Multi-Level Organizational Research

Organization is central to human achievement. The most accomplished individuals tend to have supporting groups and structures, which have major roles in celebrated successes, though they may remain hidden from superficial views. Organizations date back to the dawn of civilization, with religious institutions and military forces being enduring examples.

Organizations are inherently multi-level (Klein, & Kozlowski, 2003). Each level is connected with and dependent on the others. The individual, teams, and the organization as a whole, are the 3 significant parts of a typical structure. Organizations, though they exist in kaleidoscopic varieties of purposes, sizes, and natures, share common issues when they become subjects of systematic inquiry. Performance and effectiveness are the two most important motivations for organizational research, though the interactions between individuals and groups are matters of primary interests in fields of sociology and psychology.

This document reflects on the inevitable implications of multiple levels in organizations, and suggests best practices with respect to studying such structures and the people who function within them. The focus is on how to incorporate multi-level realities in to organizational research,

Stating the Obvious

It is common knowledge that organizations are made of individuals, and that groups of people have to function as teams. However, the implications of the obvious on how organizations should be studied and measured are complex and certainly not as obvious. It may be tempting to resort to over-simplification and ignore multi-level aspects of organizations when designing research projects to study them.

Interplay between individual perceptions and organizations are common to all types of organizations. Multi level variations have been found even in the prison structure (Camp, et al. 1997), where institutional operations and satisfaction with immediate supervision have equal influences on the social climate. The implication is that any research in to the functioning and performance of a large structure with branches at multiple levels has to incorporate distinct phases of research for individuals at the periphery, for teams that operate within the body of structures, and for the entire organization as an entity.

Each of these levels will be in dynamic inter-play with the other two. A phenomenon at a lower level, for example,  may not connect with theory validation at a higher level (Crowston, 2002). There are many examples of technologies being integrated in to organizations at lower levels, without incontrovertible evidence of the benefits at higher levels. That is why conclusions from research at a single level may not yield entirely relevant conclusions.

The best individual technical skills may come to naught if a person cannot deal with others; similarly, the best teams cannot function without adequate resources and support from above (Loo, 2003). No level of organization can be independent, take all the credit, or be assigned all of the blame. Investigative research, which aims to diagnose why an organization functions below par, or which seeks to make recommendations for improvement, has to validate its suggestions for feasibility and appropriateness at each of the multiple levels.

Past organizational research has focused more on the individual level, and not enough on teams and the organization as a unit, or on the interactions between these levels (Schnake, and Dumler, 2003). However, this trend has begun to change and researchers of today can no longer the multiple level structures of organizations they may wish to study. Organization science needs to match the integration which marks literature on the existence of multiple levels (Klein, & Kozlowski, 2003).

Mixed level research needs methods and measurements of their own (Schnake, and Dumler, 2003); the levels of data collection and analysis are often not the same.  Pluralistic ignorance, in which each individual has a special perception of the self versus that of others and of the organization, commonly plagues organizational research. That is why projects should not be based solely on the surveyed and perceived opinions of individual members about their peers, superior, subordinates, and groups.

Bottom up models which draw inferences from lower levels for the higher will yield different conclusions if a top down approach is used (Klein, & Kozlowski, 2003). Research methods must account for how perspectives change with levels. It may be best to adopt an iterative approach, thinking not micro or macro, but both micro and macro (Klein, & Kozlowski, 2003).

Research Objectives as Determinants

Given that multiple levels are ubiquitous in organizations of all types and sizes, all research in this area should take the phenomenon in to account. Organizational research may vary by objectives, and this factor of difference can help in dealing with multiple levels.

The People and Process elements of the Marketing Mix for Services (Payne, 2002) play important roles in determining strategies followed by organizations which do not have concrete or tangible product features in their offers. Research in to the internal workings of such organizations may focus on the lower levels of individual capabilities, and small team functioning, rather than concern itself with organization-level matters. Conversely, stock market analysts who are concerned with specific financial outcomes may prefer to dwell on effectiveness of organizations as a whole (Huber and Glick, 1993), rather than bother with details of issues at micro levels.

Downsizing and new designs are some major concerns of contemporary organizational research. These phenomena create most strains on the individual (Huber, and Glick, 1993) and hence projects which seek to study the effects of integration and different hierarchies should focus on the lowest level of individual members of staff.

The simplistic approach of focusing on just one level of organization will not work in all cases. Communication issues, matters related to diversity in the work force and matters concerning global organizations, all require work to be done at all levels (Huber, and Glick, 1993). Social climate studies also have to take all levels in to account, since institutional initiatives can have variations at its spatially separated sites, and individual variations as well with regard to job satisfaction and organizational commitment (Camp, et al. 1997).

Some of the research problems of multiple levels in organizations arose in the past because of the paucity of prior work in similar areas. However, there has been a cascade of organizational research in recent times, with multiple level enquiries (Huber and Glick, 1993). This new body of work can be used as templates when fresh research projects in organizational behavior are undertaken.

The Effects of Organizational Labeling

Organizations tend to fit in to stereotypes which have been created in public minds over time. Professionally managed corporations, political parties, religious institutions, and bureaucracies, are a few examples of such set patterns of our collective minds. Simultaneously, organizations are more than just people who control or work in them. Suppliers, financiers, regulators, franchisees, and customers are some of the other categories of stake holders.

