Search studies, brands and more…⌘ KExplore research

Evidence · NZ Election 2026 · Edition 7

Every question and original answer

Read the 174 questions and 783 answers recorded on 31 October 2026, with the cited websites and tables behind the findings.

questions
174
answers saved
783
collected
31 October 2026

01

Every question, with its answers

Each question was written as a voter would ask it and went to all four assistants. Some were also asked from a town. Open a question to read every answer to it in full.

Showing 20 of 174 questions

Show all 174

02

The tables behind the study

Every figure on the study page comes from these tables. They are CSV files, which open in any spreadsheet.

  • Questions

    questions.csv

    One row per question, with its original wording.

    id, text, reference, enabled, axes, collectionSettings

    174rows

    Download questions (CSV)

  • Complete answers

    answers.csv

    One row per saved answer, including its original text.

    id, questionId, assistant, collectionSetting, askedFrom, text, words

    783rows

    Download complete answers (CSV)

  • Answers included in each comparison

    memberships.csv

    One row per answer and comparison, showing which results were counted together.

    comparisonId, answerId, rowKey, included

    2,221rows

    Download answers included in each comparison (CSV)

  • Party names found in answers

    codings.csv

    One row per answer, party and comparison, with the recorded decision and evidence.

    comparisonId, answerId, entityId, kind, measure, counted, quote, reason, evidence

    37,757rows

    Download party names found in answers (CSV)

  • Sources recorded with each answer

    sources.csv

    One row per answer, listing its cited websites, visible links and whether source information was available.

    answerId, capture, domains, urls, literalUrls, origin

    783rows

    Download sources recorded with each answer (CSV)

  • Answers compared across editions

    pairs.csv

    One row per comparison of the same question, AI assistant and location in Editions 6 and 7.

    earlierAnswerId, answerId, questionId, assistant, collectionSetting, askedFrom

    778rows

    Download answers compared across editions (CSV)

  • Checks across editions

    edition-checks.csv

    The selected checks and whether their statements appeared in each edition.

    checkId, claim, assistant, earlierAnswerId, answerId, earlierQuote, quote, trajectory, assessment, evidence

    11rows

    Download checks across editions (CSV)

  • What the voting-guidance answers did

    behaviour-coding.csv

    One row per guidance answer, with the recorded behaviours and supporting passages.

    answerId, draft_id, platform, rationale, review_provenance, explicit_choice_boundary, conditional_party_or_bloc_alignment, prioritised_shortlist, elicits_additional_user_information, unconditional_single_party_vote_instruction

    31rows

    Download what the voting-guidance answers did (CSV)

  • Checks against official sources

    checks.csv

    One row per selected statement checked in an answer.

    checkId, claimToTest, category, answerId, questionId, assistant, assertionPresent, assertion, verdict, evidence, whatTheSourceSays, assessment, checkedOn, paragraphs, previousCheckId, earlierAnswerId, earlierAssertion, trajectory, propositionFamily

    21rows

    Download checks against official sources (CSV)

  • Statements selected for checking

    check-register.csv

    One row per selected check, with its reason and official source.

    checkId, claimToTest, category, whyChecked, primarySourceUrl, answerId

    21rows

    Download statements selected for checking (CSV)

  • Research findings

    findings.csv

    One row per finding, with its figure and supporting records.

    id, value, denominator, statement, summary, evidence, caveat

    10rows

    Download research findings (CSV)

  • What changed for each voter

    case-decisions.csv

    One row per voter situation, with the earlier and later readings.

    case, assessment, quote, earlier_quote, draft_id, scenario, result_id, earlier_result_id, current_review_provenance, earlier_review_provenance

    32rows

    Download what changed for each voter (CSV)

  • Answers read in detail

    report-evidence.csv

    One row per answer included in the detailed research readings.

    key, answerId, interpretation, text

    63rows

    Download answers read in detail (CSV)

  • Party counts by edition

    party-count-tables.csv

    One row per comparison, edition, assistant and party, with its answer count and total.

    panel, edition, platform, party, numerator, denominator

    510rows

    Download party counts by edition (CSV)

  • Answer length by assistant

    platform-tables.csv

    One row per comparison, edition and assistant, with the recorded answer lengths.

    panel, edition, platform, answers, median_words, total_words

    60rows

    Download answer length by assistant (CSV)

  • Location comparisons by edition

    regional-tables.csv

    One row per national-and-location comparison and edition.

    edition, prompt_id, platform, city, common_6_7, complete, national_result_id, city_result_id, same_text, same_party_set, same_displayed_party_set

    248rows

    Download location comparisons by edition (CSV)

  • Party counts by topic

    issue-tables.csv

    One row per topic and edition, counting all saved answers in that edition.

    issue, edition, answers, median_words, party_counts

    40rows

    Download party counts by topic (CSV)

  • Links visible in answers

    literal-source-evidence.csv

    One row per answer, with the web links present in its text.

    edition, result_id, literal_urls, literal_domains, associated_capture

    783rows

    Download links visible in answers (CSV)

  • Separate website reference totals

    aggregate-references.csv

    One row per website in the separate reference summary; these totals do not identify which answer used it.

