Informed, not influenced.

A source you can inspect

Official records establish specific facts, such as who holds a seat or how a member voted. Stated positions quote or summarize an individual’s own materials. Documented affiliations identify a dated membership relationship. Explanatory analysis is labeled separately and requires owner review before publication.

A vote on a procedural motion or a bill with multiple subjects is never automatically translated into a broad policy stance. Read the question, legislation, date, and linked record together.

Visitor analytics

Pollscape uses Google Analytics to understand aggregate visits, popular pages, and where visitors arrive from. Google Analytics loads only if you choose “Allow analytics.” It receives page URLs, referrers, and browser and device data and may use cookies. Pollscape records your choice in this browser. You can change it here at any time; declining stops future Analytics requests from this browser. Pollscape does not send address lookup input or ChatGPT sign-in identity as analytics events. Learn how Google uses data from sites using its services.

Missing evidence is information about our research

• Research pending: we have not completed a sufficient source review.
• No public position found in reviewed sources: a completed, dated review did not find a position in its defined source set. It does not mean no position exists.
• Qualified position: the position includes conditions, exceptions, or conflicting evidence that should be read together.

We retain historical statements and dated votes. New evidence does not silently erase an earlier position. Editorial revisions retain the prior published text.

Current coverage

House non-voting delegates and the resident commissioner are shown separately from the 435 voting seats. Independent members retain their party label. Caucus affiliations overlap; membership totals cannot be summed across organizations.

Nationwide contest inventory is separate from ballot candidate research. Known regular House/delegate and Class II Senate general elections, plus verified Florida/Ohio Class III special elections, retain official sources and election-office links. Puerto Rico’s four-year Resident Commissioner office has no regular 2026 contest. Other House special contests remain outside this inventory until verified. Candidate listings remain researched only for New York, Pennsylvania and North Carolina; FEC registrations never establish ballot certification. Inventory does not validate election geography or same-number incumbency. Missouri’s superseded Census 120th map and address assignment are withheld with the September 25 court-order notice.

Current representation uses the 119th Congress. The 2026 election map uses the 120th Congress boundaries. Census maps are simplified for display; address lookup uses a spatial boundary query. Election authorities remain the source for ballot and voting information. ZIP lookup uses the official House finder for current districts; split ZIPs list possibilities. A mailing ZIP may differ from your residence. ZIP results do not establish a 2026 House election district: use a residential street address. ZIPs go to the House, addresses to Census; Pollscape does not persist either, put them in the browser URL or include them in Analytics events. Anonymous ZIP results may be cached in memory for 15 minutes. The providers may log their requests. House ZIP lookup scope.

Issue sections are temporarily hidden from navigation, profiles, and race comparisons while we improve their design and evidence coverage. Existing issue pages remain accessible. Issue research is incomplete, especially challenger platforms. Issue-linked vote evidence remains selective. Separate profile voting histories cover published official roll calls in the 118th and 119th Congresses through their displayed retrieval dates. Candidate lists preserve each authority’s certification or unofficial status. New York District 24 has a litigation notice in the certification.

Voting history

Profiles show official House Clerk and U.S. Senate recorded floor roll calls in the 118th and 119th Congresses. This includes bills, amendments, procedural motions, nominations, quorum calls and Speaker elections when recorded. Voice votes, unanimous consent, committee votes, earlier Congresses and former-member profiles are outside this coverage. Records may lag proceedings; each session’s source and retrieval date are shown.

House records join directly by Bioguide ID. Senate LIS identifiers are mapped to current members using the official Senate member feed. We retain only votes explicitly listing that person, including service in another chamber. We never fill a missing entry with “not voting.” Members not yet sworn in, members who left office, and non-voting offices therefore do not accumulate inferred missed votes.

Recorded entries include “not voting,” “present,” and named Speaker choices. “Not voting” reproduces the official label and does not establish a reason or an attendance rate. Overall tallies include all listed voters, including former members; vice-presidential tie breakers are displayed separately. Yea/Aye and Nay/No share search filters, while the original choice stays visible. On Speaker ballots, the Clerk’s result field may name a candidate without establishing an election; the profile and JSON explicitly flag that limitation.

