COMPARISON WITH PREVIOUS CURFs
While efforts have been made to maintain comparability between CURFs where possible, changes in data collection, data item standards, analysis requirements and user preferences have resulted in changes to data items which may have an impact on data analysis and the assessment of changes over time.
Changes impacting on all items
The main changes which could impact on all data items were:
- the final sample size for the 2006 TUS was increased from 1997 to improve the coverage of weekend diary days;
- the scope of the survey was changed slightly. In 1997 visitors to the selected dwelling were considered to be within scope if they lived in a private dwelling and had not been at their usual dwelling for any part of the enumeration period. In 2006 all visitors to the selected dwelling were out of scope. In 2006 all people living in special dwellings such as hotels, hospitals or aged care establishments were also excluded;
- the allowable time for uncodeable activities within a diary increased in 2006. In 1997 across all records this averaged to 5 minutes per day and in 2006 it averaged to 7 minutes per day;
- an extra weight was added to the file in 2006. This is the person weight;
- in 1997 the population distributions (benchmarks) for each quarter of the Time Use Survey were obtained by averaging population estimates from two Labour Force Surveys on each side of the Time Use collection periods. Household benchmarks for the Time Use Survey were derived from independent estimates of the resident population for June 1997. In 2006 the two sets of population distributions (benchmarks) were used for person-day and person benchmarking for each collection period of the Time Use Survey. Similarly, for the household benchmarking, two sets of household benchmarks were used for each collection period;
- in 2006 wages and salaries information was imputed using donor records. No imputation was undertaken in 1997; and
- RSEs were calculated differently in 2006 to 1997. In 1997 a statistical model was produced that related the size of the estimates with their corresponding RSEs, and this information was displayed via a 'SE table'. For 2006, RSEs have been calculated for each estimate using the grouped jacknife method and are published individually.
Changes relating to data items
There were also changes to specific data items. As TUS is run infrequently, these changes are too substantial to list individually. CURF users who are undertaking comparison with previous TUS CURFs should take care to compare items they are using for analysis by using the previous supporting information and data item lists in User Guides and CURF related publications for previous TUS releases. A brief summary of the types of changes, along with some examples, is presented below.
Items not on both the 1997 and 2006 TUS CURFs:
The 2006 CURFs have a large number of items that were not provided on the 1997 CURFs. For example, both CURFs have the person level item 'Occupation' although the 2006 CURFs provide further household information with additional items 'Occupation of household reference person' and 'Occupation of spouse of household reference person'. The 2006 CURFs also output a number of items on the use of Internet technology at both the household and person level which were not on the 1997 CURFs. New items, 'Flag for households with imputed wages or salary' and 'Flag for person with imputed wages or salary', also did not exist in 1997 as no imputation was performed on the data. On the 1997 CURFs, the Activity level items, 'Nature of activity' and 'Purpose of activity' were coded by primary and secondary activity, but are now items in their own right for the 2006 CURFs. A new item 'Activity priority' has been added which is used to classify the activity as either primary or secondary.
Many of the new items on the 2006 CURFs have replaced what used to be single output items on the 1997 CURFs. For example, in 1997 the item 'Regular cash income sources' was collected and output as a single item, while in 2006 the same item has been broken up into multiple output items e.g. 'Whether received regular government pensions or allowance', 'Whether received regular child support or maintenance' and so on. The multiple items add more detail to the output and allow for cross comparisons between each other as well. Most of the multi-response items that were collected and output in 1997 as single multi-response items have been split into separate singular items for 2006.
The 2006 CURFs have a 'Family' level which has resulted in additional family level items which were not available in 1997. Many of these new items were output at the household level in 1997 but have been output at the family level for 2006. For example, 'Sex of first youngest child under 15' has been replaced with 'Sex of first youngest child aged 14 and under in family' and 'Whether formal child care usually used' has been replaced with 'Whether family usually uses formal child care'.
A small number of items that were on the 1997 CURFs have been removed from the 2006 CURFs, such as income items 'Annual regular cash income' and 'Annual business cash income', and household items 'Household structure' and 'Number of dependents in household'. Any information that may have been lost from the removal of these items has been made up with the new items added to the 2006 CURFs. For example, while the 2006 CURFs do not output 'Household structure' and 'Number of dependents in household', they output at the person level 'Relationship in household' and at the family level 'Number of dependent children in family'.
Other data item changes:
There are also some minor differences between items on the 1997 CURFs and the 2006 CURFs. The items are conceptually the same but have a different level of detail, different topcode or different category on either the basic or expanded CURFs. This is due to either changes in standards or in survey specific collection and output decisions. Some examples of these types of minor changes are provided below:
- there are several items which have changed from having labelled output categories in 1997 to 'Yes/No' output categories in 2006 - e.g. 'Household has a microwave oven' and 'Household has a dishwasher';
- there are items which have different ranges or categories for the data in 2006. For example, in 1997, 'Duration of unemployment' was ranged '1-26 weeks', '26-103 weeks' and '104 weeks or more'. In 2006, this has changed to comply with current standards and has been ranged from 'Under 4 weeks', '4 weeks and under 13 weeks', '13 weeks and under 26 weeks', '26 weeks and under 52 weeks' and '52 weeks and over'; and
- other items have had different topcodes applied in 2006. For example, 'Age' was topcoded at '75 years and over' in 1997, while in 2006 the data item name has been changed to 'Age of person' and is now topcoded at '85 years or over'.