74. This section discusses the following measurement issues related to income data:

• collecting income in dollar amounts rather than in ranges;
• missing data.

DOLLAR AMOUNTS VERSE RANGES

75. Ideally income should be collected in dollar amounts and not as ranges. Collection in dollar amounts enables greater precision when aggregating incomes of individuals to derive the income of income units, families and households. It also offers greater flexibility in presenting estimates. When income data is collected in actual dollar amounts, summary measures such as means and medians can be simply produced and equivalence scales can be applied. The disaggregation of the population into income quintiles and deciles is also straightforward when actual dollar values are collected.

76. Where income is collected in ranges (as in the 2006 census) it is not possible to aggregate person level income ranges directly to higher levels. To attempt to overcome this, the income range is used in conjunction with data from the most recent Survey of Income and Housing to impute an income value for each person. The imputed values for each person are then aggregated to derive imputed household and family level incomes. However there are significant methodological and resource issues in applying this approach and the results obtained are subject to substantial limitations.


MISSING DATA

77. Non-response by persons selected in the survey occurs when people cannot be contacted or are unable (or unwilling) to provide the information required. Non-response to income questions leads to a loss of information, and potentially biased estimates if ignored in analysis. The magnitude of any bias depends upon the level of non-response and the extent of the difference between the characteristics of those who responded to the survey and those who did not.

78. The following strategies can help reduce the level of non-response:

• face-to-face interviews with respondents
• the interviewer indicating how income information is important for various uses when introducing the questions
• the use of interviewers who can speak languages other than English, where necessary
• follow-up of respondents where there is no initial response.

79. The following strategies can help to reduce the impact of non-response in analysis:

• imputation of missing values
• ensuring that weighted data is representative of the population by aligning the estimates with population benchmarks.
  

Previous Page

Next Page