Projects

Workplace Transformation

Project Description

Previous research on both the determinants and the consequences of various HRM practices has suffered from several methodological limitations. First, as noted above, research has often been couched at the level of the firm, producing a wide gulf between corporate strategy and its implementation at the local, operating unit of analysis. This limitation is particularly problematic for the pulp and paper industry, whose mills are often in remote rural areas and where the linkage between corporate managers and the operating units requires particular attention.

Second, previous research has typically relied on single-respondent surveys to represent the HRM practices of whole firms, introducing serious problems of measurement error into the analysis. Even at the mill level, HR managers are often in the dark concerning actual mill practices and routines, and thus respond on the basis of assumptions that may or may not be grounded in practical realities. Third, research has often relied on cross-sectional research designs that force researchers to make a number of strong assumptions regarding the causal ordering of key variables (e.g., that workplace innovation "causes" strong financial performance, rather than the reverse). Finally, researchers have only rarely addressed the ways in which workers respond to or perceive HRM innovation -a fact that ironically reflects the very management-centered focus that workplace change claims to leave behind (see Berggren 1992). Especially in so traditional an industry as pulp and paper manufacturing, where workplace customs are often deeply institutionalized, it will be vital to understand workers' responses to new HRM practices and the factors that give rise to them (Appelbaum et al 2000).

The present research moves beyond these methodological limitations by using a multi-stage, multi-level analysis based on both survey and observational data. In stage 1, we aim to construct a random sample of 200 pulp and paper mills, stratified by region, mill size, and unionization (organized/unorganized). The sample will be limited to three major product categories in the US industry: brown paper and paperboard, white paper (including communications papers such as xerographic paper), and tissue. The mills will be identified using directories available through various trade publications. We anticipate gaining the support of the major associations within the industry, such as PIMA, AF&PA, and PACE, which will greatly aid questions of both access and response rate (see Appendix A). Prior to survey development, we will conduct site visits to the corporate headquarters of two firms, and to operating facilities selected to represent two broad types of mills: traditional, brownfield mills on the one hand, and newer, greenfield sites on the other. Communications with PACE staff has already begun to help focus in on candidate mills for such field work.

After conducting our site visits, our survey analysis will begin with the construction of a stratified random sample of 200 mills, based largely on information published in Lockwood-Post directories. Letters will be sent to the HR managers at each of the 200 mills in our sample, followed by telephone communications to elicit their support. Our stage one survey will include three respondents at each sampled mill: the HR manager, the production manager, and the local union president (where applicable). Each respondent will then be mailed a self-administrated questionnaire. Questionnaires will gather data on the prevalence of particular types of work systems (self-directed teams, High Performance Work systems, team concepts, TQM or SPC, etc.), the timing of such innovations, the success or failure of employee involvement efforts, as well as the character of labor relations, the out-sourcing of hourly worker functions (such as maintenance work), and authority relations within each mill. We will also include some semi-structured items tapping the respondents' perceptions regarding the sources of resistance to workplace change, whether from the management or the labor side. Mill performance data will be accessed from a service such as Fisher, Inc. (Fisheri.com) which specializes in the collection and dissemination of organizational information in the pulp and paper industry.

Past research on HR practices and firm performance has been criticized by Gerhart (1999) because survey responses were obtained from a single respondent at a single point in time. This approach does not allow the researcher to estimate statistically the reliability of the measurement. The lack of reliability estimates then limits the researcher's ability to statistically 'correct' for unreliability when estimating the various model parameters that link HR practices to firm performance. Our proposed design will provide multiple respondents and we will be able to compute interrater reliability estimates such as intraclass correlation (ICC).

Furthermore, as researchers we must be concerned with the possibility that the responses given by individuals describing HRM practices and other organizational characteristics can be influenced by person and role-specific biases. That is, an individual who is enthusiastic about nontraditional work systems may unconsciously 'over-report' the presence of them. It is impossible to design questioning strategies to overcome this possibility. However, by having a large sample and respondents with very different roles in the mills, we can better cancel out any particular bias.

Additionally, we are conscious of the increased opportunities for such bias when questions such as "was the employee involvement program successful?" or "what was the major source of resistance to change?" In our preliminary field work, we will work with members of the respondent population to create questioning strategies that are less open to attributional, defensive, or politically driven responding.