Vote-type filters are navigation aids derived from the official question text: quorum (including calls by states) and Speaker elections first; cloture, motions to proceed, table, recommit, reconsider, commit, instruct conferees, go to conference, close portions of a conference, require attendance, decisions of the chair, order the previous question, adjourn, recess, waive, discharge, refer, postpone, appeal and points of order are procedural. Resolutions whose titles provide for consideration are also procedural. Remaining nomination questions are nominations; passage, amendments, resolutions, concurrence, conference reports and veto questions are legislation and amendments. Unrecognized questions remain “Other.” A category does not establish a policy stance, final enactment, or importance; always read the exact question.

Search ignores case, spaces and periods in titles, questions, identifiers and results. Counts cover the entire selected Congress before filters, while matching entries describe the filtered list. Refreshes validate complete session inventories and member objects before replacing a publication; failures retain the last successful version. Party alignment uses the separate rules below; no cooperation, attendance or ideology score is introduced.

Party alignment · methodology v1

We reconstruct party tallies from every voter listed in the official roll call, including former members. The party label comes from that roll call, not today’s roster. Yea/Aye and Nay/No are equivalent for this calculation. Independents and other parties stay separate; we do not infer caucus membership. Vice-presidential tie breakers do not enter party tallies.

An eligible vote requires a Democratic or Republican member casting a yes/no vote and strict majorities of the two parties’ recorded yes/no voters taking opposite sides. Present and “not voting” do not enter either party’s yes/no denominator. Ties, no yes/no votes in a party, and agreement between party majorities are excluded. The member’s own vote is included when determining their party’s majority; the majority is not an official party instruction.

Alignment is votes with the member’s party majority divided by eligible votes, multiplied by 100 and rounded to one decimal place. The numerator and denominator remain visible. Zero eligible votes produces no percentage. Exclusions form a mutually exclusive sequence: unavailable party breakdown; quorum or Speaker election; non-yes/no member choice; independent/other/unknown member party; party tie or no yes/no votes; party majorities agreeing. Remaining votes are with or against the party majority.

Every eligible roll call has equal weight. “All vote types” includes procedural votes, nominations, legislation/amendments and otherwise unclassified yes/no questions. It excludes quorum calls and Speaker elections. The vote-type filter applies to both comparison periods; text, choice and alignment filters affect only the record list. Type labels follow the navigation rules above, so a generic motion can remain “Other.” A procedural vote is not necessarily a position on the underlying policy.

The 118th and 119th Congresses are displayed separately, further split by chamber and party label. We show the dates of actual records, sample sizes and source retrieval dates. The 119th is in progress. Different agendas, service periods and vote mixes mean a percentage change alone does not establish a position change or increasing polarization. No common-agenda adjustment or causal inference is made. Small samples remain visible rather than becoming an unlabeled rating.

Members of non-voting offices are included only where the source explicitly records a vote. Their limited voting eligibility makes comparisons with full voting members inappropriate. Missing records are never inferred as nonparticipation. The departure list and its party tallies link to the original records. These arithmetic measures are not cooperation, effectiveness or ideology scores.

Cross-party sponsorship · methodology v1

Within the selected Congress’s H.R./S. bills, we count a Democrat’s sponsored bill if it has at least one active Republican cosponsor, and a Republican’s if it has at least one active Democratic cosponsor. For cosponsorship, we count an active D/R member’s bill when its sponsor carries the opposite D/R label. Each bill counts once, regardless of the number of cosponsors. These are separate measures, not a combined score.

Parties, sponsor identity and cosponsor lists come from the retrieved GovInfo bill-status record, including people no longer in office. These records do not supply a complete party-effective-date history; we do not claim the label establishes party at introduction or joining. Conflicting party labels for repeated entries of the same cosponsor are treated as unknown. Current roster affiliations are never substituted. Independents are shown separately and not assigned an “opposite” party.

The latest dated entry determines active versus withdrawn status using the same deduplication rules as the bill list. Rejoined cosponsors count once if their latest entry has no withdrawal. The sponsored denominator includes all bills with a D/R sponsor; its numerator means at least one documented opposite-party active cosponsor, not proof that every affiliation is known. The cosponsored denominator requires active status and both member and sponsor D/R labels. Other/unknown labels and withdrawals outside these rules are explicitly excluded. Party counts, original named cosponsor links, record updates and retrieval dates are inspectable.

These are relationships in the source snapshot, not a timeline of negotiations or a measure of willingness to compromise. They do not prove joint authorship, agreement on every provision, legislative success, or policy positions. Earlier withdrawn support remains in the underlying history but does not count as an active D↔R tie.

