Cross-Cultural Biases and Inattention

Inattention (or careless responding) is prevalent in self-administered surveys, posing a threat to data quality and biasing results. Risk of low-attention is especially high in remote and self-administered settings, owing to the lack of direct monitoring and supervision by lab staff. As remote and self-administered surveys are increasingly preferred for conducting research, the risk of inattention becomes increasingly prevalent.

In this context, we will test the use of a set of established survey inattention measures (Craig & Meade, 2012; Berinsky et al. 2019), tailored to the Kenyan population but with a generalizable framework that can be used in similar contexts. These will allow us to detect low survey attention in a wider cross-cultural study while expanding the literature on survey inattention in remote surveys. Literature suggests that respondents tend to respond differently on a range of substantive indicators (Bowling et al. 2016). Removing inattentive respondents from a dataset because of data quality therefore poses a risk to the generalizability of estimates. Faced with the issue of inattention, we test two basic interventions in order to reduce survey inattention. These are:

- A small financial incentive for answering a question in the survey correctly
-The placement of an obvious attention check early in the survey (Hauser & Schwarz, 2015)

Implementing the above provides the opportunity to measure the sensitivity of attention to interventions specifically designed to manipulate it. With this, we will be able to derive recommended survey design features to reduce inattention in mobile surveys.