six variables (job satisfaction, pressure from the job, supervisor support, and satisfaction with communication) were better explained by the quadratic version of information load. Addition of the quadratic term increased the explained variance by one percent for each variable.

Generally, the results presented in Table 2 support the curvilinear hypothesis for hourly personnel. Specifically as load deviates from some level, there exists a decrease in job and communication satisfaction, and perceptions of supervisory support and pressure from the job.

Salaried Personnel

The fourth column of Table 1 presents the results of the linear regression analyses conducted for the salaried group. As above, five of the six dependent variables were significantly (p I .05) predicted by the independent variable. Like the hourly group, information load was not significantly related to pressure from the job. However, the fifth and sixth columns of Table 1 show that, unlike the hourly group, addition of the quadratic term increased the amount of explained variance for only one variable, supervisor support.

It is readily apparent that with four significant quadratic relationships for the hourly group and only one for the salaried group, information overload is more prevalent with the former than the latter. However, to confirm this, we computed a z-score comparing the two proportions which resulted from the ratio of the number of polynomial relationships which were significant to the total number of relationships examined in each employment classification (i.e., hourly [4/6] vs. salaried [1/6]).

Alternatively, we used the Pearson chi-square statistic (Berenson, Levine, & Goldstein, 1983) comparing the two proportions. The results of both tests weresignificant(z= 17.5,p< .05,andX2=3.1,p< .lo) ,andsuggest that the polynomial relationships were more prevalent in the former than the latter group. Thus, as predicted, the results presented in Table 2 indicate support for Hypothesis 2 that the linear model better explains the relationships for salaried employees while the quadratic model better explains the relationships for hourly employees.


The data testing Hypothesis 2 were also relevant for Hypothesis 3, that both groups have quadratic relationships between the variables of interest. Obviously, both groups do not have quadratic relationships. However, to test the possibility that the mean load will be higher for the salaried group than for the hourly group, a t-test was undertaken to compare the amount of information load between the groups. The results indicate that salaried ( M = 3.14) employees reported significantly ( t ( 1204) = 4.0, p < .05) higher information load than hourly employees ( M = 2.75).

Discussion

We hypothesized that as the amount of information deviated from some level, individual employees would be increasingly dissatisfied with their jobs, communication, and their work environment. We also predicted that the quadratic function would be more prevalent in the hourly than in the salaried group. The results support both of these hypotheses. As predicted, a quadratic function described the relationships of four of six dependent variables, for the hourly group regarding job satisfaction, communication satisfaction, perceptions of supervisor support, and pressure from the job. The nature of these relationships indicated up to some point or range, job and communication satisfaction, perceived pressure, and supervisor support increased as the amount of information increased. But after this point or range, additional information resulted in a decline in the level of these variables. Thus, too much information has fairly pervasive effects on the work-related attitudes and perceptions of hourly employees.

A linear effect generally typified the relationships for the salaried group, but supervisory support was better defined by a quadratic function. This latter finding indicates that both groups of employees reacted similarly to information abundance. Both reported that when information exceeded some level they perceived the need for more support from their supervisors. The obvious practical implication is that supervisors may have to provide greater assis- tance to both hourly and salaried employees when information overload is perceived. This may entail additional or clarification of job instructions, improving information channels, assignment of extra personnel, etc.

Only a portion of the third hypothesis was supported. The mean load for the two groups indicated that salaried workers, in comparison to hourly workers, did perceive more load. Yet, the quadratic regression results indicated that hourly, in comparison to salaried workers, were more likely to be negatively affected by the level of information they perceived. Several factors, including but not limited to, training and education levels, or experience dealing with large amounts of information, could account for these differences. Alterna- tively, the differences may be due to the power of the large sample size.


However, with only 1% of the variance explained by the graduate regression results, the practical utility of these findings are marginal, at best, since there is no practically significant curvilinear effect of information load on individuals’ work outcomes within the ranges tested. Additional research is needed to provide further tests of the quadratic hypothesis.

Future Research Direction

A primary area for future research is the information load construct itself because the dearth of empirical studies on this concept may be attributed to past imprecision in definition and measurement (O’Reilly, 1980). Generally, information load has been defined and operationalized, particularly in laboratory research which constitutes most of the research on the topic, unidimen- sionally in terms of either (a) the amount of information presented or avail- able or (b) the rate at which information is processed (O’Reilly, 1980).

