Using survey data quality and missingness patterns to inform questionnaire design: learnings from the Mayi Kuwayu Study — The Association Specialists
  1. Schedule
  2. Oral Presentations - Innovations in health data collection and measurement for equity and action
  3. Using survey data quality and missingness patterns to inform questionnaire design: learnings from the Mayi Kuwayu Study

Using survey data quality and missingness patterns to inform questionnaire design: learnings from the Mayi Kuwayu Study (23736)

Benjamin Harrap 1 , Christopher McKay 1 , Lauren Walduck 2 , Makayla-May Brinckley 1 , Ray Lovett 1 , Katherine Thurber 1
  1. Yardhura Walani, National Centre for Aboriginal and Torres Strait Islander Wellbeing Research, Australian National University, Canberra, ACT, Australia
  2. The University of Melbourne, Melbourne, Victoria, Australia

Background

As the largest Aboriginal and Torres Strait Islander-led and governed survey in Australia, the Mayi Kuwayu Study’s questionnaire was designed by and for Aboriginal and Torres Strait Islander peoples, measuring culture and wellbeing in ways that matter to mob.

Both the Wave 1 and 2 questionnaires were developed through an Aboriginal-led survey (re)design process and included domain experts, community focus groups, and pilot testing.

For the third Wave, the redesign process draws on a deeper analysis of data quality and missingness patterns, which we present here.

Methods

Data quality indicators including participant drop-off and item-level and question-level missingness were examined for Wave 2 respondents, using descriptive summaries and data visualisation to identify areas of the questionnaire that may need to be modified for Wave 3.

These data quality indicators were examined by questionnaire elements, such as page number or section topic, and participant characteristics like age and first language, to understand where and how data quality changes.

Results

We report on which questionnaire elements and participant characteristics were related to variation in data quality indicators and how this information was used in the questionnaire redesign process.

Conclusion

In redesigning the questionnaire, we aim to improve response rates and data quality for the next Wave of the Mayi Kuwayu Study. We demonstrate how analyses of survey data quality can directly inform the questionnaire redesign process for longitudinal studies.