Rankings Methodology
How GovScanner calculates Pollie Rankings for all 96 current WA Parliament members. Every number on this site is derived from public parliamentary records. Nothing is editorialised.
Overview
GovScanner Pollie Rankings measure parliamentary activity across 19 categories. Rankings are calculated from official Hansard transcripts published by the Parliament of Western Australia after each sitting day.
Rankings update automatically. After each sitting day, new Hansard entries are scraped, processed, and incorporated into the ranking calculations. There is no manual curation or editorial judgment involved in the ranking positions.
The underlying database function get_pollie_rankings() accepts a category parameter and returns all current members ordered by their count in that category. Ties are broken by member name (alphabetical).
Data Sources
GovScanner monitors 8 official WA government data sources. Rankings are primarily derived from Hansard, but member profiles incorporate data from all sources.
Hansard
Official transcripts of every parliamentary debate, speech, question, and procedural matter. Published after each sitting day by the Parliament of Western Australia.
Tabled Papers
Documents formally presented to Parliament - reports, petitions, correspondence, and ministerial papers.
Government Gazette
Official notices published by the WA Government - planning changes, mining tenements, appointments, and regulatory actions.
Bills
Proposed legislation currently before Parliament, including readings, amendments, and passage status.
Questions on Notice
Formal written questions from members to ministers, with published answers.
Committee Reports
Reports and inquiries from parliamentary standing and select committees.
Acts of Parliament
Legislation that has been passed by both houses and received Royal Assent.
Subsidiary Legislation
Regulations, rules, and orders made under Acts of Parliament.
AI Classification
Several ranking categories rely on AI classification of Hansard entries. Each entry is processed by a large language model (Google Gemini Flash) that assigns:
- Temperament - hot, warm, neutral, cool, or cold
- Sentiment - positive, neutral, or negative
- Urgency - low, medium, or high
- Type - speech, question, motion, petition, ministerial statement, grievance, committee, bill, or other
Classification is performed once when an entry is first processed. The same model and prompt are used for all entries to ensure consistency. The full text of each Hansard entry is provided to the model for classification - not summaries or excerpts.
AI classification is inherently imperfect. Edge cases exist, particularly for temperament and sentiment. We prioritise consistency (the same model applied uniformly) over absolute accuracy for any individual entry.
Category Definitions (24)
Each category below describes exactly what is counted and how members are ranked.
Ranked by total Hansard entries, covering all speeches, motions, petitions and parliamentary activity.
Calculation: Total count of all Hansard entries where the member is listed as a participant. Includes speeches, motions, petitions, questions, ministerial statements, and all other entry types.
Note: This is the broadest measure of parliamentary activity. A member who speaks frequently on multiple topics will rank higher than one who speaks rarely but at length.
Value shown: Entries
Members who have delivered the most speeches in Parliament.
Calculation: Count of Hansard entries with type "Speech". Includes members' statements, second reading speeches, consideration in detail contributions, and adjournment debates.
Value shown: Speeches
Members who raise the most Matters of Public Interest, the daily debate slot for urgent community issues.
Value shown: MPIs
Current members ranked by years of continuous parliamentary service.
Value shown: Years served
Members who have directed the most questions to other members or ministers in the chamber.
Calculation: Count of Hansard entries where the member is recorded as directing a question to another member or minister. Includes questions without notice and supplementary questions.
Value shown: Questions
Members who appear most frequently as the target of parliamentary questions.
Calculation: Count of times a member appears as the target of a parliamentary question. Ministers responsible for major portfolios will naturally rank higher.
Note: This category skews toward ministers, particularly those holding portfolios under scrutiny.
Value shown: Times questioned
Members who have raised the most grievances on behalf of their constituents.
Calculation: Count of Hansard entries with type "Grievance". Grievances are a formal mechanism for backbench members to raise constituency issues directly with a minister.
Value shown: Grievances
Members whose Questions on Notice take the longest to receive answers. Measures government responsiveness.
Value shown: Avg days wait
% of entries classified as "hot" in tone. Adjusted for total activity.
Calculation: Count of Hansard entries classified as "hot" temperament by AI analysis. Hot entries involve confrontational language, strong criticism, heated exchanges, or emotionally charged debate.
Value shown: % hot
% of entries classified as "cold" temperament. Adjusted for total activity.
Calculation: Count of Hansard entries classified as "cold" temperament. Cold entries are measured, procedural, factual, or deliberately restrained in tone.
Value shown: % cold
% of entries classified as positive sentiment. Adjusted for total activity.