Some recent discussions of information processing propose a multidimensional construct involving (a) the amount of information, (b) the rate at which information is presented and the time required to examine and understand the messages (Vecchio, 1989), and (c) the background noise, that is the number of simultaneous and often irrelevant stimuli that interfere with the primary task (Arnold & Feldman, 1986). Huber and Daft (1987) extend this multidimensional perspective to include (a) load or the number of symbols or messages received per unit of time, (b) ambiguity or the potential for multiple interpre- tations of a symbol or message, and (c) variety or the complexity and turbu- lence of the information stream. Earlier, we described additional dimensions which could be integrated in measures of information load and quality (e.g., information adequacy, Penley, 1982; availability, Boynton, 1988; and richness, Daft & Lengel, 1984). A multidimensional construct will increase predic- tive and explanatory power enabling researchers to learn more about the intricate problems of information in organizations.

Second, development of an information load measure needs to follow rigorous scientific steps because too often, organizational concepts lack acceptable measures (Price & Mueller, 1986). Measures typically display adequate face validity and reliability, but other considerations are necessary. For example, not only reliability of the entire measure, but reliabilities of all dimensions of a multidimensional construct must be considered (Churchill, 1979). Further, content validity of the items should be assessed by independent judges.

Construct validity also needs to be determined (Schwab, 1980). Construct validity consists of three subtypes: (a) convergent validity, (b) discriminant validity, and (c) nomological validity (Campbell & Fiske, 1959; Green, Tull, & Albaum, 1988). With convergent validity, correspondence between the pro-


posed measure and previous information load measures should be assessed. With discriminant validity, the proposed information load measure should tap a construct different from those tapped by other measures (Campbell & Fiske, 1959). With nomological validity, the proposed measure should exhibit linkages with theoretically appropriate variables (Green et al., 1988).

Third, a more complex construct of information load will make it possible to identify across a variety of work contexts those dimensions most respon- sible for the quadratic relationship. We encourage others to use power polynomials to test for the nonlinear effects of information on people. This procedure can increase the explanatory power of our models particularly when used with multidimensional constructs.

Our findings may be highly informative since they show nonlinear processes underlying relationships between information load and important work outcomes for one classification of employee. Consequently, we believe that subsequent models may have to explicitly include nonlinear relationships when describing some employees’ information processing. Daft and Lengel’s (1984, 1986) concept of richness is an important step in this direction because it implicitly contains both load and quality elements, and it identifies a situation in which information is too rich.

We would encourage subsequent research on information load to proceed along three fronts. First, the development of a reliable, multidimensional measure of information load is important. Second, this development needs to follow rigorous scientific procedures. Third, empirical investigations using a multidimensional measure and polynomial regression will allow students of organizational communication to accurately describe the influences of information load and may aid managers in the better regulation of information. Investigations sampling a variety of work departments and organizational levels, as in the present study, should provide a closer approximation of the complexities of organizational information processing.

Limitations of the Present Study

The validity of our findings must be viewed in light of two potential limitations: common method variance and a single-item measure of information load. There is widespread belief among researchers that common method variance biases the results. Further, researchers often recommend the use of multi-item measures for capturing the construct and increasing reliability (e.g., Churchill, 1979; Schwab, 1980). Though the small variance explained in this study (2%-6%) could partially result from mono-method bias, Spector (1987) recently suggested that the problems attributed to common method variance may be more mythical than real. Similarly, the usefulness of a


single-item measure that globally measures information load cannot be dis- missed, but rather requires further empirical research (cf. Scarpello & Camp- bell, 1983).

It may be tempting to argue that the 1% of additional explained variance contributed by the quadratic function is influenced by end-of-scale compres- sion effects. However, at issue is the importance of the quadratic relationship. Addressing the issue of the importance of statistically significant moderating effects, Champoux and Peters (1980) noted that the size of the increment in R‘ for the interaction effect cannot be interpreted as an index of the importance of a moderating effect. They argued that rather than focusing on the R2 increment, researchers should instead devote attention to the nature of the interaction (Stone, 1988). Similar reasoning may be applied to examining the nature of curvilinear relationships.

