Affect and managerial performance: a test of the sadder-but-wiser vs. happier-and-smarter hypotheses - includes appendix

Administrative Science Quarterly, June, 1993 by Barry M. Staw, Sigal G. Barsade

The original DDI coding scheme was revised for this research. We developed specific and objective categories for behavior on the in-basket, so as to reduce any problems of coder judgment and to increase replicability. The following decision dimensions were coded from the in-basket exercise: (1) accuracy: whether the decision made was actually the correct one; (2) additional information: whether additional data was requested from subordinates or gathered before making a decision; (3) situational contingencies: whether subjects recognized the interrelationship of problems and decisions to each other; (4) use of data: use of quantitative reports such as productivity and absenteeism indexes in making decisions; (5) timeliness: whether problems that needed quick action were addressed and others not undertaken prematurely; (6) delegation: whether tasks and decisions were delegated to the right person; and (7) follow-up: whether instructions were given to others for future reporting and meeting dates.

For each of the above coding categories, counts were made on whether the participant displayed the particular or appropriate behavior. A one or zero was recorded for each of the decision items that provided an opportunity to display the coded behavior. When a decision item did not provide an opportunity to display the behavior, no coding was done. A total score was calculated for each participant for each of the decision dimensions.

There was no scoring for the time participants took to complete the DDI exercise, because every participant appeared to use all of the available three hours. Differences in the quantity of items addressed and the thoroughness of their solution were assessed by the codings for decision quality described above.

Of the dimensions described here, the first four constitute the fairest comparative test of the sadder-but-wiser versus the happier-and-smarter hypotheses, since for each of them one could theoretically predict opposite results. On the latter three measures (timeliness, follow-up, and delegation), it is only possible to draw one-way or unitary predictions. For example, for timeliness and follow-up it is possible to predict an adverse influence of positive affect, since the sadder-but-wiser position would imply more careful timing and greater follow-up on the behavior of others. Likewise, one might argue that positive people are better able to delegate properly on a complex problem. These latter predictions are more speculative inquiries, rather than comparative tests derived directly from the affect literature.

For each of the measured dimensions of decision making a simple count was used to note whether the behavior was exhibited or not. We then calculated the number of items on which subjects exhibited a given behavior, out of the total number of items on which there was an opportunity to show the behavior (not all 21 items involved each of the aspects of decision making). Finally, a percentage mean was noted for each decision dimension.