Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

645

The British Psychological Society

Journal of Occupational and Organizational Psychology (2006), 79, 645–671 q 2006 The British Psychological Society

www.bpsjournals.co.uk

Organizational climate systems and psychological climate perceptions: A cross-level study of climate-satisfaction relationships Mathis Schulte1*, Cheri Ostroff2 and Angelo J. Kinicki3 1

Teachers College, Columbia University, USA University of Maryland, USA 3 Arizona State University, USA 2

Research has consistently demonstrated that both individual-level climate perceptions and organizational climate are related to job satisfaction; however, little work has investigated their relative importance in a single study. Using a sample of 1,076 employees from 120 branches of a US-based bank, the relative importance of individualand unit-level climate on individual satisfaction was examined. Cross-level results of hierarchical linear models indicated that individuals’ perceptions of the climate accounted for a large percentage of variance in individuals’ satisfaction. Further, unitlevel climate systems accounted for a small but significant portion of individual satisfaction above and beyond individuals’ perceptions of the climate. These results suggest that the overall climate in a work unit has some influence on individual attitudes, after accounting for individuals’ idiosyncratic perceptions of the climate.

Researchers in organizational behaviour have long been interested in understanding employees’ perceptions of the work environment and how these perceptions influence individuals’ work-related attitudes and behaviours. Early researchers suggested that the social climate or atmosphere created in a workplace had significant consequences – employees’ perceptions of the work context purportedly influenced the extent to which people were satisfied and perform up to their potential, which, in turn, was predicted to influence organizational productivity (e.g. Katz & Kahn, 1978; Likert, 1967; McGregor, 1960). The construct of climate has been studied extensively and has proven useful in capturing perceptions of the work context (Denisson, 1996; Ostroff, Kinicki, & Tamkins, 2003). Climate has been described as an experientially based description of the work environment and, more specifically, employees’ perceptions of the formal and informal policies, practices and procedures in their organization (Schneider, 2000). An important distinction has been made between psychological and organizational climate (Hellriegel & Slocum, 1974; James & Jones, 1974). Individuals’ own perceptions

* Correspondence should be addressed to Mathis Schulte, Teachers College, Columbia University, Dept. of Organization and Leadership, 525 W. 120th St., Box 6, New York, NY 10027 (e-mail: [email protected]). DOI:10.1348/096317905X72119

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

646 Mathis Schulte et al.

of the work environment constitute psychological climate at the individual level of analysis, whereas organizational climate has been proposed as an organizational or unitlevel construct. When employees within a unit or organization agree on their perceptions of the work context, unit-level or organizational climate is said to exist (Jones & James, 1979; Joyce & Slocum, 1984). A large number of studies have consistently demonstrated relationships between psychological climate and individuallevel outcomes such as satisfaction, commitment, performance and stress. Likewise, a number of cross-level studies have consistently demonstrated positive relationships between unit or organizational climate and individual outcomes such as performance, satisfaction, commitment, involvement and accidents (Ostroff et al., 2003). While past research has greatly contributed to our understanding of relationships between psychological and organizational climate and a diverse set of individual-level criteria, there are two key limitations inherent in this work. First, studies have tended to focus on either psychological or organizational climate and have ignored the relative influence of psychological and organizational climate on individual outcomes. This is an important omission because employee attitudes and behaviours may not only be influenced by one’s personal perceptions of the work environment but also by the shared perceptions of co-workers (Mathieu & Kohler, 1990). The study of emergent processes suggests that a work group’s shared perceptions might influence individual attitudes and behaviours above and beyond individual perceptions of the work environment (Kozlowski & Klein, 2000). Second, research has increasingly examined a global index representing a single strategically focused climate (e.g. a climate for service or a climate for safety) or has focused on a set of climate dimensions (Ostroff et al., 2003). Examining single dimensions or a set of independent dimensions of climate ignores the broader context in which they are operating. This is a limitation because it may be useful to examine multiple dimensions of climate together, as a system. Different organizational attributes are likely to mutually reinforce one another, making the total effect greater than the sum of individual dimensions (Bowen & Ostroff, 2004). The current study was designed to overcome the above two limitations. Specifically, we examine the relative impact of psychological and unit-level climate on individual satisfaction and evaluate whether shared perceptions in a unit are related to an individual’s level of satisfaction above and beyond their own perceptions of the climate. In addition, we adopt a configural approach for capturing organizational climate to define a more comprehensive and integrative conception of the social environment, and examine the contextual effect of organizational climate systems on individual satisfaction after controlling for individual climate perceptions.

Past research on climate In this section, we first provide an overview of climate across levels of analysis and explain a configural approach to studying organizational climate. We then address our core predictions regarding the relative impact of psychological climate and climate systems on employee job satisfaction. Climate across levels of analysis Psychological and organizational climate are conceptually related to one another. Psychological climate pertains to how organizational members perceive and make sense of organizational policies, practices and procedures in psychologically meaningful terms

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

647

(Schneider & Rentsch, 1988). Such perceptions can be idiosyncratic, even when individuals are exposed to the same work context and situation (James & Tetrick, 1986). Organizational climate emerges from these idiosyncratic interpretations of the work environment when individuals within a particular unit (e.g. group, organization) share similar perceptions of the situation. Only when individuals agree on their perceptions of the work environment can their individual perceptions be meaningfully aggregated to represent unit- or organizational-level climate (James, 1982; Klein et al., 2000). Therefore, the relationship between psychological and organizational climate can be described as compositional in that both constructs reference the same content but describe qualitatively different phenomena at the individual and unit levels of analysis (Chan, 1998; James, 1982). Psychological climate is a property of the individual, but when shared across individuals within a unit or organization, the aggregate of the responses represents the construct of unit or organizational climate (Glisson & James, 2002). As such, organizational climate is purported to be an emergent property because it originates in the cognition and perceptions of individuals, and is amplified through interactions and exchanges with other unit members to manifest as a higher-level collective phenomenon (Kozlowski & Klein, 2000). Different explanations have been offered about how individuals’ interpretations of the organizational environment emerge and are transformed into shared perceptions (cf. Ostroff et al., 2003; Schneider & Reichers, 1983). From a structural perspective, it has been suggested that unit or organizational characteristics such as size and structure (e.g. Payne & Mansfield, 1973) as well as consistency, clarity and salience in policies, practices and procedures (e.g. Bowen & Ostroff, 2004) can establish a common reality that provides the basis for shared perceptions. Further, through the process of attraction, selection and attrition (Schneider & Reichers, 1983) an organization is likely to comprise people with similar views and attributes so that individuals tend to perceive and experience the work environment similarly. Communications and repeated social interactions among members of the same unit or organization influence individual views and can also contribute to the evolvement of shared perceptions and meaning (e.g. Klein, Conn, Smith, & Sorra, 2001; Morgeson & Hofmann, 1999). The notion of within-group agreement as a precondition for unit or organizational climate does not necessarily mean that there is perfect agreement among individuals on climate. In fact, most studies that have investigated group or organizational climate have found that there is still some variability in perceptions within groups (e.g. Gonzalez-Roma, Peiro, & Tordera, 2002; Lindell & Brandt, 2000; Schneider, Salvaggio, & Subirats, 2002). A configural approach to unit-level or organizational climate A great deal of attention has been devoted to distinguishing between the objective versus perceptual nature of climate (cf. Glick, 1985; James, Joyce, & Slocum, 1988) and between psychological and organizational climate (e.g. James & Jones, 1974; Jones & James, 1979) as well as to methodological issues pertaining to the aggregation of individual climate perceptions to represent organizational climate (e.g. Chan, 1998; Klein et al., 2000). The controversies surrounding these issues have largely been resolved (Schneider, 2000). However, little attention has been directed at how best to capture climate as a system-wide variable in an organization. The notion that multiple climates exist within an organization has been widely accepted (e.g. Schneider, 2000). Yet, empirical research has tended to examine a single climate dimension or examine the relative importance of several dimensions of climate

