With regard to the first criterion, our purpose is to build and elaborate theory. A qualitative field study is especially appropriate for this research because we are developing nascent theory about anger contexts in organizations. Nascent theory “proposes tentative answers to novel questions of how and why” and proposes new constructs and suggestive models (Edmondson 2005: 5, McGrath 1964).
We used anger episodes experienced by organizational members as the context for this study. Prior research has found that respondents can identify distinct emotion episodes with a beginning and an end (Frijda 1993). We asked each respondent to recall at least two anger episodes, one in which they were the target of anger and one in which they expressed anger. A number of respondents also provided examples of anger episodes in which they observed an angry interchange between others. We focused on the respondents’ perceptions of the episode, recognizing that each party in an anger episode may have different perceptions of what happened. Interviews allowed us to expand and clarify respondents’ answers via follow-up questions. Since our focus was to identify work unit norms, it was imperative that our study be conducted in a field rather than a lab setting. Toward that end, we also drew on informants in each organization who could provide background on the normative prescriptions within each setting and who could clarify questions that arose during data analysis.
We used theoretical sampling (Glaser and Strauss 1967) to select several different types of work units which vary both in the extent to which anger display rules are prescribed by the organization and the extent to which anger expressions are valued or discouraged. Theoretical or purposive sampling (Strauss and Corbin 1990) is a theory-building technique that uses reasoning to anticipate distinctions among sites, respondents and samples in order to maximize these distinctions for the purposes of theorizing about them. Pratt (2007) refers to these as extreme cases selected to enhance the likelihood of observing the dynamics in question. Drawing on the emotional labor literature (c.f. Ashforth and Humphrey 1993,