USING THE TABLEBUILDER
This page provides specific information relevant to the Survey of Income and Housing (SIH) TableBuilder product. It will assist users in understanding and interpreting specific data items and functions relevant to specific SIH estimates.
For general information relating to TableBuilder, or instructions on how to use features of the TableBuilder product, please refer to the TableBuilder, User Guide.
Detailed information about the survey including scope and coverage, survey collection methodology, estimation method and reliability of estimates can be accessed from the Survey of Income and Housing, User Guide, Australia, 2017–18.
The Data Item List for the SIH TableBuilder product is available from the 'downloads' tab.
CONTINUOUS DATA ITEMS
The TableBuilder file contains a number of continuous data items that are available for selection from the 'Summation Options' in the 'Customise Table' panel. Continuous data items are those data items which can have a response value at any point along a numeric continuum. Examples of continuous data items are total current weekly household income from all sources, total value of household assets, usual hours of child care per week. To create tables for continuous variables, the user must first either create ranges or if interested in sums, medians or means, the user then chooses the appropriate option and adds it to the wafer, row or column.
Only one continuous item can be in the 'wafer' at any one time in TableBuilder, this is particularly important if categorical data items will be added to the row or column. A separate table needs to be created for each continuous item (one table per item with the relevant item in the 'wafer').
There are also continuous hybrid variables containing categories such as 'not applicable'. In these cases the variable can be found under 'Summation Options' and also under the appropriate level in the list of data items. Hybrid variables on the SIH file include variables such as Debt-to-gross income ratio and Capacity of solar electricity power system (Number of kW of system); a full list can be found In the data items list.
For these continuous hybrid data items there are special codes allocated for certain responses (e.g. 9999 = 'Not applicable' or 99999999 = 'Negative or zero debt'). When creating ranges in TableBuilder for such continuous items, special codes will automatically be excluded from calculations of sums, means, medians or ranges. When added to the table, it enables comparisons across populations, i.e. comparing households with a debt ratio to those without.
Limits for ranging continuous items are detailed in the data item list which can be accessed from the 'downlods' tab.
- Once a quantile or range has been created based on a continuous item, it can be found by selecting the ranges button (below summation options):
Below are TableBuilder outputs containing examples of data items with continuous hybrid variables:
Example 1: Debt-to-gross-income ratio - Number of households
Columns: Debt-to-gross-income ratio
Wafer: Number of households (default)
The number of households with a debt to gross income ratio are contained in the 'A valid response was recorded' column. The 'Negative or zero debt' column includes households that do not have a debt, or that have a negative debt to income ratio. These instances are excluded from the calculation of Quintiles and Custom Ranges as shown in the tables below.
Example 2: Debt to gross income ratio - Ranged - Numbers of Households
Columns: Debt-to-gross-income ratio
Rows: Ranged debt-to-gross-income ratio
Wafer: Number of households (default)
The 'Negative or zero debt' column includes households that do not have a debt or have a negative debt to income ratio (income is higher than debt). These instances are excluded from the calculation of Custom Ranges.
Example 3: Mean gross income by Family composition of household for households with and without a debt ratio
Wafer: Weighted mean of Total current weekly income from all sources (Household level)
Rows: Family composition of HH (brief)
Columns: Debt-to-gross-income ratio
In this example the hybrid continuous variable (debt-to-gross-income ratio) is cross tabulated with a categorical value (Family composition of HH (brief)). This will enable comparisons across populations, i.e., comparing households with a debt ratio to those without. To demonstrate, the mean value of a continuous item (Total current weekly HH income from all sources) was added to the wafer, and the mean value for both columns is displayed. The 'not applicable' category is included in the calculation of the totals.
Note: If a continuous variable has a very limited range on the Data Item List between 0 and 0 then this means that there are too few contributor to this variable and therefore only estimates of weighted sum, weighted mean and weighted median can be derived and ranges cannot be created.
Differences with Published Estimates
Ranging continuous items will not provide the same output as published data which is based on the SIH (i.e. Household Income and Wealth, Australia, and Housing Occupancy and Costs, Australia). This is due to additional confidentiality measures applied to ranged continuous items.
Quintile cut-offs in TableBuilder for continuous data items are defined according to whole numbers, whereas in published estimates they are defined according to two decimal places. This will cause slight variations between TableBuilder and published estimates based on the SIH.
Flags have been created to indicate modules in the questionnaire that have imputed data in them. If imputed values are to be removed from the analysis, this can be done by only including the 'not applicable' category. The data item list contains all the flag variables.
The 2017–18 SIH TableBuilder contains seven levels of record files: Household, Income Unit, Person, Childcare, Superannuation, Loans, and Wealth. Different information is available for each record level, and each level contains counts of the number of instances, or units, of the item in each category.
For example, if a person has more than one superannuation account, there will be multiple superannuation records for that person on the superannuation level.
Mortgage(s), motor vehicle loan(s), and personal loan(s) are all examples where a household may have multiple types of items, which means there will be multiple records for that household on the Loans level.
MULTIPLE RESPONSE DATA ITEMS
A number of questions included in the survey allow respondents to provide more than one response. The data items resulting from these questions are referred to as 'multiple (or multi) response data items'.
