Table 2

Qualitative sampling approach by informant category

Informant categoryPurposive criteria
Local government area health officials (LGAs)Positive deviants (PD)
  • At least one state per geographical zone represented in the final LGA sample, for geographical representativeness across Nigeria

  • Prioritise states with the highest number of COVID-19 cases reported to Nigeria Centre for Disease Control in each zone

  • Balance of urban and rural geographies

  • Exclude areas with known security concerns for in-person data collection

Comparators
  • Not eligible to be a positive deviant

  • Similar geography (urban or rural) as the PD (usually geographic neighbour)

  • Similar relative wealth as the PD

  • Within the same state as the PD

  • Exclude areas with known security concerns for in-person data collection

Primary health care facility (PHC) officers in chargePDs
  • If no clear PHC facility-level pattern emerged in the quantitative analysis, PHC facilities were randomly selected from a list generated by LGA officials.

Comparators
  • Randomly selected from an available facility list of non-positive deviants

Opinion leadersBoth positive deviants and comparators
  • LGA and facility informants were asked to identify additional informants critical to the local COVID-19 response

  • A list was generated of these suggestions and relevance of suggestions confirmed.

  • Individuals were randomly sampled from the final list