Calculation: Count of Hansard entries classified as positive sentiment by AI analysis. Positive entries express agreement, commendation, support, or constructive proposals.
Value shown: % positive
% of entries classified as negative sentiment. Adjusted for total activity.
Calculation: Count of Hansard entries classified as negative sentiment by AI analysis. Negative entries express criticism, opposition, condemnation, or dissatisfaction.
Value shown: % negative
Members with the highest total word count across all Hansard entries.
Calculation: Total word count across all Hansard entries for each member. Higher word counts indicate more substantive or more frequent contributions to debate.
Value shown: Total words
Members with the lowest average word count per entry (minimum 5 entries).
Calculation: Average word count per Hansard entry, ranked lowest first. Only members with 5 or more entries are included to avoid skewing by members with a single short contribution.
Note: Lower average = more concise. Members who consistently make brief, focused contributions rank highest.
Value shown: Avg words (lower = better)
Members who participate most in end-of-day adjournment debates, often used for constituency matters.
Value shown: Adjournments
Average cross-references per entry, linked to gazette notices, tabled papers, and other debates.
Calculation: Count of Hansard entries for a member that are linked to other records - tabled papers, gazette entries, bills, or other Hansard entries. Indicates how connected a member's parliamentary work is to the broader legislative record.
Value shown: Refs/entry ×10
Members involved in the most bill readings and legislative activity.
Calculation: Count of Hansard entries with type "Bill". Includes first/second/third readings, consideration in detail, and committee stages of legislation.
Value shown: Bill entries
Average bill references per entry. The most legislatively-engaged pollies.
Value shown: Refs/entry ×10
Members who have moved or been involved in the most motions.
Calculation: Count of Hansard entries with type "Motion". Includes motions without notice, urgency motions, disallowance motions, and procedural motions.
Value shown: Motions
Members most active in parliamentary committee proceedings.
Calculation: Count of Hansard entries with type "Committee". Includes appearances in committee of the whole, estimates hearings, and standing/select committee proceedings recorded in Hansard.
Value shown: Appearances
Members who have tabled the most petitions from the public.
Calculation: Count of Hansard entries with type "Petition". Petitions are tabled by members on behalf of constituents and read into the parliamentary record.
Value shown: Petitions
Ministers who have made the most formal statements to the chamber.
Calculation: Count of Hansard entries with type "Ministerial Statement". Only ministers can make ministerial statements, so non-ministers will have zero entries.
Note: This category naturally skews toward government frontbench members.
Value shown: Statements
% of entries flagged as critical or high urgency. Adjusted for total activity.
Calculation: Count of Hansard entries classified as high-urgency by AI analysis. Urgent entries relate to time-sensitive issues, emergencies, or matters requiring immediate government attention.
Value shown: % urgent
Scope & Limitations
Rankings include current members only - the 96 MPs serving in the 41st Parliament of Western Australia. Former members are excluded from rankings but remain in the database for historical queries.
Hansard coverage: Rankings are calculated from all Hansard entries currently in the GovScanner database. Historical backfill of the 40th and 41st Parliament is ongoing. As more historical data is added, rankings will become more comprehensive.
What rankings do not measure: Rankings count parliamentary activity on the record. They do not measure the quality of contributions, constituency work, committee work not recorded in Hansard, ministerial decision-making, or any activity outside the parliamentary chamber.
Ministerial bias: Ministers naturally rank higher in some categories (ministerial statements, being questioned) because those activities are exclusive to their role. This is not a flaw - it reflects the structure of Parliament.
Upper vs Lower House: The Legislative Council and Legislative Assembly have different procedures, sitting patterns, and speaking conventions. Cross-chamber comparisons should account for these structural differences.
Verify It Yourself
Every data point in GovScanner is derived from publicly available government records. You can verify any ranking by:
- Checking the original Hansard transcript on parliament.wa.gov.au
- Using the GovScanner API to query raw data - see API documentation
- Viewing individual member profiles at
/members/[name]which show all Hansard entries for that member
Example API query
GET https://api.govscanner.com.au/v1/members?current=trueReturns all current members with Hansard statistics. Full API docs →
Update Frequency
Rankings update automatically after each parliamentary sitting day (typically Tuesday to Thursday when Parliament is sitting). New Hansard entries are scraped and processed within hours of publication.
The ranking calculation is deterministic - given the same input data, it will always produce the same output. There is no randomisation, weighting adjustment, or editorial override.
Questions
If you believe a ranking is incorrect or have questions about the methodology, contact [email protected].
Purple Ghost · ABN 60 923 627 375