O’Grady (1982) applied this reasoning to all measures of explained variance in an important article on the subject. He documented the thesis that measures of explained variance are useful statistics, “but they have no claim to primacy in evaluating the potential contribution of a finding”(p. 767). Further, placing undue emphasis on the size of measures of explained variance without regard to other issues is to run a twofold risk. First, “by accepting only research that reports large measures of explained variance only trivial research is accepted” (p. 775). Second, when a small measure of explained variance results, “much research of importance that might make a contribution in the long term is ignored or rejected” (p. 775).

References

Alexander, E. R., 111, Helms, M. M., & Curran, K. E. (1987). An information processing analysis of organization information adequacy/ abundance. Management Communication Quarterly, 1, 150- 172.

Arnold, H. J., & Feldman, D. C. (1986). Organizationalbehavior. New York: McGraw-Hill.

Berenson, M. L., Levine, D. M., & Goldstein, M. (1983). Intermediate statis- tical methods and applications. Englewood Cliffs, NJ: Prentice Hall.

Boynton, A. C. (1988). Managerial scanning: The interplay between task uncertainty and information context. Manuscript submitted for publication.

Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56,

Champoux, J . E., & Peters, W. S. (1980). Applications of moderated regres- 81-105.

sion in job design research. Personnel Psychology, 33, 759-783.


Chervany, N. L., & Dickson, G. W. (1974). An experimental evaluation of information overload in a production environment. Management Science,

Churchill, G . A., Jr . (1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, 16,64-73.

Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erl- baum.

Connolly, T. (1988). Studies of information-purchase processes. In B. Brehmer & C. R. B. Joyce (Eds.), Human judgment: The SJT View (pp. 401-425). North-Holland: Elsevier Science.

Daft, R. L., & Lengel, R. H. (1984). Information richness: A new approach to managerial behavior and organization design. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (Vol. 6, pp. 191- 233). Greenwich, CT: JAI Press.

Daft, R. L., & Lengel, R. H. (1986). Organizational information requirements, media richness and structural design. Management Science, 32,

Daft, R. L., & Macintosh, N. B. (1981). A tentative exploration into the amount and equivocality of information processing in organizational work units. Administrative Science Quarterly, 26, 207-224.

Daft, R. L., & Weick, K. E. (1984). Toward a model of organizations as interpretation systems. Academy of Management Review, 9, 284-295.

Daft, R. L., & Wiginton, J . C. (1979). Language and organization. Academy of Management Review, 4, 179-191.

Driver, M. J., & Streufert, S. (1969). Integrative complexity: An approach to individuals and groups as information-processing systems. Administrative Science Quarterly, 14, 272-285.

Fulk, J., & Boyd, B. (1991). Emerging theories of communication in organizations. Journal of Management, 17,407-446.

Goldhaber, G . M., Dennis, H. S., Richetto, G. M., & Wiio, 0. A. (1984). Information strategies: New pathways to management productivity. Nor- wood, NJ: Ablex.

Green, P. E., Tull, D. S., & Albaum, G. (1988). Research for marketing decisions (5th ed.). Englewood Cliffs, NJ: Prentice Hall.

Hoppock, R. (1935). Job Satisfaction. New York: Harper & Row. Huber, G. P., & Daft, R. D. (1987). The information environments of organi-

zations. In F. M. Jablin, L. L. Putman, K. H. Roberts, & L. W. Porter (Eds.), Handbook of organizational communication (pp. 130-164). New- bury Park, CA: Sage.

Huber, G. P., O’Connell, M. J., & Cummings, L. L. (1975). Perceived environ-

10, 1335-1344.

554-571.


mental uncertainty: Effects of information and structure. Academy of Management Journal, 18,725-740.

Ivancevich, J. M., & Matteson, M. T. (1987). Organizational behavior and management. Plano, TX: Business Publication.

Jacoby, J., Speller, D. E., & Kohn, C. A. (1974). Brand choice behavior as a function of information load. Journal of Marketing Research, 11,63- 69.

Johns, G. (198 1). Difference source measures of organizational behavior variables: A critique. Organizational Behavior and Human Performance,

Katz D., & Kahn, R. L. (1978). The socialpsychology of organizations (2nd ed.). New York: Wiley.