After gathering these mill-level survey data, our data collection strategy is to move upward in the sampled organizations, contacting HR managers within the mill's parent firm. We will distribute a detailed survey questionnaire to HRM offices, gathering firm-level data on the proportion of employees who are involved with particular types of HR practices, the structural position held by HR officers within the firm, the relationship between the parent firm and the mill, as well as the company's competitive strategy and market position within the industry.

These data will allow us to test hypotheses concerning both firm and mill-level variations in the development of workplace change.

In the second stage of the study, our focus will focus on the outcomes or consequences of non-traditional work systems at the micro level, using a smaller sample of mills constructed on the basis of our stage 1 surveys. This micro level of the analysis will in turn be divided into two facets: one involving data on mill, departmental, and production crew performance, and a second involving the job rewards that accompany differing types of HR practices. Through site visits to each mill, we will select distinct production areas or departments for focused attention. Monthly data on performance will be gathered, reaching back for a 5-year period of time (or longer, insofar as is possible). Having data on temporal fluctuations in departmental performance, the introduction of innovative work systems and new process technologies, while also holding constant the age of the equipment and other relevant variables, will enable us to draw robust inferences concerning the impact of particular HR practices on performance over time. At the same time, to ensure that we can address outcomes that are most relevant to hourly employees as well as management, we will survey individuals whose jobs 1) directly involve machine operation (e.g. production crews in each of the sampled departments), 2) support the on-going operations (e.g. maintenance), and 3) involve design of the technical systems through which the work is done (e.g. relevant engineering and management staff). Doing so will enable us to understand the workers' responses to particular changes in their working environment, and to document any changes in job characteristics and rewards that seem to accompany the introduction of non-traditional work systems over time. The survey of hourly workers will be particularly helpful in understanding the character of workers' jobs, their relationship to other staff outside the immediate work group, the opportunities they perceive to use their skills on the job, as well as the attitudinal complexes they bring to bear on workplace change.

Eliciting open and trusting responses for hourly workers can be a challenge. However, we maximize our chances here by developing the instrument in conjunction with hourly worker representatives and union members (where relevant). We will need to establish communication with the local union at each site and work out the goals of the research and how it will be implemented. Local union acceptance of the research is very important. In addition, we anticipate that a research team member will have spent some time on-site prior to administration in order gain some credibility with the hourly workers. During administration, it will be important to have a research team member on site who will be clearly different from management and union. The survey instruments will have minimal identifying information associated with them. The handling of survey documents must proceed from the hourly respondents to the research team member to a secure location such as the trunk of the research team member's car.

This research represents an extensive effort to understand many economic and organizational issues underlying the implementation and effectiveness of nontraditional work systems and HR practices in the pulp and paper industry. The results should be particularly helpful to managers, workers, and labor union officials in the industry because they will provide guidance on how to adapt both firm-level practices and mill-level practices to the new competitive realities facing the industry. One of the advantages of such a concentration within the industry is that the findings and interpretations can be well-grounded in the operations, traditions, and language of the industry. We plan to publish and distribute regular reports to participating companies, trade associations, and unions. We also plan to make presentations at trade association and union meetings where possible in order to facilitate the dissemination of the findings and to allow for greater interaction with the users of the information. We also plan to prepare articles for publication in industry and union trade journals.

On the academic side, this research will represent a significant contribution to the existing knowledge base about how work redesign and related HR practices become manifested in organizations, take hold, and affect both organizational effectiveness as well as important worker-related outcomes. We will prepare reports for both presentation at professional academic conferences as well as academic journals. Because of the interdisciplinary research team, we expect that our research will be published in both sociology journals and management journals. Likewise, research results will be presented at professional meetings for each discipline.

One other possible outcome is a book along the lines of Appelbaum et al. (2000), Such a more extended written account of the research would allow us to tie many different facets of the study together in one volume making it easier for others to access and use for future research. The decision about whether and/or when to write and publish such a book will depend on the progress and findings during Stage 2 of the project. Additional benefits from the project will be that it provides the opportunity to develop the research skills and industry knowledge of 2 Ph.D. students at Georgia Tech. The range of activities and research design issues that will be considered will provide fertile ground for developing deeper understanding by the students. One of the students, Raul Necochea, already has extensive industry experience which will further facilitate our efforts. In addition, because this project represents a significant cooperative effort between two distinct colleges at Georgia Tech, it allows us to build stronger linkages between the Ivan Allen College and the Georgia Tech College of Management. The interdisciplinary nature of the research team (both faculty and students) provide an excellent combination of skills, knowledge, and perspectives. This will aid the comprehensiveness of the project as well as its impact.

Duration: 3 years