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

648 Mathis Schulte et al.

in a single study. Ostroff and her colleagues (Bowen & Ostroff, 2004; Ostroff et al., 2003) have suggested that a configural approach (Doty, Glick, & Huber, 1993; Meyer, Tsui, & Hinings, 1993) might be fruitful in this context. Configurations can broadly be defined as ‘conceptually distinct characteristics that commonly occur together’ (Meyer et al., 1993, p. 1175); they allow for examining multiple characteristics simultaneously while accounting for the interrelationships and interactions among them. Applied to the study of organizational climate, organizations or work units would be characterized by several distinct profiles across multiple climates. In this case, the focus of measurement shifts from examining independent climate dimensions to patterns or systems of interrelated climate dimensions. Configural approaches have proven useful in other areas of organizational research, particularly in human resource management (HRM). Individual HRM practices have been combined to form unique patterns of practices that depict different configurations, and these different configurations have then been related to effectiveness outcomes (e.g. Delery & Doty, 1996; Doty et al., 1993; Ichniowski, Shaw, & Prennushi, 1997). This body of research is based on the assumption that different HRM practices are interrelated and interact as a system in achieving their effects. Examining single practices or sets of practices simultaneously in a regression does not allow for capturing complementary effects and interrelations among the practices – only by examining configurations across all practices can we determine whether the entire system of practices, taken together, explains more than the sum of the effects of the individual practices (Ichniowski et al., 1997). Individual practices are believed to have limited ability to impact a particular outcome. Rather, in combination, the system of practices enables organizations to achieve higher performance (Becker & Gerhart, 1996). Further, it is also assumed that some patterns or configurations can be equally effective or equifinal (Delery & Doty, 1996; Meyer et al., 1993). Moving from HRM configurations to unit or organizational climate configurations is reasonable because climates are largely based on the perceptions of HR practices, polices and procedures (Kopelman, Brief, & Guzzo, 1990; Ostroff & Bowen, 2000). Configurations may provide a more integrative view of the overall climate in a particular unit or organization than focusing on single climates, or the independent or relative impact of several climate dimensions. Coherent patterns of multiple climates correspond to Lewin, Lippit and White’s (1939) notion of climate as a ‘Gestalt of the social environment’. It is also reasonable to assume that different climate dimensions interact and are interrelated in non-linear ways, which can be captured by a configural approach. Alternatively, all possible interactions among climates could be examined to account for non-linear effects. However, as the number of climates considered increases, the number of interaction terms increases exponentially, which not only requires very large sample sizes, but also makes the interpretation cumbersome. Relative impact of psychological climate and climate systems A great deal of research has indicated that psychological climate and organizational (or unit-level) climate is related to a variety of individual outcomes (e.g. Carr, Schmidt, Ford, & DeShon, 2003). For example, a number of studies have shown that psychological climates are related to individual satisfaction (e.g. Friedlander & Margulies, 1969; Johnson & McIntye, 1998). Results from two recent meta-analytic studies also provide strong support for this relationship (Carr et al., 2003; Parker et al., 2003). In addition, cross-level studies have demonstrated that unit-level or organizational climates are also significantly related to individual satisfaction

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

649

(e.g. Joyce & Slocum, 1984; Naumann & Bennett, 2000; Ostroff, 1993). However, there is an obvious lack of research examining psychological and higher-level unit or organizational climate at the same time to ascertain their relative impacts. Although new to the area of climate, the idea of comparing the relative importance of individual and group-level attributes on individual attitudes and behaviours has a long history in sociology and education. Sociologists, who have supported the group effects theory, have argued that groups can (and do) have effects over and beyond those of the attributes of the group members (e.g. Blau, 1960; Merton & Kitt, 1950). For example, Blau found that workers in public assistance agencies showed more service-oriented behaviour when they worked in groups with strong pro-client values than those who worked in weak pro-client-value groups, after holding constant their individual proclient values. Blau interpreted the social values that prevailed in the work groups as ‘external constraints upon the thinking and acting of its members’ (p. 182). Workers were not only guided by their own values, but also sought social approval of colleagues by acting in congruence with the prevailing group values. Similar notions are evident in social information processing theory (Salancik & Pfeffer, 1978) whereby job-related attitudes are purportedly based on both individuals’ perceptions (which are driven by their earlier experiences and behaviours) as well as on the immediate social context (e.g. perceptions of co-workers). The complexity of the work environment is believed to require people to rely on social cues in addition to their own perceptions in order to make sense of the situation. In line with this argument, theories on sense-making processes have stated that the cognitive representation of the organizational experience is not only determined by individual patterns of thinking and understanding but also by influential relationships and organizational norms (Gioia, Thomas, Clark, & Chittipeddi, 1994). A number of early studies tested the social versus individual bases for job attitudes by comparing the influence of individual demographic variables (e.g. age, gender, tenure) on social structural variables such as group, department or division affiliations (cf. Herman, Dunham, & Hulin, 1975). Results showed that group affiliations explained individual attitudes better than individual demographic characteristics. More recently, Liao and Chuang (2004) found that store-level service climate was related to individual service performance after accounting for individual-level personality traits such as conscientiousness and extraversion. However, in these studies, the individual-level and unit-level variables represented different constructs, rather than commensurate or compositional constructs at different levels of analysis (Chan, 1998). Few studies in organizational research have simultaneously examined the impact of similar constructs at different levels of analysis on individual outcomes. For example, Mathieu and Kohler (1990) demonstrated that group-level absence is positively related to individual absence above and beyond individual-level antecedents of absenteeism. Similarly, Blau (1995) reported positive effects of group-level employee lateness on individual lateness after controlling for individuallevel antecedents such as work-related attitudes, illness and accidents, weather and work-family conflict. Although these studies included similar constructs at different levels of analysis, the individual-level outcome (e.g. individual absence) was related to a compositional organizational-level predictor (e.g. organizational-level absence). Thus, it is unknown whether the analogous constructs at two levels (e.g. individual absence and organizational absence, or psychological climate and organizational climate) have independent and relative effects on separate outcomes such as job

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

650 Mathis Schulte et al.

satisfaction. A comparison between individual- and unit-level effects of functionally similar constructs is needed. We propose that an employee’s job satisfaction is not only influenced by his or her own perceptions of the climate, but also by the shared perceptions of his or her work unit. Further, we assume that due to the psychosocial processes that operate at the group level, unit-level climate differs from psychological climate in two important ways. First, perceptions of different aspects of the work environment are shared among the people in a work unit, and second, these perceptions form a coherent entity or ‘Gestalt’ (Lewin, Lippit, & White, 1939). We propose that this interplay of perceptions is particularly important in the social context and constitutes the contextual effect on individual satisfaction that cannot be explained by the individual-level perceptions. That is, we do not expect that shared dimensions alone will influence individual satisfaction beyond individuals’ own perceptions because they capture an analogous construct. However, the climate system captures a Gestalt construct that provides a better understanding of how the social context as a whole influences individuals beyond their own perceptions. Thus, in order to account for both processes (shared perceptions of single dimensions and interrelationships across dimensions), we conceptualize climate at the unit level of analysis with a configural approach instead of single climates or a set of climate dimensions. Because there is a lack of theory or empirical research to make a priori predictions regarding the patterns of climate configurations that exist, we derive them empirically in this study. As justification for this approach, it is important to note that our primary purpose is not to elucidate the types of climate systems that might exist, but rather to examine whether shared perceptions of overall climate systems explain job satisfaction beyond individuals’ own climate perceptions. Climate at the individual level can also be construed either as distinct dimensions or as systems. Depending on the conceptualization of psychological climate, the notion of unit-level climate explaining additional variance in job satisfaction beyond that explained by psychological climate takes on a slightly different meaning. First, individual-level psychological climate dimensions and unit-level climate systems have in common that they are both based on the same individual perceptions of organizational policies, practices and procedures. However, they differ in that unit-level climate systems are based on the shared perceptions within work units and further that they take into account the synergy and non-linear interrelatedness among climates. In contrast, psychological climate dimensions and unit-level climate dimensions are both based on perceptions of the work context but the unit-level climate dimensions do not account for the potential interconnections among the climate dimensions and hence are not as likely to capture the ‘Gestalt’ of the social system as well as climate systems. By controlling for the set of psychological climate dimensions, the higher-level contextual influences of climate can be examined independently from individual perceptions. Moreover, examining higher-level climate as a system allows for examining the unique contribution of contextual properties such as shared perceptions as well as the interrelatedness among organizational climates. Second, psychological climate can also be conceptualized as systems. Individual perceptions may ‘work together’ in forming an overall climate impression, or psychological climate system, for a focal person. In this case, different profiles of multiple climates reflect the different ways that individuals see the entire system of climates. Individuals within profile groups have similar ways of seeing the work environment but they may not necessarily be part of the same formal units or workgroups (Joyce & Slocum, 1984; Patterson, Payne, & West, 1996; Payne, 1990). That

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

651

is, it is possible that some people in a particular unit may have psychological climate systems that are similar to those of individuals in others units, but differ from the psychological climate systems of some individuals in their immediate work unit. Similar perceptions of the climate system across individuals in different formal units might arise from frequent interactions and communications, similar jobs (Joyce & Slocum, 1984), similar personalities or interests (Payne, 1990). Nevertheless, while the individual-level psychological climate system captures an individual’s own impression of the interrelatedness among the climate dimensions, the unit climate can still have a significant influence on the affective responses of that person. That is, the unit climate represents the shared perceptions of co-workers in the immediate surrounding of that person and therefore is likely to have a social influence on that person above and beyond his or her own climate system. Climate systems at the individual and unit level have in common that they are both based on the same individual climate dimensions and they both take into account the interrelatedness of these dimensions. However, they differ in that the interrelatedness among these dimensions is based on different processes at the different levels of analysis. One is based on individualistic views whereas the other represents an overall Gestalt across individuals within the social context or unit. By controlling for psychological climate systems, we can examine the contextual influence of shared perceptions beyond that explained by the individual processes of climate formation. In sum, although there are differences in meaning when examining psychological climate dimensions versus individual climate systems, in both cases, it is reasonable to assume that the social influence of unit-level climate systems affects an employee’s job satisfaction above and beyond his or her individual perceptions of the climate. Based on the discussion above, we hypothesized the following: Hypothesis 1. Unit-level climate systems will be related to individual job satisfaction above and beyond psychological climate dimensions. Hypothesis 2. Unit-level climate systems will be related to individual job satisfaction above and beyond psychological climate systems.

