Appendix 1: Data and methods

About our data

This report uses data from duty holders’ 2021, 2023 and 2025 workplace gender audits. Under the Gender Equality Act, duty holders must submit workplace gender audits to the Commissioner every 2 years. This is part of their requirement to develop gender equality action plans and report on progress.

Duty holders submit audit data for 7 workplace gender equality indicators:

• gender composition at all levels of the workforce
• gender composition of governing bodies
• gender pay gap
• sexual harassment in the workplace
• recruitment and promotion practices
• leave and flexibility
• gendered segregation within the workplace.

Duty holders must collect gender-disaggregated data. This is data that is broken down separately for women, men and gender diverse people. If available, duty holders are also required to report demographic data about:

• Aboriginality
• age
• disability
• ethnicity
• gender identity
• race
• religion
• sexual orientation.

Duty holders must analyse this demographic data and take steps to address any intersecting forms of disadvantage or discrimination that a person may experience. The Act conveys this requirement through the language of ‘compounded inequality.’ You can read more about how the Commission collects and analyses demographic data in the Intersectionality section below. Intersectionality at work (2023) also provides further advice about collecting and analysing demographic data with an intersectional lens.

Data types

The Commission holds 3 types of data, all of which have been used to create this report. These are aggregated workforce data, unit-level workforce data and unit-level employee experience data. Each is described below.

Aggregated workforce data

Workforce data is collected from organisations’ payroll and HR systems. Aggregated workforce data is data that the duty holder has grouped together before submitting. This data mainly consists of headcounts. For example, an organisation may have provided the total number of women working full time at the most junior level of the organisation. The Commission used this type of data for the 2021 audit, where 177 organisations met quality standards.

You can read more about the aggregated workforce data used in 2021 in the Baseline report (2021).

Unit-level workforce data

From 2023 onwards, duty holders submitted de-identified individual employee data extracted from payroll and HR systems. This data provides a range of information about each employee in an organisation, which makes it easier to calculate a variety of data measures. In the 2023 audit, 257 organisations met quality standards. In the 2025 audit, 298 organisations met quality standards. The number of organisations whose data was used to calculate each individual measure in this report varies. This is because an organisation may provide data that meets quality standards for some measures, but not others. Submitting this kind of data was optional in 2021.

Unit-level employee experience data

This type of data is collected through surveys and is used across all reporting years. Duty holders have 3 options to collect employee experience data:

• undertake the annual People matter survey or the shorter People matter survey for gender equality reporting (collectively referred to in this report as the People matter survey). The People matter survey is administered by the Victorian Public Sector Commission (VPSC)
• use a third-party provider to independently administer the survey
• conduct the survey in-house.

In this report, only People matter survey data is used to calculate survey results. This is because duty holders who do not use the People matter survey do not submit raw data to the Commission, limiting the analysis that is possible. In every reporting year, around 90% of duty holders use the People matter survey to provide employee experience data for their workplace gender audit. This means People matter survey results are broadly representative of the sector. People matter survey data in this report uses anonymous individual data from participating duty holders. This includes:

• 2021 data from 106,069 people across 271 organisations
• 2023 data from 131,852 people across 276 organisations
• 2025 data from 114,396 people across 281 organisations.

It is important to note 3 things in relation to the People matter survey.

First, because so many public sector employees take the survey, small percentage changes can mean a large number of people have experienced a change.

Second, VPSC administers the People matter survey, not the Commission. The Commission works with VPSC to include the questions required by the Commission for audits. VPSC then provides the Commission with raw data from each participating duty holder. This means that the Commission’s results database includes individual responses to the required workplace gender audit questions. For the Baseline report (2021), the Commission did not have access to the raw data for all duty holders. This is because VPSC initially reported the results to the Commission as a percentage of respondents. Items with fewer than 10 responses were suppressed by the survey provider to protect respondents’ privacy. In late 2022, the Commission received the unit level individual data. The Commission has used this raw People matter survey data for analysis in all insights reporting since the Baseline report. Raw People matter survey data allows the Commission to calculate percentages for different groups across the whole dataset.

Finally, the Commission’s People matter survey dataset is a different sample to that used by VPSC. VPSC’s reporting of People matter survey results does not include local government organisations, universities or library corporations. It does, however, include public sector organisations with fewer than 50 employees that are not duty holders under the Act. These differences can lead to discrepancies between analysis from VPSC and the Commission.

For more information on how the Commission uses the People matter survey to produce data insights, see the methodology section of Intersectionality at work (2023).

Intersectionality

Intersectionality seeks to reveal how discriminatory systems such as sexism, racism, classism and ableism interact to produce new forms of inequality. The Commission focuses specifically on intersectional gender inequality. This is where gender inequality intersects with other forms of inequality to produce unique forms of discrimination.

The Act requires that, where possible, duty holders should collect data about additional forms of disadvantage or discrimination that a person may experience based on:

• Aboriginality
• age
• disability
• ethnicity
• gender identity
• race
• religion
• sexual orientation
• other attributes.

