Kylie has over 25 years’ experience in market and social research and evaluation with specialist fieldwork suppliers and full service research houses. Kylie has designed and managed complex data collection methodologies using innovative, technological solutions and in particular has used digital technologies to overcome language and literacy barriers with Humanitarian Migrants, Prisoners and Remote Aboriginal and Torres Strait Islander people to enable them to have a voice and have their views considered in Government policy and decision-making.
Kylie has been the project director responsible for multi-million dollar government contracts for major programme and policy evaluations focused on health and wellbeing such as: Kylie is passionate about social, policy and evaluation research with a particular focus on making a difference for vulnerable and at risk community groups in Australia. Kylie is one of Australia’s leading research specialists in conducting research and consultation with Aboriginal and/or Torres Strait Islander people particularly in remote communities.
Kylie also undertakes capacity strengthening activities, workshops and strategic planning with Aboriginal owned organisations and their boards. Kylie is currently a candidate for PhD with the University of Queensland to contribute to the advancement of knowledge in digital survey research methods and interactive techniques used to collect data from mainstream and vulnerable target audiences.
Co-Presenters: Sara Dolnicar
Long surveys are problematic because respondents do not like them and get fatigued while completing them. The dislike of long surveys leads to invited respondents refusing to participate in survey studies which reduces response rates, or dropping out and terminating the survey early which reduces completion rates. Tired respondents also provide lower quality responses. A solution to these problems is to strip down item batteries to include only informative items in view of the insights sought.
This study proposes the use of an extended version of principal component analysis to (1) assess the potential of item reduction and (2) identify uninformative survey items which can be dropped. The method is explained using a prototypical artificial data set and illustrated using two empirical survey data sets.