Method Sample and procedure Survey data were collected from employees in 153 branches in the western region of a national bank in the USA. All employees were asked to participate. The survey was administered through company mail. Employees were asked to place their completed survey in an enclosed envelope to be mailed to the researchers. Surveys were returned by 3,344 employees in 153 branches. However, many employees responded in a way that did not allow for determining the bank branch in which they worked. We were able to accurately classify 1,601 employees into their bank branch. Because unit-level climate measures were based on aggregated scores at the branch level, only data from employees for which there were five or more respondents in the branch were retained for analysis. This cut-off of five per unit has been suggested by Bliese (1998) because biases in using aggregate scores start to diminish with groups of five and more employees. The final sample included 1,076 employees in 120 branches. Of these, 982 employees indicated employment status: 72% were full-time and 28% were parttime employees. Usable data on employee jobs was obtained from 1,048 employees: 15% were managers, 26% were sales/service managers or personal bankers, 4% were

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

652 Mathis Schulte et al.

teller supervisors, 18% were service quality representatives, 36% were tellers and 1% were classified as ‘other’. The sponsoring organization did not allow the collection of any additional demographic information due to confidentiality concerns. The average number of employees per branch was 9, ranging from 5 to 17.

Measures and scoring Satisfaction We used 10 items from the Minnesota Satisfaction Questionnaire (Weiss, Dawis, England, & Lofquist, 1967) to assess overall satisfaction. Items tapped an employee’s satisfaction with various aspects of his or her organization, including satisfaction with the immediate supervisor, co-workers, total compensation, work activities, work load, opportunities for promotion, job security, activities outside the job description, physical environment and interactions with clients. All items used a 5-point Likert-type scale, ranging from very dissatisfied (1) to very satisfied (5). Internal consistency reliability was .80.

Psychological climate Items were developed based on input from: (1) nine focus group meetings with employees, and (2) past studies on psychological and organizational climate. A total of 65 employees participated in focus groups. Participants brainstormed answers to the following two questions: tell me what it’s like today in your branch? What are the things we need to ask about in the survey to get a full picture of what it’s like to work in the bank? After responses had been summarized, participants were asked to sort the responses into broader categories of climate. Based on the categories identified in each group, participants were then asked ‘what specific kinds of questions would you like to see included on the survey to get at these categories?’ Respondents individually generated ideas and submitted their written suggestions to a member of the research team. This pool of potential survey questions was culled across the focus groups and edited by comparing the items to measures of climate used in other studies. The final survey included 94 items that covered the many different contextual and climate areas including management, vision, organizational change, training, career opportunities, recognition, rewards, teamwork and communication. Employees responded to all items using a 5-point Likert-type scale ranging from strongly disagree (1) to strongly agree (5). Climate scales were developed by first subjecting all items to an exploratory factor analysis with varimax rotation, which yielded an 11-factor solution. Factor scales were constructed by averaging the items that loaded unambiguously at .50 or higher on a single factor.1 Two scales with 3 and 4 items, respectively, were combined into one scale because all items focused on the perceptions of clarity/ambiguity in policies, practices and procedures and the two scales were highly correlated. Two scales were dropped due to low reliability. Thus, a total of eight climate scales were considered in further analyses: managerial support, company vision, open and clear communication, training 1 Two additional factor analyses were conducted. The first was based on the aggregated climate items at the unit level, the second was based on the within-branch component of items’ variance. All three analyses (individual-level items, aggregated item and within-branch component of item variance) revealed very similar factor structures. The results of the factor analyses can be obtained by contacting the first author.

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

653

focus, team focus, clarity, personnel support for service and rewards for service. These dimensions are similar to those used in previous research in service organizations and have been identified as important dimensions for achieving a climate for service (e.g. James & James, 1989; Schneider & Bowen, 1985; Schneider, Wheeler, & Cox, 1992; Schneider, White, & Paul, 1998). Table 1 contains the reliability coefficients, number of items, average rWG(j) scores and sample items for each of the climate dimensions. Table 1. Climate dimensions

Dimension (Definition)

Alpha

No. Average rWG(j) of items Sample items

Managerial support (manager supportive of employees)

.95

.91

14

Company vision (strategic goals and visions for future clear and visible)

.78

.77

4

Open and clear communication (clear, appropriate and adequate communication within and between units) Training focus (sufficient job and continuous training provided)

.81

.80

5

.75

.77

4

Team focus (emphasis on team structure and team processes)

.80

.63

2

Clarity (role and job clarity)

.64

.86

7

Personnel support for service (staff and support available to meet customer needs)

.83

.84

6

Rewards for service (rewards and incentives provided for attention to customer service)

.87

.73

5

My manager is in touch with employee concerns. My manager motivates me to do well I understand [name of the bank]’s vision. Senior management presents a clear vision of a positive future There is adequate communication between units. I receive clear communication about changes in the organization I am adequately trained for my job. I am sufficiently trained in new policies and procedures I feel like a part of a team within my work unit. We do fun things in my unit to build a good team I feel certain about how much authority I have. I get assignments that contradict each other (2 ) The Bank as a whole has adequate staff to service our clients. [Name of the bank] provides quality customer service to our clients The current incentive plan motivates me to attract new clients and new funds. The current incentive plan motivates me to refer business to other areas where appropriate

Aggregation of climate scores Individual climate scores were aggregated (averaged) to the branch level. Researchers have generally agreed that within-unit agreement as well as sufficient between-unit variability must be demonstrated to justify the aggregation of individual climate scores (Klein et al., 2000). Between-unit variability was first determined based on a one-way ANOVA with branch as the independent variable. Results of one-way ANOVAs for each of the climate dimensions were significant ( p , .05) indicating that all climate scores varied significantly by branch.

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

654 Mathis Schulte et al.

In addition, the intra-class correlations ICC (1) and ICC (2) were calculated (James, 1982). ICC (1) represents the ratio of the variance between units of analysis (branches) to the total variance based on individual ratings of each respondent. Across all climate dimensions, ICC (1) values ranged from .09 to .21, with a median of .11. James reported ICC (1) values ranging from 0 to .5 and a median value of .12 in the organizational literature, and Bliese (2000) indicates that ICC(1) values are typically in the range of .05– .20. ICC (2) provides an estimate of the reliability of the group means. ICC (2) values ranged from .49 to .72, with a median of .55. In general, ICC(2) values above .7 are considered acceptable, and values between .5 and .7 are considered marginal (Klein et al., 2000). Although some of the reported ICC (2) values are slightly below the recommended level, they do not seem to be low enough to prohibit aggregation. Within-unit agreement was evaluated with rWG(j) (James, Demaree, & Wolfe, 1984). The average rWG(j) for each of the eight climate scales ranged from .63 to .91 (see Table 1) with an overall median value of .79. Generally, rWG(j) values near or above .7 have been viewed as sufficient for justifying aggregation (e.g. James et al., 1984; Klein et al., 2000). Although a few branches showed rWG(j) values below .7, the overall level of rWG(j) indicated adequate agreement among employees within bank branches. Taken together, sufficient between-unit variability and within-unit agreement was demonstrated to justify aggregating the individual climate scores to the branch level.

