K. KangaMall
Research / Australian Time Use Report 2026

STUDY 001 · VERSION 1.2.0

The method.
The numbers behind it.

Sources, definitions, calculations and review for version 1.2.0.

Source

All estimates come from the Australian Bureau of Statistics (ABS), How Australians use their time, 2024, data cubes TUSDC01–TUSDC14. Tables 1–10 were first released on 17 December 2025; Tables 11–14 were added and Table 8 was reissued with additional data on 5 August 2026. Source: https://www.abs.gov.au/statistics/people/people-and-communities/how-australians-use-their-time/2024 (accessed 27 October 2026). Methodology: https://www.abs.gov.au/methodologies/how-australians-use-their-time-methodology/2024 (accessed 28 October 2026).

The ABS zip file Time-Use-Survey-data-2024.zip used for this release has SHA-256 92416fe73ac7f93f9cb5c7dd88c542be4855b24bc95a53d86aefe30c5a8563a6. The ABS files are not redistributed in the download package; download them from ABS and check the checksum before reproducing the analysis.

Scope and collection

The survey covered usual residents aged 15 and over in private dwellings in Australia, excluding very remote areas and discrete Aboriginal and Torres Strait Islander communities. The achieved sample was 6,899 households, 10,613 people and 19,319 diary days, collected from 22 July to 3 September and 16 September to 29 October 2024. Respondents recorded their activities in 5-minute intervals in an online diary over two consecutive days, or by telephone for one day.

Measures

Times are shown as hours:minutes per day. ABS uses "females" and "males" for sex recorded at birth; this report uses the same terms. "Living situation" describes a person's position in their household (for example, a parent in a couple with children under 15); it is not a household unit. "Unpaid work" is the ABS group of domestic activities, looking after children aged 14 and under, caring and support for people aged 15 and over, and volunteering. "Looking after children" counts childcare recorded as the main activity; supervision while doing something else is published separately by ABS.

Reliability and significance

Every estimate keeps the ABS relative standard error (RSE, for times) or 95% margin of error (MoE, for proportions). ABS quality markers are retained: * RSE 25–50%, use with caution; ** RSE above 50%, too unreliable for general use and not used for conclusions; # proportion with a high margin of error, use with caution.

Differences between two separate groups are tested using the ABS method: SE = RSE × estimate / 100 (or MoE / 1.96); SE(x − y) ≈ √(SE(x)² + SE(y)²); a difference is described as statistically significant when |x − y| / SE(x − y) > 1.96. The approximation ignores correlation between estimates. The findings rest on 25 distinct comparisons and the detailed tables report further tests; no adjustment is made for multiple comparisons, so a small number of significant results could arise by chance. Shares within the same group of respondents (main reasons for feeling rushed) are ranked but not tested.

ABS randomly adjusts table cells to protect confidentiality, so components may not sum exactly to totals, and one participation rate is published as 100.1%. Values are kept as published.

Data processing

analysis/extract_tus2024.py reads all 33 estimate sheets in the ABS data cubes, pairs every estimate with its error sheet, keeps the quality markers and the source cell, and converts hh:mm durations to minutes. Nothing else is changed. analysis/validate_tus2024.py checks completeness, ranges, fixed values and population consistency across tables. analysis/build_report.py generates every table, figure and claim from the extracted data. analysis/check_claims.py re-opens the original ABS files and checks every number quoted in the claim register; analysis/check_report.py checks that every number in the report text appears in the claim it cites.

Correction to the ABS file. In ABS Table 2.1, the female rows for equivalised weekly household income quintiles are printed one row above their labels: the lowest quintile's values sit on the heading row and the "Highest quintile" row is blank. The error sheet (Table 2.2) and the time tables (2.3, 2.5) are aligned correctly. The extraction moves these 55 values down one row only after confirming that each moved population estimate equals the population estimate in Table 2.3 for the target row, and labels each moved value in the data. These values are not used in the report's findings.

Data dictionary: data/tus2024_long.csv

One row per published ABS value (25,110 rows).

Column Meaning
table ABS table number, e.g. 2.5
source_file ABS data cube file
sheet_title Full ABS table title
measure participation_pct, avg_minutes_per_day_participants, avg_minutes_per_day_total_population, proportion_pct (Table 8) or population_000
row_order Order of the row within the ABS table
row_dimension Grouping heading the row sits under (e.g. Living situation)
row_sex_section Sex section in Table 2 (Males, Females, Persons)
row_top_activity Top-level activity group for activity rows
row_label ABS row label as published, footnote letters retained
col_header_1 – col_header_3 ABS column headings, top to bottom
unit_header ABS unit heading
value Published value; durations converted to minutes
value_unit minutes per day, per cent or thousand persons
abs_quality_flag caution_rse_25_to_50, unreliable_rse_over_50, caution_high_moe, not_published or blank
error_type rse_pct or moe_95_pct_points
error_value ABS error value from the paired error sheet
source_cell ABS sheet and cell, e.g. Table 2.5!C12
kangamall_adjustment Blank unless the value was realigned (Table 2.1 correction above)

The detailed report tables (outputs/tables/table-01.csv to table-11.csv) and figure data (outputs/figures/figure-0N.csv) carry the same source-cell and error fields. The claim register (claims.csv) lists each statement in the report with its ABS cells, calculation and review record.

How this report was produced

Kangamall Research reviewed and edited this report and is responsible for its content. AI-assisted tools were used for data processing, calculation checks and drafting. Quality control had three layers: automated checks that re-open the original ABS files; a check by the analyst; and an independent recalculation of every claim from the original ABS files, together with checks of 800 table cells, 28 significance results, 142 figure values, the method statements and all links. That review corrected one statement before publication (two domestic tasks were tied rather than one leading) and tightened several wordings. The report has not been externally peer reviewed.

Limitations

Version history

Downloads

All 25,110 values (CSV) · Claim register (CSV) · Source register (CSV) · How to reproduce · File checksums

Return to report & downloads