Lawler, E. E., & Rhode, J. G. (1976). Information and control in organizations. Pacific Palisades, CA: Goodyear.

Litwin, G. H., & Stringer, R. A. (1968). Motivation and organizational climate. Boston: Harvard University.

Manis, M., Fichman, M., & Platt, M. B. (1978). Cognitive integration and referential communication: Effects of information quality and quantity in message decoding. Organizational Behavior and Human Performance, 22,

McNichols, C. W., Stahl, M. J., & Manley, T. R. (1978). A validation of Hoppock’s job satisfaction measure. Academy of Management Journal,

Muchinsky, P. M. (1977). Organizational communication: Relationships to organizational climate and job satisfaction. Academy of Management Journal, 20, 592-607.

Nunnally, J . C. (1978). Psychometric theory (2nd ed.). New York: McGraw- Hill Book Company.

O’Grady, K. E. (1982). Measures of explained variance: Cautions and limitations. Psychological Bulletin, 92, 766-777.

O’Reilly, C. A. (1980). Individuals and information overload in organizations: Is more necessarily better? Academy of Management Journal, 23,

Pedhazur, E. J. (1982). Multiple regression in behavioral research (2nd ed.). New York: Holt, Rinehart, & Winston.

Penley, L. E. (1982). An investigation of the information processing frame- work of organizational communication. Human Communication Re- search, 8, 348-365.

Penley, L. E., & Hawkins, B. (1985). Studying interpersonal communication in organizations: A leadership application. Academy of Management Journal, 28,309-326.

27,443-463.

41 7-430.

21,737-742.

684-696.


Porter, L. W., & Roberts, K. (1976). Communication in organizations. In M. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 1553-1589). Chicago: Rand McNally.

Price, J. L., & Mueller, C. W. (1986). Handbook of organizational measurement. Marshfield, MA: Pitman.

Roberts, K. H., & O’Reilly, C. A. (1974a). Failures in upward communication in organizations: Three possible culprits. Academy of Management Jour- nal, 17,205-214.

Roberts, K. H., & O’Reilly, C. A. (1974b). 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.

Rogers, E. M., & Agarwala-Rogers, R. (1976). Communication in organizations. New York: Macmillan.

Scarpello, V., & Campbell, J. P. (1983). Job satisfaction: Are all the parts there? Personnel Psychology, 36, 577-600.

Schwab, D. P. (1980). Construct validity in organizational behavior. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior. (Vol. 2, pp. 3-43). Greenwich, CT: JAI Press.

Sims, H. P., & LaFollette, W. R. (1975). An assessment of the Litwin and Stringer organization climate questionnaire. Personnel Psychology, 28,

Spector, P. E. (1982). Behavior in organizations as a function of employee’s

Spector, P . E. (1987). Method variance as an artifact in self-reported affect and perceptions at work: Myth or significant problem? Journal of Applied

Steers, R. M. (1977). Organizational effectiveness: A behavioral view. Glen- view, IL: Scott, Foresman and Co.

Steers, R. M. (1984). Introduction to organizational behavior (2nd ed.). Glenview, IL: Scott, Foresman and Co.

Stone, E. F. (1989). Moderator variables in research: A review and analysis of conceptual and methodological issues. In K. M. Rowland & G. R. (Eds.), Research in Personnel and Human Resources Management (Vol. 6, pp. 191-229). Creenwich, Ct: JAI Press.

Streufert, S., & Driver, M. J. (1965). Conceptual structure, information load, and perceptual complexity. Psychonomic Science, 3 , 249-250.

Vecchio, R. P. (1989). Organizational behavior. Chicago: Dryden. Weick, K. (1979). The socialpsychology of organizing. Reading, MA: Addi-

19-38.

. locus of control. Psychological Bulletin, 91,482-497.

Psychology, 72,438-443.

son- Wesley.


Weick, K. E. (1987). Theorizing about organizational communication. In F. M. Jablin, L. L. Putnam, K. H. Roberts, & L. W. Porter (Eds.), Handbook of organization communication (pp. 97-122). Newbury Park, CA: Sage.

Information Load: A Test of an Inverted-U Hypothesis with Hourly and Salaried Employees

Documents