This paper explains why imputing missing income data is of benefit to Longitudinal Study of Australian Children and shows how two techniques have been used in combination to impute this income data.
Income is one of the most important pieces of economic information collected in Growing Up in Australia: the Longitudinal Study of Australian Children. Income is an important variable as it serves as an indicator of a family’s ability to financially invest in children’s development and access services that might support a child’s. Useful on its own, income can also be combined with other socio-economic data to provide a comprehensive overview of the socio-economic contexts within which children live and develop.
However, income tends to suffer from item non-response across all surveys. This may occur because individuals do not know or cannot accurately state their (or their family/partner’s) income, or do not wish to disclose this type of personal information). The result is that levels of missing income data tend to be higher than other variables and can pose difficulties for analyses that incorporate income.
A common approach to deal with missing items is to apply listwise or casewise deletion. However, this might lead to two main issues. First, exclusion of missing cases lowers the number of overall cases, which in turn reduces the power of the statistical analysis. Second, generalisability of the interpretation of the analysis can be limited because missing data can be unevenly distributed across sample populations and can be related to the target variable. For example, those with very low or very high earnings may be more hesitant to report their income, thereby decreasing the overall variability of the income data and underrepresenting earners at the extremes of the distribution. In order to overcome this, researchers responsible for large national household surveys increasingly impute values through reliable and robust methods rather than simply excluding missing responses from the analysis.
Therefore, this report aims to impute missing income data for Parent 1 and Parent 2 using the approach developed for another longitudinal study—the Household, Income and Labour Dynamics in Australia survey. The Household, Income and Labour Dynamics in Australia method is a well regarded and widely accepted method based on decades of previous research.