The Act requires duty holders to use this data to consider how gender inequality may be compounded. Duty holders must then use this data to drive positive and transparent progress towards intersectional gender equality.

Intersectionality focuses on the systems and structures that create and sustain inequality. You can read more about intersectionality, including how to measure and apply it, in the Intersectionality at work report (2023).

The Commission adopts a ‘gender-plus-one-other-attribute’ approach to intersectional analysis. This approach offers a practical starting point for exposing previously hidden patterns of intersectional inequality. However, it is not the final goal. As the Victorian public sector improves its ability to collect demographic data, we will also be able to analyse more complex data combinations that better reflect the experiences of all Victorians.

As was the case in 2021 and 2023, only a small proportion of duty holders returned reliable demographic data beyond gender and age in 2025. Because of this, the Commission relies on anonymous People matter survey data to analyse experiences of intersectional inequality.

This report offers intersectional analysis for 2 workplace gender equality indicators: gender pay gaps and sexual harassment. Gender pay gaps and sexual harassment are regulatory focus areas for the Commissioner. Both continue to be the cause of the most significant and persistent gendered harms. They also offer the greatest opportunity for meaningful and systemic change.

The Commission has not assigned progress ratings to measures with an intersectional lens due to complexities in the data and inconsistent reporting.

Data measures

In the 2021 Baseline report and the 2023 Intersectionality at work report, the Commission established the baseline for intersectional gender inequality in the Victorian public sector. This current report publishes a suite of measures to track progress towards a more gender equal Victoria over time.

Five years into the implementation of the Act, the Commission can now provide more detailed information about:

• what good progress looks like against measures under each workplace gender equality indicator
• how to best measure progress at industry and sector levels.

To arrive at the suite of data measures presented in this report, and produce their corresponding progress ratings (where relevant), the Commission:

• undertook an extensive review of local and international literature, conducted internal consultation, and commissioned expert advice from Associate Professor Leonora Risse
• assessed which datapoints collected under the workplace gender audit are the most meaningful to understand the state of gender (in)equality against each workplace gender equality indicator
• considered how progress could be rigorously assessed for each measure across multiple audit years.

Understanding what ‘good’ looks like on each data measure and how to interpret changes in the data as evidence of progress is a complex process. It involves first determining what the numerical gender equality ‘goal’ is on each measure. It then involves identifying whether changes in that measure can be understood as ‘positive’ for gender equality, based on the direction and size of the change. Overall, the Commission designed 4 types of progress ratings. These include matrix ratings, holistic ratings, proportionality ratings and gender-composition ratings.

Duty holders can use these measures to track their progress against their industry group to understand how their performance aligns or deviates from industry averages. However, it is important to note that measures at the industry and sector level do not always meaningfully translate at the organisational level. Duty holders should consult the Commission’s progress reporting guidance when seeking to assess organisational progress.

Gender gaps

This report presents several ‘gender gap’ measures to reveal structural inequalities. The most familiar of these will be the gender pay gap measure. These measures are calculated differently depending on the type of data being reported. For measures where women’s or men’s experience sits along a spectrum (for example, pay, or weeks of parental leave taken), the Commission calculated the difference between the average for men and women (for example, average earnings or average weeks of parental leave). This outcome was then divided by men’s average (earnings or parental leave) and expressed as a percentage.

A different approach is used for measures based on proportions. These measures compare the proportion of women with the proportion of men who experience, access or participate in a particular workplace arrangement or opportunity. Examples include:

• the uptake of part-time work
• promotion rates
• access to development opportunities
• the uptake of flexible work.

For these measures, the gender gap is expressed as the percentage point difference between men and women.

Values greater than zero indicate that men have the more favourable outcome for the measure being reported (that is, they are promoted at a greater rate, or are paid more), while values less than zero indicate that women have the more favourable outcome.

Matrix ratings

The Commission developed matrices to assign progress ratings for most measures. Three types of matrixes are used in this report to assess progress on three different types of data measure:

• Matrix 1: Gender composition ratings. Used to assess progress on measures that look at the gender split of a population. This includes, for example, changes in overall gender composition at sector and industry levels.
• Matrix 2: Directional change ratings. Used to assess directional change against measures such as the sexual harassment occurrence and reporting rates, perceptions of fairness in recruitment, and promotion rates.
• Matrix 3: Gender gap ratings. Used to assess whether gender gaps are narrowing or widening.

Each type of matrix was designed to rate both the movement (the degree of change over time, or how far an industry or the sector has moved towards or away from gender equality) and the overall performance (the final landing position in 2025).

The matrices calculate the degree of change over each 2-year reporting period. This means that the 4-year percentage point change presented in the Tables in the report is divided by 2 before being entered into the matrix.

Below, each matrix cell provides further information about how to interpret progress scores. Each cell includes:

• The overall assessment (Excellent – Moderate – Poor – Very poor)
• The movement assessment (Substantial progress – Good progress – Stable – Stagnant – Slight decline – Worsening)
• The performance assessment (Equality achieved – Equality maintained – On track – Not on track)
• The numerical assessment (1 – 20).