Sponsored and cosponsored bills

Member profiles cover House bills (H.R.) and Senate bills (S.) in the 118th and 119th Congresses. Resolutions and amendments are excluded. Counts apply to the selected Congress, not a career, and can include a member’s service in another chamber. A member who did not serve during a Congress may have no recorded bills.

Congress.gov bill-status records are distributed through GovInfo bulk data. We match official Bioguide IDs, check each archive against its source inventory, and publish a complete member snapshot. Archive retrieval dates and individual record-update dates are separate. Latest actions reproduce source text and may lag proceedings; a missing action remains explicitly unavailable.

We count each bill once per member. Sponsorship, active cosponsorship and withdrawn cosponsorship are separate. Active means the latest dated cosponsorship entry has no withdrawal, not that the bill is still pending. Repeated entries are deduplicated; withdrawal/rejoin history is retained, ordered by join date and then withdrawal date. Source chronology inconsistencies are labeled without inventing corrected dates.

Sponsorship is not a vote, agreement with every provision, an ideology score or a measure of effectiveness. Failed refreshes retain the previous publication; unavailable coverage is never displayed as zero. Congresses outside this two-Congress window are not covered.

Primary sources

• House Clerk: membership and roll calls
• Senate: current senators and roll calls
• Congress.gov: legislation and Congressional Research Service
• New York Board of Elections
• Pennsylvania: unofficial post-primary candidate report
• North Carolina State Board of Elections
• Census TIGERweb: state and congressional boundaries
• GPO Congressional Pictorial Directory: portraits · reuse information

Supplemental office links and term identifiers use the unitedstates/congress-legislators community dataset and are checked against the official roster. Portraits are government materials unless a source indicates otherwise; unavailable images have a labeled fallback.

Districts and candidate photos

Map colors describe current seats. Election outlines never imply a current officeholder or future winner. Senate state bands represent one seat each. County descriptions distinguish whole and partial counties; address lookup resolves a particular home.

Candidate photos use verified government or licensed images with attribution. When a usable portrait cannot be verified, an explicit fallback appears. Campaign website discovery is separate from ballot certification.

People and representation

Committees, leadership, contacts, and service

Institutional assignments and office contacts come from official House and Senate sources. Committee pages and member profiles share the same listed assignments. Joint rosters include both chambers; the Helsinki Commission’s congressional roster excludes its three executive seats, whose dated status is explained separately. The inaugural roster records January 20, 2025 event service, not ongoing assignments. Official hearing and event-page supplements preserve source dates and stop for review if their roster blocks change. Partial coverage and source notes remain visible; unlisted ex officio participation is not inferred. Leadership covers serving members, not non-member officers. A retrieval date is not an appointment date.

Term history is a separately attributed community-maintained congress-legislators compilation. Approximate combined service unions recorded terms across chambers, excludes gaps and overlaps, and counts no service beyond the source retrieval date. Scheduled term ends remain labeled. Official office links provide contact options; they are distinct from campaign contacts.

Member characteristics come from public biographies and documented institutional records, including House History. We never infer identity from names, photographs, party, or caucus membership. Each characteristic has its own unknown count and source dates. Vacancies are excluded from serving-member denominators; non-voting delegations are separate.

Race identities can overlap; Hispanic or Latino ethnicity is separate. Member gender is not equated with Census sex. The initial institutional source documents women; other gender values remain unknown until sourced. We do not assign a combined Asian or Pacific Islander institutional category to a narrower Census category.

2024 ACS 1-year population estimates provide demographic context. Race figures include that race alone or in combination. Age bands use adults 18 and older. Single-cell margins of error are retained; combined age-band margins are not calculated. Representation age bands use documented birth dates and the roster snapshot date. Individual profile ages use the current UTC date. Official congressional biography dates take precedence when sources disagree. Where no official date is available in our records, profiles use an explicitly attributed Congress Legislators community compilation. Selecting a profile age reveals its birth date and source.

Platform review and changing positions

Every pilot candidate has the same issue-question checklist. Capturing a campaign page is source discovery, not editorial review. Discovery excerpts stay private until an owner verifies attribution, scope, date, and qualifications and writes a contextual summary. A completed search with no attributable statement is explicitly limited to its cited sources and dates. Other gaps remain research pending.

