STEM plays an important part in the Australian Public Service (APS) across a many STEM and non‑STEM roles.
The annual APS Employee Census supplies data on workforce composition, including people working in STEM roles.
Publicly funded research agencies (PFRAs) are APS agencies whose main focus is research. PFRAs typically employ a high proportion of STEM-qualified people.
Compare numbers and employment levels of women and men working in the APS in STEM roles, non-STEM roles and health roles.
Source: Australian Public Service Commission (unpublished)
In 2023, 61% of APS Employee Census respondents were women.
16% of respondents (about 19,000 people) reported a STEM role as their main type of work. These STEM roles included:
Only 36% of APS employees working in STEM roles were women. This proportion has remained stable since 2020.
In comparison, women made up 66% of employees in non-STEM roles and 78% of health roles.
Most STEM roles in the APS had a classification level of APS 5 or 6. The proportion of women in STEM roles at that level and above was substantially lower than men.
Overall there was a higher proportion of women in EL1 and EL2 roles (57%). However, only one third (34%) of EL1 and EL2 STEM roles were filled by women.
The overall proportion of women and men at Senior Executive Service (SES) levels was about equal (55% women compared to 45% men). However, only 38% of SES-level STEM roles were filled by women.
Compare numbers and employment levels of women and men working at PFRAs in STEM, non-STEM and health occupations since 2020.
Publicly funded research agency data, 2023 (unpublished)
In 2023, 70% of all people working in the sampled PFRAs were in STEM occupations. This is the same as 2022.
Women made up 30% of people in STEM occupations, a one percentage point increase from 2022. In comparison, women made up 60% of non-STEM occupations in PFRAs.
The largest number of women in STEM were working at the EL1 level (974, or 35% of total women in STEM occupations). The largest number of men in STEM were also at the EL1 level (2,301 or 35% of total men in STEM occupations).
The gender split for STEM employees at the EL1 level was 30% women and 70% men. Although the EL1 level has the highest number of STEM employees, it has one of the lowest proportions of women in STEM roles.
The EL2 level had the largest difference between men and women in STEM roles – 22% of EL2 STEM roles were held by women and 78% were held by men. This was a slight improvement on 2022, where the gender split was 20% women and 80% men.
In 2023, women held 17 of 53 SES positions for STEM occupations. That’s 32% of SES STEM roles, a decrease from 2022. 0.6% of women in STEM were working at an SES level, compared to 1.5% of women in non-STEM occupations. 0.6% of men in STEM were working at an SES level, compared to 3.6% of men in non-STEM occupations.
In 2023, all participating PFRAs included output categories for people who did not identify as a man or woman. Limited analysis on these non-binary and gender diverse people is shown below.
There were 48 non-binary and gender diverse people across the sampled PFRAs. Three quarters (75%) of these people were in STEM roles.
While the largest number of women and men in STEM were working at the EL1 level, the largest number of non-binary and gender diverse people in STEM were working a level below, at the APS 5-6 level. This included 18 people or 50% of all non-binary and gender diverse people in PFRA STEM roles.
Other non-binary and gender diverse people in PFRA STEM roles were at APS 3-4 level, EL1 level, and EL2 level.
The APS Employee Census is a voluntary annual survey for all eligible APS employees. The data is self-reported and reflects the classifications respondents held at the time of the survey, rather than their substantive classification level.
The Employee Census asked respondents to best describe the type of work they do. We defined whether these roles were STEM, health or non-STEM based on our methodology. People in senior positions may describe their role as ‘senior executive’, in which case their responses would not be captured in STEM.
The response options for job families on the APS Employee Census correspond to the APS Job Family Framework. The 2022 Employee Census was the first to employ the 2021 edition of this framework. The STEM Equity Monitor uses the updated framework (except the ‘science and health’ job family) and has applied it to past data. This means some job families have been renamed. The job family ‘data and research’ has changed from non-STEM to STEM, as most roles in that family are now STEM-qualified occupations.
While the 2021 edition of the APS Job Family Framework combines ‘science and health’ into one job family, we split these in the STEM Equity Monitor. This is determined using benchmarks from previous years.
In 2023, 127,436 APS employees responded, representing 80% of the eligible APS workforce. This response rate was lower than 2022, when 120,662 APS employees responded, representing 83% of the eligible APS workforce. In 2021, 109,537 APS employees responded, representing 77% of the eligible APS workforce.
See Employee Census explanatory guides and agency reports on the APSC website:
This data does not cover all publicly funded research agencies. Workforce data was supplied by the following agencies:
Participating agencies defined which occupations in their agency were STEM, health and non-STEM. They based this either on the department’s methodology or through a self-determined analysis.
Agencies also aligned classification levels in their organisation to equivalent levels in the APS if they do not use standard APS classifications. In some cases classification levels were approximately aligned to reporting broadbands of APS classifications based on publicly available APS classification band descriptors.
Agencies reported numbers of employees who preferred not to disclose gender. These employees are not presented in the analysis.
Some staff in senior positions may be described as being in management or leadership occupations, so may not be captured in STEM roles/fields in the visualisations above.
Read more about our methodology and this data.