All organizations do not have to be equally responsive to the environment (Huber and Glick, 1993). Monolithic, powerful, and strongly entrenched organizations may demand research of how to further their aims, but worry less about the proclivities of employees and suppliers of goods and services at the bottoms and fringes of their hierarchies.

A highly evolved organization will be dynamic and open minded: it deserves and needs organizational research at multiple levels, with plenty of iteration, and the objective validation of hypotheses. Others will be directive and in a hurry for results: they may appreciate the values of enquiries at multiple levels less or not at all.

Organizational research driven by purely scholastic intent is a rare privilege. The professional researcher will more often have a restrictive brief set by a paying client with a specific and selfish perspective. Organizational studies may mention all the levels in passing or for the sake of completeness, rather than address each of them in detail and in concert.

The multiple levels within organizations should always be kept in mind when studying structures and group working. Even though studies may be sponsored by the most apical level of organization, valid findings and scientific rigor require that the study extends to teams and to individuals as well.

Towards Best Practices

Multiple levels are inevitable in organizational studies. How can the complexities of this reality be incorporated for better research project design, and to ensure findings on which sponsors can rely? It is best to start with the environment in which the organization operates (Huber, and Glick, 1993).

Such a prior effort will put research fully in context. It is necessary to spend time to collect primary data on how the organization functions; and to focus on processes which link the various levels (Crowston, 2002). It is risky to plunge headlong in to researching an organization: every investment in understanding its needs and to profile it accurately, will pay off in terms of relevant findings on which action can be taken.

Levels in organizations can be conceived in terms of internal customers. This is a standard concept in Services Marketing (Payne, 2002). A higher level in an organization, or a branch of the same level, is a customer of a team which supports it through its function.

Thus, a Sales Manager is a customer of his or her sales people, and the entire sales function is a customer of the production or manufacturing department. Since dealing with all levels can make organizational research very complicated, a workable option is to use the internal customer concept to delineate levels of research.

Another approach is to set up microcosms of the organization in external settings, to act as laboratories for experimentation. Groups of employees may be placed on extended sabbaticals, and asked to play roles equivalent to the levels of their parent organizations. This approach has the advantage of eliciting more objective input data from people inside the organization, but can be expensive and time consuming. Research projects which adopt this ‘laboratory’ approach require patience and sustained support, but yield better and more applicable findings at the same time.

A relatively quick and simple compromise could be to adopt a problem-solving approach (Loo, 2003). The median level is a good optimum in such cases, with a focus on actual teams from the structure, with additional participation by representatives from other levels and branches. This team approach can use direct and indirect benchmarking to support its recommendations.

Wherever a choice is available, the organizational researcher should try to go as micro as possible as uncontrollable factors and complexities keep increasing as the analysis goes to higher levels (Camp, et al. 1997). Aggregate measures tend to mislead, and reduce the chances of useful and valid findings.

Regardless of how organizational research is approached, a common requirement is to specify qualitative organizational performance measures as closely as possible. What for example is motivation or how will supervision be assessed?

This will also address the issue of phenomena at lower levels not being confirmed by theories at higher levels-perhaps the higher level has not specified the benefits it seeks (Crowston, 2002). Such specification will also reduce the pitfalls of data from a lower level from pluralistic ignorance, affecting analysis of a higher level (Schnake, and Dumler, 2003).

Finally, the organizational researcher must always think of multi-level implications of proposed work and suggestive conclusions whether or not all levels are included in the work (Klein, & Kozlowski, 2003). Such an approach should persist throughout the research, including the critical stage of sampling.

Conclusions

While multiple levels in organizations and their influences cannot be denied, including all of them comprehensively in actual projects may present.  It may often be best to make research manageable with clear statements of assumptions and limitations.

Organizations are in flux in any case and will evolve towards the median level. This is the historic trend (Huber, and Glick, 1993). Excellence and thoroughness in studying the team level may be a good compromise.

The perspective of each category of stake holders is distinct, yet valid. The researcher would do well to keep the expectations and profile of his or her customer in mind! Multiple levels, in the end, are integral to all significant organizations, and all research in to such groupings must take these multiple levels in to account.

References

Camp, R. et al. (1997) Aggregating Individual-Level Evaluations of the Organizational Social Climate: A Multilevel Investigation of the Work Environment at the Federal Bureau of Prisons. Justice Quarterly 14: 4: 739-762.

Crowston, K, (2002) Process as theory in information systems research, Proceedings of The IFIP WG 8.2 International Conference: The Social and Organizational Perspective on Research and Practice in Information Technology, 10–12 June 2000, Aalborg, Denmark.

Huber, G, and Glick, W, (1993) Organizational Change and Redesign: Ideas and Insights for Improving Performance, Oxford University Press

Klein, K & Kozlowski, S, (2003) A Multilevel Approach to Theory and Research in Organizations: Contextual, Temporal, and Emergent Processes, Chapter 1 in Multilevel Theory, Research, and Methods in Organizations: Foundations, Extensions, and New Directions, JOSSEY BASS

Loo, Robert (2003) A multi-level causal model for best practices in project management, Benchmarking, Vol10, Issue 1; pg. 29, 8 pgs

Payne, A, (2002), The Essentials of Services Marketing, Prentice-Hall

Schnake, M. and Dumler, M. (2003) Levels of measurement and analysis issues in organizational citizenship behavior research, Journal of Occupational and Organizational Psychology. 76(3):283