UCSB PGDCSM (Civil Service Management) study materials
UCSB PGDCSM (Civil Service Management) study resources

UCSB PGDCSM (Civil Service Management): Government Data Governance and Information Management

UCSB PGDCSM (Civil Service Management) · Work through your exam, one topic at a time.

All guide sections

Government Data Governance and Information Management

This domain tests whether you can treat data as a managed asset across its lifecycle and against legal constraints - from creation, through sharing and retention, to publication or disposal.

Core rules and concepts

  • Assign roles explicitly: an accountable data owner, data stewards who maintain definitions and quality, and a chief data officer for strategy. Governance is a set of decisions and rights, not a software purchase.
  • Metadata is a deliverable: data dictionary, business definitions, lineage, and sensitivity classification. Undocumented data cannot be defended, shared, or safely joined.
  • Measure quality on stated dimensions - accuracy, completeness, consistency, timeliness, validity, uniqueness - rather than asserting that data are clean.
  • Records management: apply retention schedules, honour legal holds, and dispose defensibly. Deletion on schedule is compliance, not loss.
  • Privacy: collect the minimum, publish a notice, secure it, and keep it no longer than needed. Aggregation is not anonymisation when cells are small.
  • Access: freedom-of-information regimes create a presumption of release subject to enumerated exemptions; privacy law separately governs an individual's access to their own record.

Worked example

An agency wants to publish a dataset of service locations. Sequence: confirm no legal hold applies; classify each field for sensitivity; suppress small counts and generalise precise addresses near sensitive facilities; attach metadata and a licence; log the release and the reviewer. Reasoning: publication is irreversible, so sensitivity assessment must precede de-identification, and de-identification must be tested against differencing attacks rather than assumed because names were dropped.

Traps that cost marks

Treating metadata as optional; assuming aggregated tables are anonymous; confusing freedom-of-information access with a privacy right of access to one's own record; retaining data indefinitely "just in case" instead of following the schedule.