Statements, recorded actions, and attributed analysis remain separate. Interpretations name their author. Distinct dated evidence is retained when positions change or conflict. Candidate and congressional records join through verified stable identities; they never match through an empty identifier.

Briefing and social sources

The civic briefing initially uses Congress.gov feeds and Federal Register notices. Government publication is not automatically independent reporting. Sources, dates, original links, and distinct publication-type labels stay visible. Topic filters describe subject words, not positions.

Congress.gov’s weekly report measures page views on that website, not all public attention. Its report period stays attached. Feeds refresh on demand after 30 minutes; outages retain successful results and caches older than 24 hours are flagged. Reporting feeds require suitable reuse permission and are not enabled in the initial release.

X timelines load only after a reader chooses to view them and only for accounts linked by an official congressional website. X controls availability and ordering. Social posts enter platform evidence only through source review.

Reported polling results

Pollscape displays reviewed original releases and the VoteHub polling compilation, licensed under CC BY 4.0. We filter exact election subjects and reformat the data; inclusion does not imply endorsement. Survey percentages describe voter responses, not win probabilities.

Field dates, sample/population, sponsors and reported response names are preserved. Campaign/partisan context is shown when supplied. Compiled records may omit question wording, interview method, uncertainty or some response categories; those gaps remain explicit and we never invent the remainder. Original extractions take precedence over matching compiled results. Different populations and tested matchups remain separate.

Older or hypothetical matchups may differ from the current ballot. District generic ballots are labeled as party preference; statewide generic results and primary polls are not assigned to these general-election district races. The September ECU forced-choice result is distinguished from its question allowing undecided responses. No polling average is calculated.

Polling refreshes on demand after six hours. No matching result in VoteHub and our reviewed releases means no poll located in those sources, not proof that none exists elsewhere.

Prediction market odds and volume

All 58 pilot elections have been reviewed for Polymarket and Kalshi coverage. Available contracts display market-implied percentages directly on Pollscape. These are traders’ price-based assessments of the named party outcome, not vote shares, candidate recommendations or a Pollscape forecast. Missing provider coverage remains explicit.

For verified $1 binary contracts, a $0.63 price implies 63%. Kalshi values are inferred from the last traded price when trading volume is positive; they are not claimed to match Kalshi’s headline algorithm. Polymarket uses a usable bid/ask midpoint when the spread is at most $0.10, otherwise a last trade when available, following its documented method. Sparse trades or wide spreads can limit the usefulness of these figures.

Kalshi volume counts contracts, including fractional contracts; Polymarket reports dollar volume. All-time and past-24-hour amounts are separate. Volume measures activity, not confidence, unique traders, voter support or campaign funds. We do not sum across outcomes/platforms, force odds to total 100%, or infer another party’s odds from a No price.

Contract rules determine settlement. Kalshi’s matched party contracts concern the member sworn in for the term beginning in 2027 and allow early determination; Polymarket’s election contracts have their own rules. Provider candidate display names do not change our party outcome identity. Unpublished placeholders are excluded.

Market data refreshes on demand after 15 minutes. Retrieval and provider record-update times are shown separately; neither necessarily timestamps the last trade. Last-success data remains during an outage, with refresh-failure and stale notices. Closed/non-trading and post-election contracts have no live percentage. Missing quotes remain unavailable, never fabricated zeroes.

Numeric data is fetched on the server through public read-only APIs. Polymarket charts also appear in provider-hosted frames, with buy buttons disabled and no affiliate codes. These connect your browser to Polymarket and refresh independently, so their values can differ from our dated snapshots. Charts are hidden for closed/non-trading contracts and after election day. The public race JSON contains the same native numeric observations and source context as the page; chart history is supplied by Polymarket.

Published election ratings

Inside Elections supplies the actual qualitative assessment for each pilot race through its public House/Senate rating exports. We show the original category and rating date, previous category when changed, source retrieval date and a publisher link. Solid, Likely, Lean and Tilt describe a party advantage; Toss-up describes a competitive assessment. These categories are not numerical probabilities.

Ratings refresh on demand after 24 hours. A source access failure retains the dated snapshot and is disclosed. A rating may legitimately predate the retrieval date if the publisher has not changed it. Cook Political Report and Sabato’s Crystal Ball ratings are not currently ingested. We do not average analyst assessments or generate our own forecast.

