We conducted over a hundred in-depth structured interviews with people across the country. We identified interview subjects through professional and personal contacts, articles, and news stories. We made every effort to include a diverse group of men and women, of varying age, race, socio-economic status, vocation, sexual orientation, and geographic location. Our procedure was guided by the classic case study, qualitative approach. This is an intensive open-ended study of individuals in order to provide rich data, often used in the early stages of developing theories and new ideas in social science. The questions we used were standardized but designed to be open-ended and flexible. This encouraged interview subjects to speak freely and allow for deeper conversations to emerge.
Participants were told that we were interviewing a range of people to learn more about how people face challenges in their lives. Participants were informed their stories could be included in a book. They were told that the goal of these interviews was to record real-life examples about the behaviors and the ways of thinking that they used when coping with difficult situations to identify priniciples of agency in action.
Interviews were supplemented with archival information found in public records, books, journals, newspaper and magazine articles, and online sources such as recorded speaking events. Interviews were arranged in various locations: people’s homes, their workplace, or a mutually convenient location, such as a public library. Some interviews were conducted via video conference or by telephone when face-to-face meetings weren’t feasible. All interviews were conducted in one or two sessions and lasted between one and four hours in length.
The development of the Agency Practices Inventory involved an adapted version of the survey design process recommended by Artino and colleagues (2014). This multi-step process began with a search of the research literature related to human agency and a series of interviews with a diverse sample. Candidate inventory items were developed based on a synthesis of the literature and interviews; we then conducted pilot testing of candidate items with a sample of 380 adults recruited from Amazon MTurk. In addition to completing candidate items, participants in this pilot study also completed measures of self-efficacy and hope, which we expected would correlate with agency, and a brief demographic questionnaire. Item range and variance, reliability, and convergent validity were evaluated using the pilot data. Based on this analysis, some candidate items were removed from the inventory, some were revised, and some new items were developed. The revised inventory was then piloted with a new sample of 327 adults from Amazon MTurk, and statistical analyses were again used to evaluate the psychometric properties of the inventory. The revised inventory demonstrated acceptable psychometric properties and, as expected, indicated that agency was positively correlated with both self-efficacy and hope.