The example below displays 'Types of formal child care income unit used in the last 4 weeks'.
When a multi-response data item is tabulated, the same record is counted against each response provided. As a result, some household units utilising multiple forms of formal child care are counted multiple times. Consequently, the sum of individual multi-response categories can be different to the population or actual number of people applicable to the data item, as respondents are able to select more than one response.
Multi-response data items are identified in the data item list (available via the 'downloads' tab).
COUNTING UNITS AND WEIGHTS
The Summation Options section in the Customise Table panel contains the counting units/weights that are available. It is critical that the correct weight (or summation option) is used when specifying tables for counts of persons or proportions. As a general rule of thumb, use the weight which corresponds to the level of analysis you are undertaking, i.e. household level weight with household level items, the person level weight with person level items. To analyse persons in households, use the person level weight with household level items.
The Household level weight is the default in SIH TableBuilder. The weights are located under the relevant level in the Summation Options. Below is a list of the levels and corresponding weights. To change the weight from the default, click on the "Sum" box in the appropriate level:
The default summation option will be automatically added to a table when the table is being specified, therefore care needs to be taken that this is the correct weight for the particular tabulation. If the default weight is not the required weight, select the appropriate weight from the Summation Options list and add either to the 'filter' or the 'wafer' (either will override the default of household level in the 'filter') for categorical data items. This needs to be done when obtaining counts and proportions. It is not necessary (or possible) for tables which have a continuous item in the wafer.
For example, if analysing counts of persons for person level items, select 'Person - Sum' as shown below, and add to the 'wafer', before running the table.
Weighting and Defining Quintiles
Quintiles need to be weighted according to the corresponding level of the continuous item. For example, for quintiles based on 'Total current weekly HH income from all sources' (a Household level item), set the 'Equal distribution of ' box to 'Household level':
Using the 'Filter by' option
The 'Filter by' option enables quintiles to be defined further, according to a categorical variable.
To illustrate, in the example above, the quintiles for gross household income are based on the Australia wide population. This means that the entire population is divided into five equal sections and the quintile cut-offs will be based on the mean values for all households in Australia. Datacubes 13 and 14 in the Household Income and Wealth, Australia publication are state based, and the distributional analysis is also undertaken at the state level for these tables. This means that the quintiles are further defined by the relevant state.
To replicate this in TableBuilder, use the 'Filter by' to add the relevant state as in the following image:
This means that the quintiles will be based on the population for NSW only.
Weighting Equivalised Items
Equivalised items are available at the household and the person level in TableBuilder. Equivalised items are household items which have been equivalised to take all persons in the household into account. All analysis using equivalised items in publication Household Income and Wealth, Australia, is done using the equivalised items on the person level with person weighting. Some equivalised analysis in the publication Housing Occupancy and Costs, Australia, is done at the household level with household weighting.
Example 4: Mean equivalised disposable household income (EDHI) by equivalised disposable household income quintiles
This example recreates a section of Table 5.4 in publication Household Income and Wealth, Australia 2017–18 where mean equivalised disposable household income is used at the person level, with person level weighting on the income quintiles:
Creating the EDHI table in TableBuilder
1. Range 'Current weekly HH equivalised disposable income (Person level)' into quintiles and apply the person weight (Person level in the 'Equal distribution of' box). Select 'Create'.
2. Add the EDHI quintiles defined in step 1 to the columns via the 'Ranges' drop down box (below summation options) in the data item list.
3. Add 'Main source of current HH income' from the Household level categorical items to the rows.
4. Add 'Current weekly HH equivalised disposable income (Person Level)' - Mean - to the wafer.
5. Select 'Retrieve Data':
Note: Slight variations between TableBuilder and estimates published in the Household Income and Wealth, Australia publication are evident in this example. This is because quintile cut-offs in TableBuilder are defined according to whole numbers, whereas in the publication they are to two decimal places.
More information regarding the use of household or person weights for equivalised items is located in publication Survey of Income and Housing, User Guide, Australia, 2017–18 (refer to the 'Summary indicators of income and wealth distributions' and the 'Weights' sections).
To minimise the risk of identifying individuals in aggregate statistics, a technique called perturbation is used in TableBuilder to randomly adjust cell values. Perturbation involves small random adjustments to the statistics and is considered the most satisfactory technique for avoiding the release of identifiable statistics while maximising the range of information that can be released. These adjustments have a negligible impact on the underlying pattern of the statistics. After perturbation, a given published cell value will be consistent across all tables. However, adding up cell values to derive a total will not necessarily give the same result as published totals.
ZERO VALUE CELLS
Tables generated from sample surveys will sometimes contain cells with zero values because there were no respondents who satisfied the parameters of the cell in the survey sample. There may, however, still be persons or households within the general population with these characteristics. If all persons within the population had been enumerated (ie, a Census), there may have well been a value for that cell. This is an example of sampling variability which occurs with all sample surveys. Relative Standard Errors cannot be generated for cells with zero values.
SEARCHING DATA ITEMS
It is possible to search for data items using the search function below the data item list. When searching for a data item label using the search box under summation options, the search will default back to displaying household level, but will be performed across all levels.