3. Research method
The research method is based on a qualitative approach through semi structured interviews. The use of open ended questions in this qualitative study is due to the fact that this allows respondents to express their opinion, what they think, what they know, and what they practice without the influence of the researcher, thus not being limited to a set of alternatives. It facilitated the collection of information whenever occurred
deviant responses, or unclear, derived from the questions. Each phase of the research method and its outcomes are shown in Figure 1.
3.1 Planning
This research was conducted in a large public sector IT organization which develops systems aimed to automate public policies of Brazilian government. The systems range from Business Intelligence products to web systems, handling a large amount of data. The projects are characterized by the existence of various stakeholders, both internal and external, which require a large ability from project managers to deal with
various perspectives and so achieve the expected results. In this context, cognitive biases, if poorly controlled, can lead projects to failure.
The sample of this study consists of seven experienced project managers. On average, their working experience managing projects was five years. Other three project managers, who did not participate of the interviews, were selected for the triangulation phase. An invitation was sent to each project manager contextualizing the research objective, the interview script and the cognitive biases to be analyzed with their respective descriptions in order to make them think beforehand and thus optimize the time of interviews.
3.2 Execution
During the interviews, for each bias, after presenting its description, it was asked the following questions: “Have you ever experienced a situation related to this bias in the context of project management?” and “If so, describe the situation and actions that was taken.”. By using keywords selected by the researchers related to each bias, not accessible to the interviewees, the interviewer avoided deviant responses that were unrelated to the question.
In order to avoid inconsistent responses not related to reality, it was emphasized that the interviews would be recorded for the sole purpose of being transcribed to support the analysis, and that the anonymity of each interviewee was guaranteed. The interviews were recorded using the software QuickTime Player and were analyzed after each one in order to improve the following interviews. On average, the duration of each interview was 30 minutes.
3.3 Analysis
The content analysis of individual responses was performed by using two complementary techniques: (i) Strategic Options Development and Analysis approach (Eden and Ackermann, 1998) for the creation and treatment of cognitive maps, which includes aggregating and clarifying the concepts and its relations, and (ii) triangulation technique as a way to prevent the influence of individual analysis based on interviewer researcher’s personal opinion (Northcutt and McCoy, 2004). Three independent project managers participated of this phase. The analysis of semi structured interviews was conducted by the following three steps: :
Grouping the responses for each question: As a starting point of creative thinking, all responses of each question were grouped to extract information about susceptibility, root causes, tools and techniques, with the objective of building a map with all the concepts.
Discussion of initial concepts map: During the triangulation process, the concepts map was discussed with the three project managers who did not participate of the interviews in order to, interactively, review it by including, aggregating or disaggregating the concepts. This process was facilitated by one of the researchers.
Validation of the final congregated concepts map: After the interactions developed in the previous step, it was obtained an updated concepts map with the contributions of each project manager. Through a group meeting, also facilitated by one of the researchers, they discussed, fitted and sought consensus in the final congregated concepts map.