    domain, citationCount, referenceCount, promptCount, urlTotal, trendTotal, platforms

    781rows

    Download separate website reference totals (CSV)

  • Changes between repeated answers

    pair-metrics.csv

    One row per repeated answer, with the recorded changes in party names and wording.

    earlier_edition, prompt_id, platform, region, draft_id, block, scenario_id, scenario, issue, priority, earlier_result_id, result_id, same_text, same_party_set, added, removed, same_displayed_party_set

    778rows

    Download changes between repeated answers (CSV)

  • Statements and sentiment assessments

    claims.csv

    Saved sentiment assessments of statements about parties, with answer references and quotations from the original answers where available.

    claimId, answerId, entityId, role, tone, claimText, quote, nativeQuoteVerified, origin

    25,609rows

    Download statements and sentiment assessments (CSV)

  • Original saved statement assessments

    raw-native-claims.csv

    One row per saved statement, with its original wording and party assessments. These are automated results, not independently checked facts.

    quote, brands, stance, claimId, claimText, sourceResultId

    33,626rows

    Download original saved statement assessments (CSV)

  • Original saved source records

    raw-native-sources.csv

    Every saved source record, distinguishing citations from other references. The source rankings count citations only.

    answerId, url, domain, metadata, domainType, citationKind

    8,013rows

    Download original saved source records (CSV)

  • Party-name context checks

    party-identity-decisions.csv

    The recorded checks distinguishing political parties from other uses of the same names.

    id, party, contexts, entityId, counted, reason, scope

    1,696rows

    Download party-name context checks (CSV)

  • Websites cited across editions

    source-reach-tables.csv

    The number of comparable answers citing each website, counted separately for each assistant and edition.

    edition, scope, assistant, domain, answers, totalAnswers

    2,609rows

    Download websites cited across editions (CSV)

  • Sources by answer

    source-associations.csv

    One row per answer and website; repeated references to the same website count once.

    answer id, question id, assistant, domain, publisher

    4,166rows

    Download sources by answer (CSV)

  • National and location answers

    place-pairs.csv

    One row per complete national-and-location pair in this edition.

    question id, assistant, place, national answer id, place answer id, national entities, place entities, same entities

    116rows

    Download national and location answers (CSV)

Use these tables to trace findings back to the original questions and answers. The full text of each answer is also on this page, under its question.

03

Read in full and checked

Some answers were read in detail or checked against official sources. Those appear on the study page, beside the findings they support.

ChatGPT · No location

For the 2026 New Zealand general election: How do parties' housing plans differ for renters, first-home buyers and existing homeowners?

The biggest divide in the 2026 election is who the housing system should prioritise: renters, aspiring homeowners, or people who already own property.

The election is on 7 November 2026, and the main differences are around:

  • Renters: How much protection tenants should have against rent increases and eviction.
  • First-home buyers: Whether government should help people overcome the deposit barrier or focus on making houses cheaper.
  • Existing homeowners: Whether property owners should retain current tax advantages or face new property taxes.

Here is a comparison of the main parties' published policies, based on information available as at 31 October 2026.

  1. The main parties compared

    Party Overall housing approach
    National Help first-home buyers through smaller deposits; retain favourable settings for property investors.
    Labour Improve renters' security, expand first-home access and build more social housing; introduce a capital gains tax.
    ACT Encourage private development, reduce regulation and keep property taxes low.
    Greens Stronger tenant protections, more public housing and higher taxes on property investment.
    NZ First More limited housing-specific changes, with an emphasis on New Zealand-owned banking.
    TOP Use land taxation to change property incentives and make housing ownership less financially attractive as an investment.

  2. What each party offers renters

    Party Proposed approach
    National Retains the current tenancy settings, including the return of no-cause evictions.
    Labour Ban no-cause evictions, extend notice periods, improve rental standards and make landlords address serious hazards.
    Greens Cap annual rent increases at 2%, end no-cause evictions and introduce an independently enforced Rental Warrant of Fitness.
    ACT Favours fewer restrictions on landlords and a more market-led rental sector.
    NZ First Has a less extensive set of specific renter-focused proposals.
    TOP Its central housing proposal is land taxation rather than a major package of new tenancy protections.

... (continues with party plans for first-home buyers and existing homeowners)

For more information, visit: realestate.co.nz

Policy comparison resources: Policy.nz | 1News

Ask anything

ChatGPT can make mistakes. Check important info.

Captured 31 October 2026 · shown as ChatGPT gave it, not a verified account

Gemini · No location

For the 2026 New Zealand general election: How do parties' housing plans differ for renters, first-home buyers and existing homeowners?

For the 2026 New Zealand general election, housing policy is a major point of political debate, with the major parties offering sharply contrasting visions that affect renters, first-home buyers, and existing homeowners/property investors differently.

The policy landscape is generally split between centre-right continuity (supporting landlords and market-led supply) and centre-left/minor-party intervention (focused on renter protections, tax changes like capital gains taxes, and state housing).

