“Do the Right Thing:” Diverging Effects of Accountability in a Managerial Context

The need to justify one's decisions is a signal characteristic of decision making in a managerial environment. Even chief executives must communicate reasons for their actions. Yet, despite a significant amount of laboratory research on the effects of accountability on decision making, few studies have attempted to assess what affects accountability might have outside the lab for actual managers. In this paper, we use as subjects actual members of the professional account, research, and creative staffs of several advertising agencies in an experimental simulation of an advertising copy meeting. We demonstrate that accountability effects in complex, managerial decision contexts diverge considerably from those found in the lab. In particular, we find that accountability does not always lead managers to emphasize the most “objectively” accurate information, but instead can also encourage emphasis on information which is socially acceptable to the manager's key “constituencies” (those to whom they feel accountable, see Tetlock 1991). The social constituencies of accountable managers can differ significantly. For example, members of an advertising agency's account service or research departments are likely to favor the quantitative, analytical decision styles of their supervisors, while members of the same agency's creative department are likely to favor empathy and personal judgment. Requiring managers with different constituencies to justify their responses can cause “policy bolstering,” the tendency to favor a process of decision making particular to one's own constituency. Thus accountability may sometimes make it more difficult for a diverse group to reach a consensus, and may not have the desired positive impact on decision quality. In an experiment simulating an advertising copy meeting, we first solicited managers' private judgments of 12 ads, then presented them with research data and asked each of them to make numerical predictions of how much consumers would like each ad. The experiment manipulated the diagnosticity of the available research and the degree to which managers expected to be held accountable for their predictions. We used hierarchical linear regression to analyze the results. Hierarchical linear regression allows us to estimate simultaneously a within-subjects model capturing individual weighting coefficients for private judgment and research, and a between-subjects model of the effects of department membership and experimental conditions on these judgment-policy weights. This analysis also controls for the varying degrees of precision in the within-subjects estimates, as well as for correlations between the within-subjects weighting coefficients. . 1) Accountability had different impacts on the weighting schemes of members of account services and creative departments. Accountable research/account staff weighed research more in making their predictions than nonaccountable research/account staff. Accountable creatives, on the other hand, actually weighed research less than did nonaccountable creatives. 2) Because accountability caused subjects to attend assiduously to the research, the weight given private judgment reflected whether the data was diagnostic of the problem at hand. Accountable subjects presented with low-diagnosticity research reacted by emphasizing private judgment more heavily. 3) Accountability increased confidence in predictions; this effect was not dependent on the accuracy of subjects' predictions. . We provide several suggestions and observations regarding current managerial practices, which may improve the impact of individual accountability on decision making.