Issue timelines and comparisons

Profile issue timelines are temporarily hidden while we improve evidence coverage and usefulness. Published records and the timeline API remain available under the rules below.

Timelines place selected published evidence under a shared policy question, oldest first. Votes retain their roll-call dates, choices, bill identifiers and procedural questions. Statements, platforms, affiliations and attributed analysis retain their distinct labels, qualifications and citations. Those interpretive items appear only after editorial publication.

For evidence without a vote date, we use the source publication date when supplied. It does not establish when a position was first adopted. Retrieval and review dates never substitute for an event date. Missing or unusable dates are labeled and listed after dated records. Edits to an existing evidence item are corrections, not new political events. Timelines show current published versions, not the private editorial revision log.

The timeline API supports two records for the same person and question. The default pair uses the earliest and latest dated records available; undated items are used only when needed. Records may concern different bill versions, packages, stages or circumstances. We do not infer a reversal, consistency, compromise or ideology score from the pair. Conflicting or qualified published evidence remains visible; a reviewed finding that no public position was found is labeled as research, not a stance.

Coverage is selective. Selected Ukraine, immigration, tax, healthcare, abortion-related and Iran roll calls include every current member recorded in those official sources. Members who did not serve or were not recorded in a selected vote receive no invented entry. The daily refresh maintains this selection and retains the last successful data on failure. A zero count, a single record or a gap does not establish silence or an unchanged position. The full roll-call history is separate and is not automatically classified into issue questions.

Question-specific notes explain which proposals were selected and where research remains incomplete. Existing AI evidence concerns a Senate amendment; no directly relevant data-center roll call is selected. Broad energy or permitting bills are not substitutes. Born-alive legislation is not a general abortion-access vote; package votes are not separate votes on every provision. Cloture and discharge motions are not passage. Source qualifications remain visible when comparing records.

National finance & organizations

The finance hub covers 2025–2026 FEC congressional registrations, with a separate current-Congress view. Registration does not establish ballot eligibility. Party aggregates sum reported campaign figures with reporting counts and date ranges; they are not unique dollars entering politics. Registrations sharing an authorized committee are excluded from aggregates. Future-dated summaries are withheld.

Organization pages distinguish connected PAC contributions, the organization’s own disclosed contributions, separate refunds, and independent spending. FEC connected-organization declarations and dated reviewed identity rules establish the included relationships. Current source relationships do not prove historical affiliation. Similar names, subsidiaries, and employees are not automatically combined. Unresolved organizational disclosures remain inspectable with consolidated amounts withheld.

Committee contributions use donor-side records; linked congressional campaigns use PAS2 and other recipients use OTH. Organizational receipts use eligible non-memo ORG disclosures, including super PAC receipts. Mirrored reporting is not added twice. Signed corrections remain and repeated report-local identifiers withhold subtotals. These are covered records, not complete net-donation or amendment-chain reconciliations.

A donor’s contribution to a committee and that committee’s later spending are separate disclosed steps. We do not allocate the donor’s dollars to subsequent candidate spending or combine the steps into one total. Independent spending excludes rapid notices and memo records from its subtotals and remains separate from campaign receipts. Employee giving and undisclosed original-funder tracing are excluded.

The hub preserves reported summaries and separately shows transfer-adjusted receipts and disbursements following the FEC authorized-committee transfer method. Missing values remain unavailable. Source acquisition dates, report-through dates, identities and publication generation accompany the data.

Finance directory totals and ordering

Finance lists default to highest-to-lowest reported amounts, with unavailable subtotals last. Organization and race totals cover all matching result pages. Contributions, campaign spending, refunds and independent support/opposition spending remain separate. A connected committee is counted through its organization when both match the filters. These are sums of disclosed financial activity, including transactions between committees, rather than unique dollars entering politics.

Only usable profile subtotals enter directory amounts and party breakdowns; reporting counts disclose excluded subtotals. Race figures include only verified linked campaigns with usable summaries. Unknown affiliations and missing candidates remain explicit. Opposition spending is grouped by the party of the candidate opposed, not a presumed beneficiary.

A missing committee party field can be supplemented by its verified authorized candidate’s party or a separately labeled, sourced PAC alignment. Committee profiles retain the original missing value and link to the classification evidence. Spending ratios and similar names do not establish party affiliation.

