2954.0 - Census Working Paper 96/1 - Income, 1996  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 27/02/1997   
   Page tools: Print Print Page Print all pages in this productPrint All

Census Working Paper 96/1

1991 CENSUS DATA QUALITY: INCOME

Population Census Evaluation
June 1996




INTRODUCTION

Although Income was not originally chosen for analysis as part of the 1991 Census Working Paper Series, the Evaluation group has received a number of queries about the quality of Income data. Income was of interest because of a number of data quality problems observed with this variable in previous Censuses. These included the tendency for people to understate their income and a high non-response rate in relation to the other Census questions.


This paper provides a brief assessment of the quality of the data obtained from the Income question in the 1991 Census (Q.29). It should ensure users have a basic appreciation of the limitations of the data. The quality of the data was evaluated by considering intercensal changes, including non-response rates, and comparison with the 1990 Survey of Income and Housing Costs and Amenities. In addition, the changes to the question design implemented in the 1996 Census are described because they deal with a number of issues identified as data quality problems in the 1991 Census.


BACKGROUND

Income is an important socio-economic measure, providing an indication of individual, family and household well-being in different areas of Australia and between different population groups. It also provides information to assist policy makers determine which areas and social groups need social services. Income data is required by many Government departments (including the Department of Social Security, Department of Veteran's Affairs, and Department of Employment, Education and Training), academics and researchers, and welfare organisations, among others.

A question on Income was first included in the 1933 Census in response to the Great Depression of 1929. It was not included again until the 1976 Census but has been included in every subsequent Census. Since 1976, the Census has collected information on the gross income from all sources of people aged 15 years and over. Every Census question on income has asked respondents to select the range in which their gross income falls rather than writing in a numeric amount. This is to improve response rates as many people regard income as a sensitive topic. Also, many people may have difficulty in answering a question asking for an actual dollar amount.


Income was considered to be one of the more contentious questions within the 1991 Census. For the 1991 Census, 9.2% (216) of a sample of 2360 queries to the Census Hotline involved the Income question
(a). In the same Census, debriefings of collectors and other feedback indicated that 'Income is seen as being too personal a question for some people, even though we only ask for an answer in ranges(b).

The Census Income question has been the subject of two previous Census Working Papers. The first, ' Working Paper No. T32, 1981 Census Development Programme -Topic Evaluation: Income', examined the quality of data from the 1976 Census question and the results of testing for the 1981 Census. The second paper, 'Census Working Paper 89/3, 1986 Census Data Quality Assurance Review: Income' focussed on the question used in the 1986 Census and the recommendations for the 1991 Census. These papers provided much useful background information for this report.


(a) Tables from 1991 Hotline Operations, NSW (Folio 46, File 91/1183)
(b) Evaluation of Census Forms and Information Booklet (Folio 47, File 91/1187)