The following is an example of a field experiment. The management of a large company is interested in determining the impact that a four day work week would have on employee absenteeism. To be more specific management wants to know if employees working four 10-hour days have lower absence rates than similar employees working the traditional five day week of 8 hours each day. Because the company is large, it has a number of manufacturing plants that employ essentially similar work forces. Two of these are chosen for the experiment, both located in the greater Cleveland area. Obviously, it would be appropriate to compare two similar sized plants of one is in rural Mississippi and the other is in urban Copenhagen because factors such as national culture, transportation and weather might be more likely to explain any difference found than changes in the number of days worked per week.
In one plant, the experiment was put into place – workers began the four day week. At the other plant, which became the control group, no changes were made in the employees’ five day week. Absence data were gathered from company’s records at both locations for a period of 18 months. This extended time period lessened the possibility that any results would be distorted by the mere novelty of changes being implemented to the experimental plant. After 18 months, management found that absenteeism had dropped by 40 percent at the experimental plant, and by only 6 percent in the control plant. Because of the design of this study, management believed that the larger drop in absences at the experimental plant was due to the introduction of the compressed workweek.
The field experiment is similar to the laboratory experiment, except it is conducted in a real organization. The natural setting is more realistic than the laboratory setting, and this enhances validating but hinders control. In addition, unless control groups are maintained, there can be a loss of control if extraneous forces intervene – for example, an employee strike, a major layoff or corporate restructuring. Maybe the greatest concern with field studies has to do with organizational selecting bias. Not all organizations are going to allow outside researchers to come in and study their employees and operations. This is especially true of organizations that have serious problems. Therefore, since most published studies in OB are done by outside researchers the section bias might work toward the publication of studies conducted almost exclusively at successful and well managed organizations.
Our general conclusion is that, of the research designs the field experiment typically provides the most valid and general findings and except for its high cost, trades off the least to get the most.
What’s the overall effect of organizational behavior modification (OB Mod) on task performance? There have been a number of field experiments that have sought to throw light on this question. Unfortunately, the wide range of effects from these various studies makes it hard to generalize.
To try to reconcile these diverse findings, two researchers reviewed all the empirical studies they could find on the impact of OB Mod on task performance over 20-year period. After discarding reports that had inadequate information had non-quantitative data, or didn’t meet all conditions associated with principles of behavioral modification, the researchers narrowed their set to 19 studies that included data on 2,818 individuals. Using an aggregating techniques called meta analysis the researchers were able to synthesis the studies quantitatively and to conclude that the average person’s task performance will rise from the 50th percentile to the 67th percentile after an OB Mod intervention.