Across the accounts, contributors commonly describe interconnected concerns involving perceived favouritism, access to development and promotion, bullying or belittling behaviour, limited responses to concerns raised internally, and impacts including anxiety, distress, resignation or loss of employment.

A single account can be coded to several themes at once, which is why the shares add to well over 100% and cannot be added together.

The clustering visible here is a property of the material received, not a measurement of the workforce. A theme raised less often has not been ruled out or ranked as less serious — it has simply been described less often by the people who chose to submit.

• 01Submitted records were preserved in the state received.
• 02Test, filler and other non-substantive entries were removed before analysis.
• 03Each remaining account was read and coded to one or more recurring themes.
• 04Counts were tallied per theme and expressed as a share of the accounts reviewed.
• 05No account was quoted, attributed or identified in this update.

Contributors chose to come forward, so the material cannot be treated as a representative sample of the workforce. Nothing here has been verified unless expressly stated otherwise.

Submission volumes are not published at this stage, so the shares describe the accounts reviewed and nothing beyond them. Theme labels summarise what a code covers — they are not quotations, and they do not assert that any described event occurred.

Careful follow-up, evidence review and clarification with contributors who have given permission to be contacted.

Contact permission, verification, consent and publication permission continue to be treated as four separate decisions. Giving one does not give the others, and any of them can be withdrawn.

No contributor has been identified or quoted in this update.