Last weekend’s Wall Street Journal carried an article by Safeway’s CEO, Steven A. Burd, addressing the national health-care reform debate. Safeway has been able to hold the per-capita health-care costs flat since 2005, while for most American companies these costs have increased by 38% during the same period.  Mr. Burd attributed this to a program that Safeway designed in 2005 to reward healthy behavior.

In this post, I focus on the following:

• The use of contingent rewards at Safeway.
• Carrot or stick? Does it make a difference?
• What does research say about rewards and punishment?
• Does rewarding or punishing individuals make a difference at the team level?
• What does it mean for virtual teams?

The use of contingent rewards at Safeway
The Safeway program emerged out of a realization that about 70% of all health-care costs are the direct result of behavior and, therefore, preventable. Another realization was that cardiovascular disease, cancer, diabetes and obesity cases constitute 74% of all health-care costs and majority of these cases are preventable. To motivate healthy behaviors, Safeway instituted discounts in health-care premiums and tied these discounts to tobacco usage, healthy weight, blood pressure, and cholesterol levels.The discounts have helped Safeway build personal accountability for health among its employees. This accountability has created a culture of health and fitness within Safeway.  According to Mr. Burd, Safeway’s obesity and smoking rates are roughly 70% of the national average. Employees have reported better health as a result of the program and are asking for more rewards.

What caught my attention in the Safeway article was that contingent rewards were used to bring about change in people’s behaviors. Employees are tested for tobacco usage, weight, blood pressure, and cholesterol levels and receive discounts off a “base level” premium for each test they pass.  So rather than punishing employees who don’t take care of their health by adding to their premiums, Safeway rewards those who pass the health tests by providing them with discounts.

Carrot or Stick? Does it make a difference?
Does it make a difference in the ultimate behavior shown by people when we reward good behavior or punish bad behavior? If my teenage children provide any indication of how others would respond, it seems that rewards would be more powerful than punishment. As a scientist, however, I wanted more robust evidence, for which I turned to past research.

There is a fair amount of research on this topic. I focused on research about rewarding and punishing behaviors by leaders. There were two notable studies by Philip Podsakoff and his colleagues that I found extremely useful. One of them examined the effects of a leader’s reward and punishment behaviors on group processes and productivity and the other reported an analysis of several studies on this topic.

What does research say about rewards and punishment?
The general finding is that both leader reward and punishment behaviors have a significant relationship with a variety of employee attitudes, perceptions, and behaviors. Contingent rewards, or rewards that are contingent upon appropriate task behaviors, reduce role ambiguity and improve employee satisfaction, effort, conscientiousness and performance, among other things. The rewards may be in the form of merit increases, bonuses, praise, commendations, and social approval. Contingent punishment, or punishment that is contingent on the absence of appropriate task behaviors, too reduces role ambiguity and increases employee satisfaction, although to a smaller extent than contingent rewards. Additionally, contingent punishment reduces employee conscientiousness and performance. Punishment may take the form of negative feedback, low or no merit increase, withholding of commission, dismissal, and social disapproval.  These results suggest that contingent rewards are superior to contingent punishment.

Part of the superiority of contingent rewards stems from the fact that for contingent rewards to work, a leader needs to clarify behaviors that would lead to superior performance and, hence, rewards. This clarity enables the “appropriate” behaviors. On the other hand, contingent punishment simply signals what is undesirable but does not clearly indicate the effective behaviors that the leader expects from the employee.

Does rewarding or punishing individuals make a difference at the team level?
What I found interesting in the papers is that when a leader displays reward and punishment behaviors with individual members in a group, the group’s processes and performance are also affected.  Contingent reward behavior has been found to promote group drive, cohesiveness, and productivity. Contingent punishment behaviors also have been found to be beneficial by promoting group drive and productivity but their effects are weaker than those of contingent reward behaviors.

It is hypothesized that the effects of contingent reward or punishment behavior on group drive, cohesion, and performance may be occurring, in part, due to positive effects on individual worker attitudes and performance, as indicated above. The effects on the group may also be occurring due to most members seeing that the leader is equitable and seeks to improve the performance of everyone in the group. These members may themselves encourage poorly performing members to straighten up. It could also be argued that when a leader employs contingent punishment, the leader may not only be identifying unproductive behavior for the individual receiving the punishment but also for other group members, who then attempt to avoid such behavior.

What does it mean for virtual teams?
What research seems to suggest is that virtual team leaders would benefit from clarifying role and task expectations to team members and rewarding team members on the basis of their performance. While research suggests contingent punishment could also have beneficial effects, leaders would help their teams more by focusing on contingent rewards. Leaders should also note that if they simply reward performance but don’t take the time to clarify role and task expectations, the rewards will most probably lose their efficacy. Since members of a virtual team may not have enough opportunities to share their experiences with one another and, therefore, learn from each other about their leader’s rewarding behaviors, leaders may want to make sure that the whole team is aware of their rewarding behaviors.