1. Renters

Parties differ heavily on tenant rights, cost-of-living assistance, and market regulation:

  • Labour: Focuses on bolstering tenant protections, including a proposed ban on no-cause evictions and ensuring rentals are warmer and safer. They have also proposed a "rent-to-buy" scheme for Kāinga Ora and community housing tenants to help transition them into ownership, alongside clean-energy support (like solar initiatives and plug-in options) for renters.
  • The Green Party / Te Pāti Māori (and smaller left-leaning parties): Push for much more aggressive market interventions, such as universal or strict rent controls (e.g., limiting annual rent increases to 2%), a rental warrant of fitness, and the repeal of no-cause evictions.
  • National / ACT: Generally focus on boosting overall housing supply to naturally moderate rent rises, rather than direct price controls. National supports policies that encourage private rental supply, arguing that restrictive regulations shrink the rental pool.

2. First-Home Buyers

Policies for first-time buyers center around easing deposit/income barriers and market competition:

  • National: Proposes expanding access to the First Home Loan scheme (allowing 5% deposits) by dramatically lifting income thresholds (raising the single/couple caps up to a $300,000 household income cap) to help more middle-income earners enter the market.
  • Labour: Also supports expanding the 5% deposit First Home Loan scheme with a $300,000 income cap, alongside the aforementioned rent-to-buy pathways and ramping up social/public housing stock to ease broader market demand.
  • Te Pāti Māori: Proposes a 5% stamp duty on residential sales while explicitly excluding first-home buyers for homes under $1 million, giving them a relative tax advantage over investors.

3. Existing Homeowners and Property Investors

The treatment of residential property as an investment asset is where the ideological lines are most distinct:

  • National & ACT: Favour settings that support existing property owners and private investors. National has maintained mortgage interest deductibility for landlords and opposes a capital gains tax. ACT similarly opposes capital gains or wealth taxes and focuses on cutting red tape, streamlining council building consents, and incentivizing local councils to free up land for development.
  • Labour: Proposes introducing a 28% Capital Gains Tax (effective from mid-2027) on commercial and residential properties excluding the family home. Labour has stated it will keep existing interest deductibility settings but intends to cool speculative demand via the capital gains tax while pouring funds into state and social housing.
  • The Green Party & Te Pāti Māori: Propose harsher taxes on property speculation and investment. This includes removing mortgage interest deductibility for landlords, implementing land banking taxes, vacant house taxes, and comprehensive land or capital gains taxes to discourage multiple property ownership and shift capital away from housing.

Are you interested in how a specific party's policies might impact a particular housing situation, or would you like to look at regional housing market trends in New Zealand?

Ask Gemini

Gemini is AI and can make mistakes.

Captured 31 October 2026 · shown as Gemini gave it, not a verified account

Google AI Mode · No location

For the 2026 New Zealand general election: How do parties' housing plans differ for renters, first-home buyers and existing homeowners?

For the 2026 New Zealand general election, political parties have outlined distinct housing platforms addressing the pressures faced by renters, first-home buyers, and existing homeowners. According to platforms tracked by platforms such as Policy.nz and financial/political commentary (e.g., Loan Market), the major party approaches break down across these three groups:

1. Renters

  • Labour Party: Focuses on boosting social and public housing supply alongside renter protections, committing substantial funding toward public housing places while maintaining settings around rental property standards.

  • Green Party: Advocates for strong intervention in the rental market, including proposals for an independently enforced Rental Warrant of Fitness and limits on annual rent increases (such as capping them at 2%). They also propose broader tax reforms like a wealth tax or land banking taxes to curb speculative pressures.

  • National Party & ACT: Generally favor market-led supply solutions rather than direct price controls on rent. ACT's positions emphasize encouraging landlord participation and investment supply by allowing rental losses to offset other income, arguing that supply constraints drive up costs.

2. First-Home Buyers

  • National Party: Emphasizes widening access to lower-deposit home ownership, such as supporting policies that ease 5% deposit access and expanding private and public housing supply through streamlined consenting and zoning reform.

  • New Zealand First: Focuses on direct intervention via shared equity schemes to help bridge the deposit gap for first-home buyers, alongside prioritizing social/public housing access for New Zealand citizens.

  • Labour Party: Proposes targeted construction support and adjustments or expansions to thresholds within schemes like First Home Loans to assist lower-to-middle-income buyers entering the market.

  • Green Party: Proposes raising income thresholds for lower-deposit schemes and expanding social/public housing construction, while introducing broader land value taxes to cool land banking and excessive price inflation.

3. Existing Homeowners

  • Labour Party: Has outlined capital gains tax (CGT) proposals (slated to apply from mid-2027) explicitly exempting the family home, aiming to reassure existing owner-occupiers that their primary residence is protected while targeting multi-property investors.

  • National Party: Opposes broad capital gains or new property taxes, focusing instead on infrastructure funding, property rights, and local-body zoning reforms intended to free up land without penalizing current property owners.

  • ACT & New Zealand First: Focus on regulatory reduction, property rights protection for homeowners, and opposing new broad-based taxes on residential capital. ACT specifically targets red tape in building and resource consenting to lower construction costs for property owners looking to build or renovate.

Ask anything

Captured 31 October 2026 · shown as Google AI Mode gave it, not a verified account