Cohre Home   

Reinventing the Workplace:
How Business and Employees Can Both Win:  Measuring
the Links Between Participation and Performance

by Dr. David I. Levine

Factors That Understate the Participation-Performance Relationship
Factors That Overstate the Participation-Performance Relationship
Reinventing The American Workplace Home

Factors That Understate the Participation-Performance Relationship

When you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind.
-Lord Kelvin

Determining the true effect of work practices such as employee involvement on productivity is difficult. Measurement problems can lead researchers to find no relationship when one does exist or vice versa.

  • Measuring work practices such as participation is difficult because plans differ according to the level of participation (frontline employees versus representatives), the number of people involved, the topics covered, the power the employees have on each topic, and the gap between official and actual involvement and empowerment. Measuring most forms of performance, such as productivity or worker satisfaction, also is problematic. The measurement of both work practices and performance typically leads to the estimated effect of a work practice being less than the true effect.
  • Many samples of organizations are small, making the detection of any effect difficult. Work practices are most useful in combination (for example, participation plus gainsharing plus training). If the researcher does not specify the interrelationships correctly, a true relationship may not appear. The effects are particularly difficult to detect in small samples, because testing for all possible combinations is hard.
  • Failing to control for relevant aspects of the environment or context can lead to understatement of the importance of employee involvement. For example, if unionized companies are likely to adopt high involvement practices when the company is in trouble, then employee involvement will become correlated with troubled companies.
  • Mismeasurement of the lags between the upfront costs of implementing a new program (training, for example) and the benefits biases researchers against finding any true effects.

back to top

Factors That Overstate the Participation Performance Relationship

  • Successful participatory schemes are more likely to be reported in the empirical literature than are failures. For example, Richard A. Guzzo, Richard D. Jette, and Raymond A. Katzell found in their meta-analysis of studies of changes in work organizations that studies with better designs (for example, experiments with true control groups) revealed smaller effects than studies with other designs (for example, before-after comparisons).(1) The latter forms of research are less rigorous and expensive, and they are probably rarely published when not successful.
  • Most samples of companies that adopt a work practice will disproportionately include those who find it most useful. Thus any correlation between work practice and performance does not predict what would happen to additional companies that adopt that work practice. Because companies differ, no single true effect of a work practice can be measured.
  • Failing to control for relevant aspects of the environment can lead to overstatement of the importance of employee involvement. Prosperous companies could (for example, because of a technological advantage) share their good fortune with workers via employee involvement or profit sharing and also have high productivity and profits. In this case, employee involvement would falsely appear to raise performance.
Note:  In response to Lord Kelvin, economist Frank Knight reportedly said, "Yes, and when you can express it in numbers your knowledge is of a meager and unsatisfactory kind." 1. Richard A. Guzzo, Richard D. Jette, and Raymond A. Katzell, "The Effects of Psychologically Based Intervention Programs on Worker Productivity: A Meta-Analysis," Personnel Psychology, vol.38 (Summer 1985), pp.295-91.

back to top

© Institute for Research on Labor and Employment, University of California, Berkeley. All rights reserved.
  Last update 2/9/99.