Analysis
All in depth interviews were audio-taped and transcribed verbatim in Kinyarwanda. Transcripts of interviews were made anonymous using participants’ codes. Analysis of in-depth interviews was conducted using atlas ti. Software. Data analysis for in depth interviews about mother-child interactions during feeding (responsive feeding) followed thematic analysis method which involved reading the transcripts for familiarization with data, generating initial codes, sorting the different codes into potential themes, reviewing and refining themes as well as reporting [18]. Relevant verbatim quotes, translated from Kinyarwanda language to english, were incorporated in the presentation of the results to help with data understanding.
For the analysis of mealtime observations, the number of times a single practice occurred was counted. An action or a practice that happened at least two times during the meal was considered as present while practices which occurred less than two times were counted as absent [3]. No statistical test was performed, because the sample size was too small. The transcripts of lunch meal were coded by two independent raters that are the principal investigator and the research assistant.