Unit-level climate systems The eight aggregate climate scales were used as the basis for a cluster analysis to identify the unit-level climate systems. The goal of cluster analysis is to identify homogenous groups based on the similarity between profiles of the variables (Aldenderfer & Blashfield, 1984). Ward’s method was used for the clustering procedure because it produces more easily interpretable clusters than those produced by other procedures (Ketchen & Shook, 1996). A four-group solution was found with cluster analysis. Each branch was categorized into one of the four climate systems. Figure 1 graphically presents the results of the cluster analysis for the four unit-level climate systems. Scores on climate dimensions were standardized to facilitate the interpretation of the graph. Two of the configurations that resulted from the cluster analyses captured different arrays of perceived support for service in organizations. That is, support for service can come from people within the immediate work environment (e.g. from employees’ direct managers and from their peers) or from the organization as a whole (e.g. rewards for service and communication). In the first configuration, branches with a climate system of organizational support for service were lower on climate dimensions that reflected support that originated within a specific branch (such as managerial support and team focus) but higher on climate dimensions more reflective of the amount of support for service provided by general practices in the organization (e.g. personnel support such as adequate staffing levels for service and communication). In contrast, in the second configuration labelled the work environment support for service systems, managerial support and team focus were relatively strong and the more macro aspects of support for service were lower. Branches that were high across all dimensions (and hence strong on both organizational and work environment support for service comprised the third system, labelled the comprehensive system. In these branches, all climate dimensions are relatively high. In contrast, the poor climate system shows that climate is low across all dimensions, and hence branches with this profile have the weakest climates relative to branches with other configurations. These latter

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

655

Figure 1. Unit-level climate systems.

two profiles can be considered as indicating that most employees in the organization share the view that climate is either generally positive or generally weak overall. A set of dummy variables was created to represent the climate systems during hypothesis testing. Personal climate systems We used the same procedure to derive climate systems at individual level of analysis as we did for the unit-level systems. Here, however, instead of the aggregates climate dimensions, the eight individual-level climate scales were used in a cluster analysis. A four-cluster solution resulted and each individual was assigned to one of these four clusters groups. As shown in Figure 2, the four climate configurations bore a reasonable resemblance to those at unit level of analysis (see Figure 1). A comprehensive system with relatively high climate dimensions contrasted with an average system where all climate dimensions were slightly above their mean scores. This latter configuration is similar to the poor climate system at unit level of analysis but showed an overall higher elevation than its counterpart. Corresponding to the systems at unit level, two complementary configurations emerged. One was relatively high on team focus and managerial support (work environment support for climate) whereas the other one was relatively low on these dimensions (organizational support for climate). Overall, climate systems at individual and unit level had very similar shapes but differed in their relative elevations.

Control variables Employees from two different kinds of bank branches participated in the study: Banking centres that offered all available services to customers and express centres that had a limited number of services available to customers and that were predominantly located

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

656 Mathis Schulte et al.

Figure 2. Individual-level climate systems.

in supermarket and department stores. A dummy variable (0–banking centres, 1– express centres) was used as a control variable in subsequent analyses. The average rWG(j) value for each dimension of climate was calculated for each branch and was included as a control variable in further analyses because the extent to which employees within a unit (branch) agree on the climate might influence individual satisfaction independently from the content of climate itself (Gonzalez-Roma et al., 2002).

Results The means, standard deviations and intercorrelations of psychological climates, satisfaction and control variables are presented in Table 2. Intercorrelations among the eight psychological climate scales were significant, ranging from .24 to .56 with a median r of .39 indicating that relatively separate dimensions of climate were assessed. As expected, correlations between psychological climates and satisfaction were high, ranging from .44 to .64 with a median r of .51. Cross-level correlations between individual satisfaction and climate systems revealed a negative relationship for a poor climate system (r ¼ 2:26, p , .01) and a positive relationship for a comprehensive climate systems (r ¼ .29, p , .01). Although not significant, the correlation between satisfaction and the work environment support for service system was negative (r ¼ 2:02) and the correlation with the organizational support for service system was positive (r ¼ :07). In terms of individual-level or personal climate systems, satisfaction was significantly related to each system. Positive correlations resulted for the individual-level comprehensive system (r ¼ :50) and average system (r ¼ :19) while negative correlations resulted for the work environment support (r ¼ 2:30) and organizational support for service systems (r ¼ 2:39).

0.80 0.80 0.78 0.69 0.99 0.60 0.79 0.90 0.63 0.40 0.08

3.28 0.20 0.79

SD

3.79 3.51 3.06 3.53 3.48 3.03 2.52 3.34

M

.05 .06

.46

(.95) .33 .29 .28 .56 .38 .24 .29

1

.17 .15

.51

(.78) .46 .41 .33 .40 .51 .41

2

.04 .11

.53

(.81) .45 .28 .51 .51 .37

3

.21 .11

.47

(.75) .30 .39 .39 .37

4

.15 .12

.44

(.80) .28 .26 .30

5

.16 .14

.58

(.64) .53 .39

6

.28 .15

.64

(.83) .41

7

.18 .18

.51

(.87)

8

.18 .16

(.80)

9

– .22

10

N ¼ 1,076 and correlations $ j.05j are significant at p # .05 for all variables except control variables where N ¼ 120 and correlations $j.17j are significant at p # .05. Diagonal entries are alpha coefficients.

Individual climates 1. Managerial support 2. Company vision 3. Communication 4. Training focus 5. Team focus 6. Clarity 7. Personnel support for service 8. Rewards for service Outcome variable 9. Satisfaction Control variable 10. Branch type 11. Average agreement on climates

Variable

Table 2. Means, standard deviations and intercorrelations of individual level climate dimensions, satisfaction and level unit control variables

Copyright © The British Psychological Society

Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems 657

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

658 Mathis Schulte et al.

Hierarchical linear model analysis The hypotheses predicted that unit-level climate systems would be related to individual satisfaction once individual percepts of climate dimensions were accounted for. This relationship was predicted to hold whether psychological climate is conceptualized as a set of independent climate dimensions (Hypothesis 1) or as a personal climate system (Hypothesis 2). Hierarchical linear modelling analysis (HLM; Raudenbush, 2002) was employed to test the hypotheses because it is particularly appropriate for testing cross-level relationships. First, relationships between individual climate perceptions and individual satisfaction were assessed. Next, unit-level climate systems were explored after accounting for individual climate. Further, significant differences between unit-level climate systems in explaining individual satisfaction were addressed. This procedure was used first with psychological climate conceptualized as independent dimensions then with psychological climate as systems. Finally, although we did not expect significant relationships, we also tested the extent to which the set of independent unitlevel climate dimensions explained variance in satisfaction beyond that explained by the set of individual climate dimensions. HLM follows a two-stage approach. First, Level 1 relationships among individual-level variables are estimated for each group separately. This is equivalent to conducting ordinary least squares (OLS) regression analyses for each of the units or bank branches independently. The overall individual-level (Level 1) relationships are analogous to pooled parameter and standard error estimates across all groups. The parameter estimates (intercepts for each group) from the first stage are then used as outcome variables in Level 2 to examine group-level effects. Several models were estimated, each differing in the number of predictors that were included in the analysis. In the first model (Model 0), no predictor variables were added. Model 0 was used to partition the variance in satisfaction into between- versus withinunit (branch) variability. This procedure is equivalent to a one-way ANOVA with branch as the independent variable and satisfaction as the dependent variable. This first step is necessary to examine the amount of between-unit variance that can be explained by the predictors. As can be seen in Tables 3 and 4, Model 0 reveals that a modest but significant proportion of total variance in satisfaction (12.9%) was explained by branch membership. Significant variance between units justifies the inclusion of predictors at the unit level of analysis. Once significant between-unit variance is demonstrated in Model 0, individual-level predictors (i.e. either psychological climate dimensions or systems) are included in Model 1. Next, unit-level predictors are included, with control variables being included in Model 2 and unit-level climate variables included in Model 3. Models 2 and 3 represent Level 2 analyses whereby unit-level variables are used to predict the intercepts derived from the Level 1 analyses. The first set of HLM analyses was used to test Hypothesis 1 that unit climate systems will explain additional variance in individual satisfaction beyond that explained by psychological climate dimensions (Table 3). In Model 1, all eight psychological climate dimensions were added. Further, parameter estimates from a fixed effects model for each of the variables are contained in the lower portion of Table 3. Significant results were observed for each of the psychological climate dimensions indicating that psychological climates are significantly related to job satisfaction. These results are consistent with much prior work on the relationship between individual-level psychological climates and individual attitudes (e.g. Friedlander & Margulies, 1969; Johnson & McIntye, 1998). Comparing the variance components of Model 1 with the

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

659

Table 3. HLM of the effect of individual-level psychological climate dimensions and unit-level climate systems on satisfaction

