Table 2.

Techniques for longitudinal and cross-sectional analysis of in-depth interview data

Techniques	JustHouHS	Stigma and social cohesion
Case-based tracking and synthesis documents  Facilitated data management Enabled quick familiarization with data at each time point  Aided systematic preparation of follow-up questions	A matrix was created for each participant summarizing data at each of the four waves including demographic information, survey data, and interview summaries  Organized by the primary study domains (e.g. housing, economic situation, health, criminal justice involvement, substance use, partners and sex, HIV)  Contributed to a master topical matrix that combined information from different cases, with separate tabs for each substantive domain, allowing the identification of themes, and comparison of context and experiences across participants, over time	An analytic summary was created for each participant that included socio-demographic data and viral suppression status, and the following sections with space for each of the three interviews  Contextualized summary of HIV narrative including diagnosis, family and relationship dynamics, and managing life with HIV  Key quotes and themes (inductive and deductive) used to develop codebook  Summary of content of topical modules included in the interview guide at each wave (e.g. mobility, perceptions of long acting injectable ART, mindfulness)  Follow-up questions to clarify content from previous interviews and probe on emergent areas
Indexing  Identified large sections of text by topic  Facilitated easy retrieval of information	20 indices related to main study domains applied to transcripts in Nvivo  Facilitated development of codes and writing of analytical memos	Section within analytic summaries indexed data from topic-specific modules (long-acting injectable anti-retroviral therapy (ART), mindfulness, mobility). This facilitated efficient review of topical information while keeping the data within the overall context of the participant  Example: mixed methods analysis of perceptions and attitudes of long-acting injectable ART at baseline [26]
Coding  Labeled thematic content systematically using software	Used in conjunction with indices  Example: for an analysis of the residential experiences of people returning from prison, excerpts labeled with the index “housing” were further coded with sub-codes related to types of residences (halfway houses, homeless shelter, etc.) in Nvivo	Codes were developed to label sections of the HIV narrative elicited in baseline interviews including diagnosis, disclosure, initiating care/treatment. Coding these chapters using Atlas.ti facilitated comparative analysis of HIV trajectories based on viral suppression status across participants  We plan to use time-focused codes to systematically track processes related to stigma and social cohesion
Memo writing  Facilitated analysis and interpretation of original research questions and emergent themes  Bridged analytic and interpretive processes	Could be specific to one transcript, participant, group of participants, or thematic  Facilitated sharing analytic thinking prior to research team meetings and summarized ideas generated during meetings  Assisted with more deeply understanding the data  Example: in an analysis of gender and HIV risk, material indexed at the intersection of relationships, HIV, and sex was used to write participant-specific analytical memos related to themes of interest	Within the analytic summary, a memo-within-a-memo to describe emerging ideas regarding processes related to stigma and social cohesion over time within participants  Thematic memos will be used to facilitate cross-sectional and longitudinal analysis of key themes across participants