and Hulin, 1997). The vertical axis reflects the level (micro to macro) of the outcome. Previous studies have generally focused on a particular level of analysis such as interpersonal (Averill, 1982) or group/team (Jehn, 1995), while this schema provides a more inclusive coverage of the domain and recognizes that anger episode outcomes may span individual, interpersonal and work unit levels simultaneously. This conception fits the data better than using a focused approach on one dimension. Although these levels of outcomes, from individual to interpersonal to group and organizational, have not been used theoretically to describe anger outcomes, they are common classifications in the organizational literature. It seems logical that the anger episodes would also manifest their effects on these levels. Thus, using these two dimensions to classify the anger outcomes allowed us to develop four categories of outcomes that are qualitatively distinct, best fit the data, and have not previously been identified in anger research: negative individual, positive individual, negative social and positive social.
Notably, most anger episodes generated outcomes in multiple categories, such as both stress—a negative individual outcome—and solving the problem—a positive social outcome—with a total of 542 outcomes out of 558 reported. Negative and positive outcomes were each reported exactly half the time or 271 times each. Negative individual outcomes (including stress, guilt, physical illness, humiliation, hurt, withdrawal, and avoidance) were reported 161 times (30%). Negative social outcomes (including damaged relationships, decreased morale, and retaliation) were reported 110 times (20%). Positive individual outcomes (including let go, stood firm, changed my behavior, felt satisfaction and upheld principle) were the least common outcomes, reported 82 times (15%). Positive social outcomes (including improved relationships, increased group motivation, and problem solved) were the most common outcomes, reported 189 times (35%). This analysis shows that it is important to consider multiple levels of analysis when studying anger outcomes. While sometimes the outcomes reported were all negative or all positive, many respondents reported a combination of both positive and negative outcomes from the same episode. It appears individuals easily recall and report negative individual outcomes because of their salience. Often as they continue to identify additional outcomes, more positive outcomes are recalled.