Level 1 residual variance Model 0 – No predictors Model 1 – Individual-level climate dimensions (level 1) Model 2 – Individual-level climate dimensions and Level 2 Controls Model 3 – Individual-level climate dimensions, Level 2 controls, Unit-level climate systems Model 3 Parameter estimates Level 1 Individual-level climate dimensions Constant Managerial support Company vision Communication Training focus Team focus Clarity Personnel support/service Rewards for service Level 2 Controls Branch type Average agreement Level 2 Unit-level climate systems Constant Comprehensive Work environment support for service Poor

Level 2 residual variance

Level 2 variance explained by model (%)

Percentage of Level 2 to overall variance (%)

.345 .124

.051 .008

84.54

12.90 6.04

.124

.008

20.29

6.14

.123

.007

2.01

5.44

Coefficient (SE)

t value

3.28 (.01) .10 (.02) .07 (.02) .05 (.02) .07 (.02) .06 (.02) .22 (.03) .25 (.02) .11 (.02)

228.42** 5.07** 3.30** 2.10** 2.91** 3.37** 7.80** 11.44** 5.70**

2.06 (.05) .07 (.19)

2 1.32** 0.35**

3.13 (.16) .18 (.06) .11 (.06)

19.83** 3.03** 1.90†

.11 (.06)

1.96**

**p # .01, *p # .05, †p # .10. Organizational support for service system was the omitted dummy variable.

psychological climate dimensions to Model 0 indicated that a large proportion (84.24%) of between-unit variance in satisfaction was accounted for by the set of psychological climate dimensions. The proportion was calculated by subtracting the between-unit variance in Model 1 from the between-unit variance in Model 0, divided by the betweenunit variance in Model 0 (Raudenbush, 2002). The remaining between-unit variance was still significant after taking the individual-level predictors into account, which indicates

.008

.221

2 0.69** 0.92** 13.40** 1.18 2 1.75† 2 2.28**

2.03 (.04) .23 (.25) 2.77 .07 2.10 2.13

(.21) (.06) (.06) (.06)

187.24** 23.26** 1.02** 13.48**

(.02) (.04) (.05) (.04)

3.28 1.01 .05 .54

t value

.010

.221

Coefficient (SE)

.051 .011

.345 .221

Level 2 residual variance

*p # .05, **p # .01, †p # .10. Organizational support for service system was omitted dummy variable for climate systems at both levels.

Model 0 – No predictors Model 1 – Individual-level climate systems (Level 1) Model 2 – Individual-level climate systems and Level 2 controls Model 3 – Individual-level climate systems, Level 2 controls, unit-level climate systems Model 3 Parameter estimates Level 1 Individual-level climate systems Constant Comprehensive Work environment support for service Average system Level 2 Controls Branch type Average agreement Level 2 Unit-level climate systems Constant Comprehensive Work environment support for service Poor

Level 1 residual variance

4.1

3.1

77.6

Level 2 variance explained by model (%)

Table 4. HLM of the effect of individual-level climate systems and unit level climate systems on satisfaction

3.5

4.5

12.90 4.7

Percentage of Level 2 to overall variance (%)

Copyright © The British Psychological Society

Reproduction in any form (including the internet) is prohibited without prior permission from the Society

660 Mathis Schulte et al.

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

661

that other higher-level variables (e.g. climate systems) might help explain the remaining variance. In Model 2, only control variables (branch type, average agreement on climates) were included in addition to the individual-level predictors. Adding the control variables in Model 2 over and above the individual climate dimensions (see Table 3) did not explain between-unit variance in satisfaction (2 .29%). A negative proportion in explained variance after including further predictors seems to be paradoxical. However, Raudenbush (2002) pointed out that this might occur if all added predictors are nonsignificant and their coefficients are rather small. This was the case in Model 2 for branch type (b ¼ 20:05; p ¼ :31) and the extent of agreement on climates (b ¼ 0:12; p ¼ :51). In Model 3, the set of dummy variables for the profile groups at unit level were added. The comparison of variance components between Model 2 and Model 3 revealed that an additional 2% of the between-unit variance in satisfaction was explained by the set of climate systems beyond the variance explained by the individual-level climate dimensions and the control variables. This small but significant increase in explained variance supports our Hypothesis 1 that branch-level climate systems explain variance in individual satisfaction above and beyond psychological climate dimensions. The overall proportion of explained between-unit variance was 86.36%. HLM procedures were also used to test Hypothesis 2 that higher-level climate systems will explain variance in satisfaction beyond that explained by psychological climate systems. Results of these analyses are presented in Table 4. Here, results indicated that the individual-level climate systems accounted for a significant proportion (77.60%) of variance in satisfaction (see Model 1). In addition, parameter estimates from a fixed effects model for each of the variables are contained in the lower portion of Table 4. Significant results were obtained for relationships between individual climate systems and satisfaction. These results are consistent with some prior work on the relationship between psychological climate systems and individual attitudes (e.g. Joyce & Slocum, 1984). Further, unit-level control variables explained 3.1% additional variance in satisfaction after controlling for individual climate systems (see Model 2). Finally, in support of Hypothesis 2, climate systems at the unit level of analysis explained an additional 4% of the between-unit variance in satisfaction after controlling for individual climate systems, branch type and agreement (see Model 3 in Table 4). Note that when individual climate systems were used to represent climate perceptions, unit climate systems explained slightly more additional variance in satisfaction than when psychological climate dimensions were used to represent climate perceptions (4% additional variance compared with 2% additional variance). Finally, we also tested whether the eight aggregate climate dimensions explained variance above and beyond the eight individual climate dimensions. That is, instead of the dummy variables for the climate systems, we added the eight aggregate climate variables in Model 3 of our HLM analyses. In this case, unit-level climate dimensions explained no incremental variance in satisfaction beyond that already explained by psychological climate dimensions and control variables. As noted above, the set of unit-level climate systems was significantly related to satisfaction above and beyond the individual climate dimension or systems. Thus, significant differences between the various unit-level configurations were examined to determine the types of climate systems that are significantly different from one another in their relationship to individual satisfaction (see Cohen, Cohen, Aiken, & West, 2003, p. 315, for details on calculating these coefficients). Different results were found depending on the conceptualization of psychological climate.

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

662 Mathis Schulte et al.

When individual climate dimensions were controlled for (see Table 3), the unit-level comprehensive, work environment support for service and poor climate systems did not significantly differ from one another in their relationship to satisfaction but all were significantly more positively related to satisfaction than the organizational support for service system (although the difference between organizational support and work environment support for service was only marginally significant). These results suggest that, after taking psychological climate dimensions into account, an employee tends to be less satisfied in a unit (or bank branch) where the team and managerial support dimensions of climate are relatively low. When individual-level climate systems were controlled for (see Table 4), the unit-level comprehensive service system fared significantly better than the poor climate system ( p , .01) and the work environment support for service ( p , .01). In addition, the organizational support for service system was significantly higher than the poor climate system ( p , .05) as was the work environment for service (although only with p , .10). In order to determine whether the effects of the unit-level climate systems change with the inclusion of individual-level dimensions or systems, a third HLM analysis was conducted using only unit-level predictors. Individual climate perceptions were not included in these analyses. In this case, the set of unit-level climate systems accounted for 84% of the between-unit variance in satisfaction after the variance due to control variables was accounted for (see Table 5). A comparison between Model 1 in Table 5 and Model 3 in Tables 3 and 4 reveals that the amount of explained variance in satisfaction due to climate systems dropped substantially when psychological climates were included in the model. That is, when psychological climate was included as a set of dimensions or as systems, most of the variance in satisfaction was explained by these

Table 5. HLM of the effect of unit-level climate systems on satisfaction

Level 1 residual Level 2 residual variance variance Model 0 – No predictors .346 .345 Model 1 – Controls and unit-level climate systems (Level 2) Model 1 Parameter Coefficient (SE) estimates Level 2 Controls Branch type 2.03 (.05) Average agreement .34 (.24) Level 2 Unit-level climate systems Constant 3.13 (.20) Comprehensive .28 (.05) Work environment support 2.13 (.05) for service Poor climate 2.31 (.05)

.051 .008

Level 2 variance Percentage of Level 2 to overall explained variance (%) by model (%)

84.33

t value

20.57** 1.41**

15.83** 5.43** 22.77** 26.05**

*p # .05, **p # .01. Organizational support for service system was omitted dummy variable.