Race finance identities include reviewed challengers and open-seat candidates, not only current members. Reporting denominators include every election-guide candidate, even when a finance identity or report is missing. The coverage roster shows ballot parties; financial totals preserve FEC candidate parties.

Campaign finance & donor records

Finance profiles cover the 2025–2026 reporting window, not a career or a full six-year Senate election cycle. FEC candidate summaries combine authorized committees and retain reported receipts, disbursements, cash, debt, contributions, loans, transfers, and refunds. Transfers may inflate combined activity. We show reported totals and transfer amounts separately rather than silently adjusting them. Summary report-through dates differ from source retrieval dates and can differ across campaigns.

We link congressional identities through the community-maintained congressional FEC-ID crosswalk, then check FEC candidate name, office, state, and authorized committee links. Unresolved matches and missing summaries remain explicit gaps. Each snapshot accounts for every current member; it does not claim every match or summary is available. Dated, sourced review notes explain specific identity exceptions and former campaign committees converted to PACs. PAC finances are not substituted for missing campaign summaries.

The committee contribution list uses the FEC committee-to-candidate bulk file and includes only non-memo contribution codes 24K and 24Z to linked authorized committees, dated within the reporting window. The filing committee is the donor; the other committee ID identifies the recipient. Signed corrections, report amendment labels, transaction identifiers, and filing links are retained. Repeated copies of the same FEC record ID are removed; transactions are not collapsed by name, date, amount, or a reused filer transaction ID.

Each refresh replaces the source snapshot rather than adding successive downloads together. We preserve the records supplied in the current FEC bulk release, including amended and termination reports. Committee overviews group these records by FEC committee ID and order them by contribution subtotal (or refund subtotal when selected), highest first, with subtotals including signed negative corrections. A repeated transaction ID within the same filing withholds that committee’s totals until reconciliation; reuse in a different report is not automatically a duplicate. This is not a reconstructed amendment history or a net-donation ranking.

Refunds appear separately, using non-memo 22Z records in the FEC committee-to-committee file. The filer is the authorized campaign returning money; the other committee ID identifies who received the refund. Shared campaign committees are counted once. Negative refunds remain visible as reported corrections. Refunds can relate to earlier donations, and donor-side corrections may overlap with recipient reporting, so we do not subtract this column to claim a net donation. Subtotals cover only the included records. Memo and conduit attribution, individual donors, transfers, and independent or coordinated spending remain outside these totals. Different reporting periods prevent direct reconciliation to the campaign summary.

The separate detail view includes itemized individual and other noncommittee contributions from the FEC individual file, with reported name, state, employer and occupation. Residential addresses are omitted. The file is a subset of disclosed contributions and does not name unitemized donors. Memo and earmarked records remain visible because they explain attribution; we do not sum them or merge people by name. Individual refunds are separate.

Outside spending is not campaign income. Regular independent-expenditure records use the FEC API’s most-recent reports, excluding rapid notices and memo entries from subtotals. Support and opposition are separate. Rapid 24/48-hour notices may recur in regular reports and are never added again. Electioneering disclosures may mention several candidates; we show the full ad cost and the FEC’s calculated equal share, without inferring support or opposition or producing a category total. Communication costs and coordinated party-paid expenses have separate covered-record subtotals. Signed corrections, source dates, filing links and coverage gaps remain visible. Reporting delays and disclosure limits prevent complete tracing of every original funder.

Funding does not prove allegiance, influence, or why a member voted. Employer names are not treated as donations by an employer. Net donor rankings, industry aggregation, complete original-funder tracing and longer history remain unimplemented.

FEC campaign summary fields · Committee contribution records · Transaction definitions · Refund reporting

Publication & corrections

Structured imports validate identities, seats, sources, and relationships before switching the published snapshot. Inconsistent updates are quarantined; a failed run retains the last successful data. Interpretive drafts remain private until the owner reviews them. A source retrieval date is not an editorial review date.

We do not publish candidate ratings, ideology scores, voting recommendations, or our own election forecasts. Attributed market-implied percentages are reported separately from polls and evidence. Single-seat scenarios require compatible current and election boundaries. North Carolina House districts are excluded until a valid mapping is established; district numbers alone do not assign current officeholders. Available scenarios hold all other seats constant and make no claim about how a candidate will vote or which caucus they will join.

Pollscape has an owner-only editorial workspace for reviewing drafts, correcting text, and recovering revisions. Public readers do not need an account.