The researchers employed a quota sampling technique. The underlying target population was estimated to be 283,937 adults (over 18 years old), ‘new’ Rohingya (those who had arrived in Bangladesh after the events of 25 August 2017) and currently residing in refugee camps in Cox’s Bazar district. This number was determined through simple calculations based on the most current data available from the UNHCR (15 March 2019) at the time of data collection preparation. The target sample was 1,300 new adult Rohingya. The data provided by the UNHCR allowed the researchers to establish the population and gender distribution for each refugee camp, and thus choose the number of respondents to include in the survey from each camp proportionately. Each subgroup within the target population received representation within the sample and no subpopulation was overrepresented in the results
Out of a total of 1,455 surveys conducted, 1,277 were deemed sufficient for further analysis. The sample of 1,277 respondents can be considered broadly representative of the total adult, new Rohingya population residing in the selected refugee camps across Cox’s Bazar district
The enumerators were given instructions to select respondents randomly in the chosen camps which was not always possible for several reasons. Hence, even though the sample can be considered representative of the population under study, the potential for selection bias cannot be excluded.
The enumerators, as Rohingya refugees themselves and fluent in Rohingya, could clarify questions posed by the respondents. However, the questions and the responses were translated from English to Rohingya and from Rohingya to English respectively by each enumerator on the spot. This might have resulted in misinterpretations and/or negatively influenced the accuracy of some responses.
Data regarding the respondents’ views of Rohingya involvement in illegal activities, substance use as a coping mechanism, and of violence in the camps should be interpreted with caution because of the potential for bias due to respondents trying to preserve the image of their community with regards to these sensitive topics