12.90 12.90

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

663

individual-level perceptions and unit-level climate systems explained a small additional but significant amount of variance in satisfaction. Examining the differential relationships of the unit-level climate systems to satisfaction (when individual perceptions were not taken into account) revealed that comprehensive system (high across all climate dimensions) was significantly more positively related to satisfaction than the organizational support for service system, which, in turn, was significantly more positively related to satisfaction than the work environment support for service systems. Poor climate systems had the most negative relationship to satisfaction. Note that this pattern of relationships between unit-level climate systems and individual satisfaction was similar to what was found when individual climate systems were also controlled for. The rank ordering of the climate systems was the same, and the comprehensive and organizational support systems fared significantly better than work environment support and poor systems in both cases. In contrast, when individual climate dimensions were held constant, the organizational support for service systems tended to be related to lower individual satisfaction scores than other systems. Here, the inclusion of psychological climate dimensions changed the rank ordering of the climate systems to some degree as well as the significant differences between them.

Discussion This study was designed to examine the relative impact of psychological and unit-level climate on individual satisfaction and to explore whether an employee’s satisfaction is not only related to his or her own perceptions of the work environment but if systems of shared climate perceptions add additional explanatory power. Results from hierarchical linear models supported our hypotheses that unit-level climate systems account for significant variance in individual satisfaction beyond that accounted for by individual psychological climate dimensions and climate systems. The set of unit-level climate dimensions did not explain additional variance in satisfaction beyond that explained by the set of individual-level psychological climate dimensions. Taken together, the pattern of results suggests that contextual social domains of the environment are important for understanding individual attitudes in organizations beyond individuals’ own perceptions of the context, and further that examining higher-level climate as a system (as opposed to independent dimensions) may better capture the ‘Gestalt’ of climate. A number of authors have long argued that while both person and contextual variables are important for understanding individual responses and attitudes, the context should be defined and measured based on individuals’ perceptions because the context matters to the degree that it is perceptually filtered by the individual (e.g. Endler & Magnusson, 1976; Mischel, 1977). The results from this study also emphasize the importance of individuals’ perceptions in relationships to attitudinal outcome variables such as job satisfaction. More importantly, however, results suggest that social, contextual information plays a small but noteworthy role in explaining attitudinal responses. Unit-level climate systems accounted for 2% of the between-unit variance in individual satisfaction above and beyond that accounted for by psychological climate dimensions and accounted for 4% above and beyond that of psychological climate systems. While a 2% to 4% increase in criterion variance seems small, this finding needs to be considered in light of two issues that serve to attenuate this relationship. First, both individual-level and unit-level climate are expected to have shared a considerable amount of variance because of the compositional relationship (e.g. Chan, 1998)

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

664 Mathis Schulte et al.

between the two constructs. Unit-level climate is derived directly from individual climate. Further, the climate systems were based on the aggregated (mean) scores of the psychological climate scales within each unit. In order to justify aggregation, individuals within each unit (bank branch) must have some degree of similarity in their perceptions. As such, only a small portion of the unit-level climates can be statistically independent from psychological climates. The fact that the unit-level climate systems did produce a significant increment to an equation that already accounted for a very large percentage of variance provides some indication that social systems may add a small but unique and important effect that is not solely filtered through individuals’ own perceptions. Second, from a conceptual perspective, individual perceptions might already hold group influences. Perceptions of co-workers may have an influence on both individual attitudes as well as individual perceptions. For example, perceptions of people within groups tend to become more similar over time because of socialization processes (Salancik & Pfeffer, 1978). Because such group influences might already be incorporated in individual perceptions, even a small amount of explained variance due to unit-level climate can be considered significant. Moreover, collective employee satisfaction has been linked to unit and organizational performance (e.g. Ostroff, 1993; Ryan, Schmit, & Johnson, 1996). Thus, a better understanding of the group-level factors that influence satisfaction is important as even small increases in variance explained can translate into large bottom-line productivity or performance results. Future research is clearly needed to replicate these finding. It is likely that stronger results might emerge if different organizations, rather than independent branches within the same organization, were examined because between-group variability would likely be higher across organizations than within a single organization. Further, we proposed that by using climate systems instead of aggregated independent dimensions of climate, additional variance could be captured because the interrelations of the aggregate dimensions are taken into consideration. Results from this study provide some preliminary support for this assumption. After controlling for psychological climate dimensions, the aggregate dimensions did not explain a significant amount of additional variance in satisfaction. Significant levels of additional variance were only explained by unit-level climate systems. As noted by Schneider and his colleagues (Schneider, Bowen, Ehrhart, & Holcombe, 2000), ‘climate is a gestalt, a whole that is constructed and can be identified based on specific activities, behaviors, and experiences. Climate is thus inferred from the presence of parts : : :’ (p. 28). In our case, the focus on climate systems allowed us to focus on more than just an inference based on the parts of climate. By relating entire systems of climates to outcome variables, non-linear relationships among the climates were acknowledged. That is, variables that are found to be causally related in one system may be unrelated or even inversely related in another (Becker & Gerhart, 1996; Meyer et al., 1993). Hence, not only the shared perceptions but also the ‘Gestalt’ of the system was related to individual satisfaction above and beyond individual percepts of independent climates. Moreover, when examining the relative effects of individual- and unit-level climates on satisfaction, we used two alternative measures of psychological climate – individual-level dimensions and individual-level systems. Climate systems at the unit level accounted for more incremental variance after controlling for individual systems (4%) than after controlling for a set of psychological climate dimensions (2%). One explanation for this difference is that there is a smaller overlap or differences in how individual- and unit-level climate are interrelated at different levels of analysis.

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

665

The significant effect of unit-level systems over and above individual dimensions can be ascribed to the shared properties of the climate aggregates and the interrelatedness of climate systems at the higher level. While the same argument can be made for the incremental effect of unit-level climate systems over individual-level climate systems, in addition, individual systems are not necessarily commensurate with the branch system of shared climates. That is, some individuals within a branch may have psychological climate systems that are more similar to individuals in other branches than to individuals within the same branch, suggesting that how shared climates are interrelated in a bank branch is not necessarily congruent with how climates are interrelated at the individual level. It is also conceivable that some of the differences between individual- and unit-level climate systems may be due to the sampling error or statistical artifacts that resulted from using cluster analysis at both levels of analysis. Nevertheless, the results revealed that no matter how psychological climate was conceptualized and operationalized, unit-level climate systems explained additional variance in individual satisfaction. Overall, these results support the notion of emergent properties (Bliese, 2000; Firebaugh, 1978) implicit in the higher-level climate variables. That is, by aggregating and clustering psychological climates, the unit-level climate systems have properties that are partly independent from the individual-level psychological climates. A person is influenced in his or her affective responses by the shared perceptions of those with whom he or she interacts, even if he or she does not fully share their perceptions. Because organizational climate was conceptualized as a system of interrelated climates, it is the ‘Gestalt’ of the system that appears to be related to individual satisfaction above and beyond individual perceptions of climate. Although it was not the primary purpose of the study to delve into the nuances of differences between climate systems, particularly due the paucity of theoretical work in this area, it should be noted that some differences in the relative importance of the various climate systems were found depending on whether individual climate dimensions or systems were controlled for in the analyses. In particular, the comprehensive system and organizational support for service system tended to be more positively related to satisfaction than the work environment support and the poor climate systems when either no psychological climates were controlled for or when psychological climate systems were controlled for. In these cases, the poor climate was also lowest in the rank order, and was negatively related to satisfaction. This pattern of results, particularly that the organizational support for service system was significantly better than the work environment support for service system and the poor system, is consistent with the notion that at least some modest level of concern for employees and the internal work environment is necessary for the development of service climates (e.g. Burke, Borucki, & Hurley, 1992; Schneider et al., 2000). Climate systems that were strong in the amount of support for service provided by general practices in the organization (e.g. personnel support such as adequate staffing levels for service) and had at least a modest level of attention to internal work environment support (e.g. teams, managerial support) were generally superior to those where the emphasis was largely on attention to the immediate work environment but ignored the general organizational practices in support of service. Hence, it appears that attention to internal work environment and team issues alone is not sufficient, but more general support for service across the organization is also necessary. However, results for the unit-level climate systems differed when individual dimensions were controlled. In this case, the organizational support for service system

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

666 Mathis Schulte et al.

fared the most poorly and the poor climate system was significantly better than the organizational support for service system (there were no significant differences between the comprehensive, work support and poor climates). There are several possible reasons for this aberration compared with the other pattern of results (when no individual-level climates or when individual-level climate systems were controlled). First, these differences may be due to suppressor effects. Individual climate dimensions seem to act as a suppressor in the relationship between satisfaction and the organizational support for service climates because the coefficients for the poor climate system changed from negative to positive values suggesting. A suppression effect occurs if the zero-order correlation between a variable and an outcome is negative but the beta weight for this variable becomes positive after including another variable (the suppressor variable) in the regression equation (Conger, 1974; Krus & Wilkinson, 1983). The negative zero-order correlation between the poor climate systems and satisfaction, in contrast to its positive beta weights when individual dimensions are controlled for, supports the notion of a suppression effect. Here, the entire system of psychological climate variables may have acted as a suppressor (Tzelgov & Henik, 1991). A second possibility is that by relying on empirically derived individual-level climate systems, more error is created in capturing an individuals’ system and classifying individuals into a particular system because more variance within a branch is produced. As such, these individual-level systems produce less overlap with unit-level climate systems, which then allows for more additional and unique variance to be accounted for by the unit-level climate systems as well as for more similar patterns of results to emerge across the levels of analysis. Finally, it is possible, though probably less likely, that individual-level systems and a set of psychological climate dimensions capture qualitatively different aspects of psychological climate, and hence the additional effect of the social context (unit climate systems) produces different results beyond these individual effects. Clearly, much more research is needed in conceptualizing and measuring climate systems at both levels of analysis.

