After the interviews were transcribed, the analysis was conducted by the researcher without using any software. The process of analysis was time-consuming, taking months to complete. It was also challenging and required a lot of organization and support from my committee and fellow researchers. Conducting the analysis allowed me the opportunity to gain coding experience, interpret the data, and identify words or phrases that are not used in the English dictionary (Bright & Connor, 2007). In addition, software can be expensive, and time-consuming to learn how to appropriately use the software (Padgett, 2008). I started with some predetermined categories for data coding, though some of the codes changed during analysis (Marshall & Rossman, 2011, Miles & Huberman, 1994). Each phase of the data analysis entailed the following: data reduction, where collected data was brought into manageable chunks, and interpretation, by bringing meaning and insight to the words and acts of the participants in the study (Marshall & Rossman, 2011). I started with first level coding by coding each transcription sentence-by-sentence, which included single terms (Miles & Huberman, 1994). There were 201 codes following the first cycle of coding. Some examples of codes include; self-confidence, perseverance, support, sacrifice, shock, financial capital, distraction, insecure, balance, power, financial security, overwhelmed, ownership, dual role, and safe space. Next, I engaged in second level coding, also known as pattern coding in which the codes were combined into smaller number sets (Miles & Huberman, 1994). In order to do this, codes were organized under categories and color-coded on large sheets of drawing paper and posted on the wall. Codes that were redundant were eliminated while others with the same meanings were combined. Themes were developed from the categories that came forth from the data and were constant in all the interviews. Overall, five themes were identified and codes were reduced to 56. Codes were reviewed again and eliminated based on the number of times it was mentioned; this produced five themes, twenty-five (25) secondary themes, and twenty-two (22) subthemes.
In keeping with the nature of phenomenological research, a synopsis of each study participant’s experiences, examination of the context and setting of these experiences, and a condensation or summary of major themes with associated excerpts from the interview was developed. (Padgett, 2008).
Credibility and Ethical Considerations. This section will describe the steps I took to enhance the credibility and trustworthiness of this study. This study utilized the following strategies for rigor: peer debriefing and support; member checking; and reflectivity.
Credibility.In order to achieve rigor in a qualitative research study, a researcher must maintain credibility, protect data from contamination, and establish trustworthiness. Trustworthiness is a word that can be used, to sum up credibility, transferability, audibility, and conformability (Padgett, 2008). Trustworthiness deals with how one can determine the trust that we have in research (Marshall, & Rossman, 2011). Threats to trustworthiness include reactivity, researcher biases, and respondent bias (Padgett, 2008).
Reactivity is the chance that the researcher’s presence might have a distorting effect on the participants’ beliefs and behaviors. Researcher biases can emerge when observations and interpretations are clouded by the researcher’s preconceptions and personal opinions. Respondent bias can happen when the respondent withholds information or even lies to protect their privacy or to avoid revealing unpleasant truths (Padgett, 2008).
These threats to trustworthiness are present in all research studies, but there are many ways to build confidence and rigor (Padgett, 2008). Strategies for rigor include prolonged engagement, triangulation, peer debriefing/ support, member check
