Improving hard-to-reach sub-group coverage in longitudinal surveys
|Day:||Tue 2 Jul|
There is a growing recognition that the absence of certain hard-to-reach populations from survey data (such as people living in institutional settings, homeless, etc.) can substantively affect key policy-relevant estimates (such as poverty rates, health and educational outcomes) since those who are missing are likely to be the most vulnerable.
Using Understanding Society, we investigate: i) the extent to which such groups flow in and out of the Study; ii) the extent to which field operations are successful in areas, where hard-to-reach groups are more concentrated; and iii) what direct and indirect information is already collected by the Study that would allow us to identify those who are likely to be in (or fall into) one of the categories of vulnerable/hard-to-reach subgroups. We begin with comparing estimates of small, hard-to-reach groups based on Understanding Society data with relevant administrative data sources and review the information available in the Study that would allow us to identify such groups indirectly. We then look at household response rate by different types of small area characteristics, known to be strongly associated with the hard-to-reach populations. We conclude by providing recommendations on the type of improvements that could be made to increase coverage of the hard-to-reach subgroups.