Limitations and future research It is important to note that the relationship between individual and branch-level climates is based on an additive model in this study. That is, the relative climate effects at the different levels of analysis are linear and combine to produce their overall relationship to satisfaction. However, interactions between psychological climates and climate systems at the branch level are also conceivable such that group-level variables might moderate how individual-level variables are related to outcome variables. For example, climate systems might not only be directly related to the satisfaction of an employee but might also influence the relationship between psychological climates and satisfaction. Although we tested for this possibility in the current sample and found no significant moderating effects, future research with different samples across different organizations might produce different results. Further, the relative effects of climates at different levels of analysis were examined only in relation to individual satisfaction. In addition, all measures were collected using the same survey, which could result in inflated correlations due to common-method variance. This is particularly likely in individual-level relationships between psychological climates and satisfaction (Carr et al., 2003). Hence, future

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

667

studies should consider employing alternate designs such as separating measures of the independent and dependent variables over time or using split-sample designs (Ostroff, Kinicki, & Clark, 2002), and should include a variety of different attitudinal and performance-related outcomes such as commitment, turnover and individual performance. Finally, future research is needed to explore climate systems to determine if such systems meaningfully exist with other measures, in other samples and other organizational settings. First, the interpretation of climate systems depends on the climate dimensions that are included in them. In this study, we used empirically derived climate scales that were based on a customized survey instrument. Although these dimensions have been identified as important in the literature, other dimensions or other measures of similar dimensions may have changed the systems and their interpretation. Future research needs to consider different climate dimensions and different measures. More work is needed that provides a conceptual framework for the selection of climates that should be integrated in systems. Existing taxonomies of climate (e.g. Ostroff, 1993) may be helpful in this context. Furthermore, we would expect that different samples could produce different systems depending on various characteristics of employees and organizations. Our sample was based in the service industry, particularly banking. Different types of climate systems are likely to be relevant for different industries. For example, the importance of social aspects in producing different systems in this study might be characteristic of the service industry where relationships between managers, employees, and customers play a decisive role (e.g. Schneider & Chung, 1996). In other industries (e.g. manufacturing), different prominent aspects might emerge in the climate systems. From a practical standpoint, the findings suggest that organizations should attend to the totality of the climate system, as opposed to focusing on single dimensions. Changing single aspects of the work environment, without attention to other related dimensions, may have unintended negative effects on individuals’ attitudes. In addition, by examining the Gestalt of the climate system from a configural perspective, organizational decision makers can target the most appropriate array of climate dimensions that are most important for fostering positive employee attitudes. Finally, the effectiveness of organizational climate above and beyond psychological climate emphasizes the importance of attention to social processes that may lead to sharing of climate perceptions and to the creation of a positive social environment at the unit or organizational level. Unit leaders are likely to play a major role in these processes because they can act as ‘climate engineers’ (Naumann & Bennett, 2000; Rentsch, 1990) and influence the overall climate system. Overall, this study begins to shed light on the effect of group-level perceptions on individuals’ cognitions and attitudes. Individuals may be influenced by group perceptions even if they do not fully share these perceptions. These results support the notion of pervasive group influences on individuals. Future research should address potential mediators and moderators of this effect. More specifically, which psychological mechanisms are responsible for these group effects beyond individual perceptions? Which organizational context factors foster or inhibit the influence of group perceptions on individuals? Although this study presents some promising initial results, further investigation of the influence of unit versus psychological climate is warranted.

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

668 Mathis Schulte et al.

References Aldenderfer, M. S., & Blashfield, R. K. (1984). Cluster Analysis. Beverly Hills, CA: Sage. Becker, B., & Gerhart, B. (1996). The impact of human resource management on organizational performance: Progress and prospects. Academy of Management Journal, 39, 779–801. Blau, G. (1995). Influence of group lateness on individual lateness: A cross-level examination. Academy of Management Journal, 38, 1483–1496. Blau, P. M. (1960). Structural effects. American Sociological Review, 25, 173–193. Bliese, P. D. (1998). Group size, ICC values, and group-level correlations: A simulation. Organizational Research Methods, 1, 355–373. Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and analysis. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research and methods in organizations (pp. 349–381). San Francisco: Jossey-Bass. Bowen, D., & Ostroff, C. (2004). Understanding HRM – firm performance linkages: The role of the strength of the HRM system. Academy of Management Review, 29, 203–211. Burke, M., Borucki, C., & Hurley, A. (1992). Reconceptualizing psychological climate in a retain service environment: A multiple stakeholder perspective. Journal of Applied Psychology, 77, 717–729. Carr, J. Z., Schmidt, A. M., Ford, J., & DeShon, R. P. (2003). Climate perceptions matter: A metaanalytic path analysis relating molar climate, cognitive and affective states, and individual level work outcomes. Journal of Applied Psychology, 88, 605–619. Chan, D. (1998). Functional relations among constructs in the same content domain at different levels of analysis: A typology of composition models. Journal of Applied Psychology, 83, 234–246. Cohen, P., Cohen, J., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences. New Jersey: Erlbaum. Conger, A. J. (1974). A revised definition for suppressor variables: A guide to their identification and interpretation. Educational and Psychological Measurement, 34, 35–46. Delery, J. E., & Doty, D. (1996). Modes of theorizing in strategic human resource management: Tests of universalistic, contingency, and configurational performance predictions. Academy of Management Journal, 39, 802–835. Denisson, D. R. (1996). What is the difference between organizational culture and organizational climate? A native’s point of view on a decade of paradigm wars. Academy of Management Review, 21, 619–654. Doty, D., Glick, W. H., & Huber, G. P. (1993). Fit, equifinality, and organizational effectiveness: A test of two configurational theories. Academy of Management Journal, 36, 1196–1250. Endler, N. S., & Magnusson, D. (1976). Personality and person by situation interactions. In N. S. Endler & D. Magnusson (Eds.), Interactional psychology and personality. New York: Hemisphere. Firebaugh, G. (1978). A rule for inferring individual-level relationships from aggregate data. American Sociological Review, 43, 557–572. Friedlander, F., & Margulies, N. (1969). Multiple impacts of organizational climate and individual value systems upon job satisfaction. Personnel Psychology, 22, 171–183. Gioia, D. A., Thomas, J. B., Clark, S. M., & Chittipeddi, K. (1994). Symbolism and strategy change in academia: The dynamics of sensemaking and influence. Organization Science, 5, 363–383. Glick, W. H. (1985). Conceptualizing and measuring organizational and psychological climate: Pitfalls in multi-level research. Academy of Management Review, 13, 133–137. Glisson, C., & James, L. R. (2002). The cross-level effects of culture and climate in human service teams. Journal of Organizational Behavior, 23, 767–794. Gonzalez-Roma, V., Peiro, J. M., & Tordera, N. (2002). An examination of the antecedents and moderator influences of climate strength. Journal of Applied Psychology, 87, 465–473. Hellriegel, D., & Slocum, J. W. (1974). Organizational climate: Measures, research and contingencies. Academy of Management Journal, 17, 255–280.