Matrix 1 – Gender composition ratings

Matrix 2 – Directional change ratings

Matrix 3 – Gender gap ratings

Holistic ratings

The Commission identified 3 areas that require a more holistic interpretation of progress. These are:

• uptake of part-time work
• uptake of flexible work
• uptake of parental leave.

In these cases, it was not possible to assign an ideal outcome and produce a corresponding matrix. This was because assessing progress relied on interpreting 2 distinct measures in relation to one another. For example, in the case of parental leave, there is no ideal number of weeks that men and women should be taking. Instead, to assess progress, expert assessors considered changes in the duration of leave by gender, combined with whether the gender gap in the average number of weeks of parental leave taken was getting bigger or smaller.

The progress rating scale for measures assessed holistically mirrors the matrix rating scale of:

• Excellent
• Moderate
• Poor
• Very poor

Holistic measures are indicated in the tables in the report by including (Holistic) after the progress rating.

Proportionality ratings

Proportionality ratings are used in Indicator 1 to assess progress on the gender composition of senior leaders in relation to the overall gender composition of the workforce. In these cases, the progress rating scale is:

• Improving (assigned when the 2025 data for the measure has moved closer to the overall gender composition since 2021)
• Proportional (assigned when the 2025 data for the measure is within +/-5% of the overall gender composition)
• Worsening (assigned when the 2025 data for the measure has moved further away from the overall gender composition since 2021).

Gender-concentration ratings

Gender-concentration ratings are used in Indicator 7 to indicate how gender-concentrated an occupation or industry is. They are comprised of two parts. The first part summarises the change:

• More segregation
• No change
• Less segregation

The second part relates to the density of the gender concentration:

• Highly men-concentrated (more than 75% men)
• Moderately men-concentrated (more than 65% to 75% men)
• Slightly men-concentrated (more than 55% to 65% men)
• Gender-balanced (45% to 55% women; 45% to 55% men; up to 10% any gender)
• Slightly women-concentrated (more than 55% to 65% women)
• Moderately women-concentrated (more than 65% to 75% women)
• Highly-women concentrated (more than 75% women).

Data caveats

The Commission holds the most comprehensive public sector workplace gender equality dataset in Australia. There are certain complexities to such a large dataset and the way the Commission handles it.

Data quality and consistency

The Commission’s data can shift slightly over time. This is because duty holders can update their audit submissions if they identify ways to improve or correct their data. Resubmissions can lead to changes in the underlying data used in the Commission’s reports. As such, data in this report may be different from previous releases. The data used in the current report was locked on 11 May 2026.

For each measure calculated, the Commission only includes the audit results from a given duty holder if their underlying data passed the Commission’s data quality standard. For example, if there were issues in an organisation’s base salary data, their audit is excluded from base salary pay gap results.

Survey participation and comparability between years

Participation in the People matter survey is not identical between survey rounds. The mix of organisations responding, as well as the proportion of employees from each organisation who complete the survey, differs from year to year. The most notable variation in this cycle was in public health, which made up a smaller share of responses in 2025 than in 2023. Total responses were broadly stable across rounds.

The Commission tested whether this affected the sexual harassment experience results by reweighting 2025 to the earlier participation mix and by excluding public health from all years. Results were consistent across all approaches, and the findings reported here hold regardless of method. Because public health was under-represented in 2025 and reported rates in that part of the sector are above average, the 2025 sexual harassment experience estimates are likely to be conservative.

Improvements to the workplace gender audit

Since the Act came into effect in March 2021, the Commission has continued to refine its data collection through improved data definitions and processes. This means that some measures are newer than others and may not be available across all reporting years. For example, the senior leader category was introduced to the audit in 2023. You can review the definition of senior leader in the audit guidance.

In other cases, the Commission updated guidance materials to improve duty holder interpretation of a measure, resulting in vast differences in data reported. For example, from 2023 onwards, the guidance for calculating the average number of weeks of parental leave taken was updated. For this reason, 2021 parental leave data is excluded from this report.

Many fields will become mandatory to report fully in 2027 audits. As a result, 2025 results may not be of a sufficient quality to publish for many duty holders. See the list of fields changed to mandatory in 2025 to learn more.

While it has always been mandatory to provide remuneration data, the Commission has improved its standard in 2025. This means that the Commission has excluded data from some organisations in 2021 and 2023 where it didn’t meet this revised standard.

Gender diverse and intersectional data

The data in this report is limited in relation to employees of self-described gender. Many duty holders still only collect gender data as male/female or men/women. This means that often the sample size for people who self-describe their gender is too small to analyse meaningfully. Where the data was reliable, the Commission included insights beyond just the categories of women and men.

The Commission does not assign progress ratings for data about people who self-describe their gender. This is due to incomplete data and complexities in calculating progress for this group.

Similarly, many organisations are unable to provide demographic data beyond gender and age. This limits the extent to which the Commission can undertake intersectional analysis. The People matter survey provides the key data source to consider forms of intersecting inequalities.

Please email: enquiries@genderequalitycommission.vic.gov.au if you would like more detail on the Commission’s methodological approach.