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

669

Herman, J. B., Dunham, R. B., & Hulin, C. L. (1975). Organizational structure, demographic characteristics, and employee responses. Organizational Behavior and Human Decision Processes, 13, 206–232. Ichniowski, C., Shaw, K., & Prennushi, G (1997). The effects of human resource management practices on productivity: A study of steel finishing lines. The American Economic Review, 87, 291–313. James, L. A. (1982). Aggregation bias in estimates of perceptual agreement. Journal of Applied Psychology, 67, 219–229. James, L. A., Demaree, R. G., & Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69, 85–98. James, L. A., & James, L. R. (1989). Integrating work environment perceptions: Explorations into the measurement of meaning. Journal of Applied Psychology, 74, 739–751. James, L. A., & Jones, A. (1974). Organizational climate: A review of theory and research. Psychological Bulletin, 18, 1096–1112. James, L. A., Joyce, W. F., & Slocum, J. W. (1988). Comment: Organizations do not cognize. Academy of Management Review, 13, 129–132. James, L. R., & Tetrick, L. E. (1986). Confirmatory analytic tests of three causal models relating job perceptions to job satisfaction. Journal of Applied Psychology, 71, 77–82. Johnson, J. J., & McIntye, C. L. (1998). Organizational culture and climate correlates of job satisfaction. Psychological Reports, 82, 843–850. Jones, A. P., & James, L. R. (1979). Psychological climate: Dimensions and relationships of individual and aggregated work environment perceptions. Organizational Behavior and Human Decision Processes, 23, 201–250. Joyce, W. F., & Slocum, J. W. (1984). Collective climate: Agreement as a basis for defining aggregate climates in organizations. Academy of Management Journal, 27, 721–742. Katz, D., & Kahn, R. L. (1978). The social psychology of organizations (2nd ed.). New York: Wiley. Ketchen, D. J., & Shook, C. L. (1996). The application of cluster analysis in strategic management research: An analysis and critique. Strategic Management Journal, 17, 441–458. Klein, K. J., Bliese, P. D., Kozlowski, S. W. J., Dansereau, F., Gavin, M. B., Griffin, M. A., et al. (2000). Multilevel analytical techniques: Commonalities, differences and continuing questions. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research and methods in organizations (pp. 512–556). San Francisco: Jossey-Bass. Klein, K. J., Conn, A. B., Smith, D., & Sorra, J. S. (2001). Is everyone in agreement? An exploration of within-group agreement in employee perceptions of the work environment. Journal of Applied Psychology, 86, 3–16. Kopelman, R. E., Brief, A. P., & Guzzo, R. A. (1990). The role of climate and culture in productivity. In B. Schneider (Ed.), Organizational climate and culture (pp. 282–318). San Francisco: Jossey-Bass. Kozlowski, S. W. J., & Klein, K. J. (2000). A multilevel approach to theory and research in organizations. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research and methods in organizations (pp. 3–90). San Francisco: Jossey-Bass. Krus, D. J., & Wilkinson, S. M. (1983). Demonstration of properties of a suppressor variable. Behavior Research Methods, Instruments, and Computers, 18, 21–24. Lewin, K., Lippit, R., & White, R. K. (1939). Patterns of aggressive behavior in experimentally created social climates. Journal of Social Psychology, 10, 271–299. Liao, H., & Chuang, A. (2004). A multilevel investigation of factors influencing employee service performance and customer outcomes. Academy of Management Journal, 47. Likert, R. (1967). The human organization; its management and value. New York: McGraw-Hill. Lindell, M. K., & Brandt, C. J. (2000). Climate quality and climate consensus as mediators of the relationship between organizational antecedents and outcomes. Journal of Applied Psychology, 85, 331–348. Mathieu, J. E., & Kohler, S. S. (1990). A cross-level examination of group absence influences on individual absence. Journal of Applied Psychology, 75, 217–220.

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

670 Mathis Schulte et al. McGregor, D. M. (1960). The human side of enterprise. New York: McGraw-Hill. Merton, R. K., & Kitt, A. S. (1950). Contributions to the theory of reference group behaviour. In R. K. Merton & P. F. Lazersfeld (Eds.), Continuities in social research: Studies in the scope and method of the American soldier (pp. 40–105). Glencoe: Free Press. Meyer, A. D., Tsui, A. S., & Hinings, C. (1993). Configurational approaches to organizational analysis. Academy of Management Journal, 36, 1175–1195. Mischel, W. (1977). The interaction of person and situation. In D. Magnusson & N. S. Endler (Eds.), Personality at the crossroads: Current issues in interactional psychology (pp. 333–352). Hillsdale, NJ: Erlbaum. Morgeson, F. P., & Hofmann, D. A. (1999). The structure and function of collective constructs: Implications for multilevel research and theory development. Academy of Management Review, 24, 249–265. Naumann, S. E., & Bennett, N. (2000). A case for procedural justice climate: Development and test of a multi-level model. Academy of Management Journal, 43, 881–889. Ostroff, C. (1993). The effects of climate and personal influences on individual behavior and attitudes in organizations. Organizational Behavior and Human Decision Processes, 56, 56–90. Ostroff, C., & Bowen, D. E. (2000). Moving HR to a higher level: HR practices and organizational effectiveness. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research and methods in organizations (pp. 211–266). San Francisco: Jossey-Bass. Ostroff, C., Kinicki, A. J., & Clark, M. A. (2002). Substantive and operational issues of response bias across levels of analysis: An example of climate-satisfaction relationships. Journal of Applied Psychology, 87, 355–368. Ostroff, C., Kinicki, A. J., & Tamkins, M. M. (2003). Organizational culture and climate. In W. C. Borman & D. R. Ilgen (Eds.), Handbook of psychology: Industrial and organizational psychology (Vol. 12, pp. 565–593). New York: Wiley. Parker, C. P., Baltes, B. B., Young, S. A., Huff, R. A., Altmann, R. A., Lacost, H. A., & Roberts, J. E. (2003). Relationships between psychological climate perceptions and work outcomes: A meta-analytic review. Journal of Organizational Behavior, 24, 389–416. Patterson, M., Payne, R., & West, M. (1996). Collective climates: A test of their sociopsychological significance. The Academy of Management Journal, 39, 1675–1691. Payne, R. L. (1990). Madness is our method. A comment on Jackofsky and Slocum’s paper, a longitudinal study of climates. Journal of Organizational Behavior, 11, 77–80. Payne, R. L., & Mansfield, R. (1973). Relationships of perceptions of organizational climate to organizational structure, context, and hierarchical position. Administrative Science Quarterly, 18, 515–526. Raudenbush, S. W. (2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Thousand Oaks, CA: Sage. Rentsch, J. R. (1990). Climate and culture: Interaction and qualitative differences in organizational meanings. Journal of Applied Psychology, 75, 668–681. Ryan, A. M., Schmit, U. J., & Johnson, R. (1996). Attitudes and effectiveness: Examining relations at an organizational level. Personnel Psychology, 49, 853–882. Salancik, G. R., & Pfeffer, J. (1978). A social information processing approach to job attitudes and task design. Administrative Science Quarterly, 23, 224–253. Schneider, B. (2000). The psychological life of organizations. In N. M. Ashkanasy & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. XVII–XXI). Thousand Oaks, CA: Sage. Schneider, B., & Bowen, D. E. (1985). Employee and customer perceptions of service in banks: Replication and extension. Journal of Applied Psychology, 70, 423–433. Schneider, B., Bowen, D. E., Ehrhart, M. G., & Holcombe, K. M. (2000). The climate for service: Evolution of a construct. In N. M. Ashkanasy & M. F. Peterson (Eds.), Handbook of organizational culture and climate (pp. 21–36). Thousand Oaks, CA: Sage.

Copyright © The British Psychological Society Reproduction in any form (including the internet) is prohibited without prior permission from the Society

Organizational climate systems

671

Schneider, B., & Chung, B. (1996). Service quality. In C. L. Cooper & D. M. Rousseau (Eds.), Trends in organizational behavior (Vol. 3, pp. 65–80). Chichester: Wiley. Schneider, B., & Reichers, A. E. (1983). On the etiology of climates. Personnel Psychology, 36, 19–39. Schneider, B., & Rentsch, J. R. (1988). Managing climates and cultures: A future perspective. In J. Hage (Ed.), Futures of organizations (pp. 181–200). Lexington, MA: Lexington. Schneider, B., Salvaggio, A. N., & Subirats, M. (2002). Climate strength: A new direction for climate research. Journal of Applied Psychology, 87, 220–229. Schneider, B., Wheeler, J. K., & Cox, J. F. (1992). A passion for service: Using content analysis to explicate service climate themes. Journal of Applied Psychology, 77, 705–716. Schneider, B., White, S., & Paul, M. C. (1998). Linking service climate and customer perceptions of service quality: Test of a causal model. Journal of Applied Psychology, 83, 150–163. Tzelgov, J., & Henik, A. (1991). Suppression situations in psychological research: Definitions, implications, and applications. Psychological Bulletin, 109, 524–536. Weiss, D. J., Dawis, R. V., England, G. E., & Lofquist, L. H. (1967). Manual for the Minnesota satisfaction questionnaire. Minneapolis: University of Minnesota, Industrial Relations Center. Received 1 April 2004; revised version received 2 September 2005