Public Opinion and Messaging Research

THE HEADLINE FINDINGS

On public opinion: Ontarians see homeless encampments as a serious problem (84%, Abacus Oct 2024) and rate the Ford government's handling of homelessness badly (~53-65% poor/bad, consistent across Abacus and Leger, multiple waves) — but they also prefer a compassionate, housing-focused response over enforcement by a wide margin (80% vs. 20%, Abacus). No professional polling exists, from any pollster, specifically naming HART Hubs, the Safer Municipalities Act, or supervised consumption site closures — this is unanimous across all eight independent research passes, making it about as solid a "gap in the public record" finding as this library has produced.

On messaging: the clearest, most consistently-confirmed pattern across six of eight sources is that Vancouver 2022 (Ken Sim) and Toronto 2023 (Olivia Chow) both won on paired messaging — not pure enforcement and not pure compassion, but a specific combination (Sim: policing and mental health nurses; Chow: housing supply and tenant protection). Calgary 2025 (Jeromy Farkas) adds a third, independently-confirmed data point: a candidate can run explicitly on Housing First and public safety simultaneously and win. The single most useful cautionary finding: Vancouver's ABC lost its council supermajority in a 2025 by-election specifically because it failed to deliver on the "care" half of its own paired promise (100 mental health nurses promised, 35 delivered) — a concrete example of what happens when paired messaging isn't matched by paired delivery.

A finding independently re-surfaced and now properly cross-referenced into this document — correction to this document's own first draft: this document originally described the 234,000 figure as "not previously in this library." That was wrong, and worth being explicit about rather than quietly fixing: this library's own registry already contained this exact source and figure (CL-254 through CL-256, added in earlier research). What's actually new here is not the fact itself but its connection to this document — the messaging/public-opinion context — which hadn't previously cross-referenced it. This is exactly the kind of check the multi-source review protocol exists to catch, and it caught it on this document's own first pass, not a later one.

PART 1 — POLLING, ORGANIZED BY CONFIDENCE LEVEL

1.1 High confidence (independently confirmed by 3+ of the 8 sources, consistent figures)

1.2 Genuinely new, single-source but well-attributed (worth using with normal citation care, not extra caution)

• Leger, Oct 2025: 64% support clearing encampments only if alternative housing/shelter is available; 61% say the province, not municipalities, should lead; only 38% believe camping should be allowed when shelters are full.
• Angus Reid, Oct 2024: granular issue-approval breakdown — 8% good/84% poor on poverty/homelessness specifically; 10%/84% on housing affordability; 14%/79% on drug use/addictions; 15%/75% on street crime/public safety.
• Pollara/CMHA, June 2024, n=1,053: 80%+ prefer health/social-service solutions over criminal-justice punishment for the opioid crisis; 73% concerned the crisis is worsening; 71% believe government should prioritize it.
• Liaison Strategies, Feb 20-22 2026: the most granular multi-issue Ford approval breakdown found — overall 30%/66%; housing 19%/70%; cost of living 20%/71%; healthcare 25%/70%; crime 30%/62%. Note: a separate wave (same pollster, April 2026) found 27%/68% overall — cite by exact date, not as a single number.
• Research Co., Feb 2026 (national, with Ontario breakout): 67% say homelessness has increased in their province (73% in Ontario specifically); 46% support/42% oppose immediately dismantling tent cities; 75% support better mental health investment; 79% support temporary housing; primary-cause attribution: 50% affordable housing, 46% mental health, 40% poverty.

1.3 One exception worth flagging: an SCS-closure poll may actually exist

Seven of eight sources state no SCS-closure-specific polling exists. One source (only) found a specific exception: a Leger poll, May 2025, on the closure of a "waterfront" supervised consumption site specifically — 47% opposed, 30% supported. This is flagged, not yet confirmed — a single-source find on a claim seven other independent passes missed entirely deserves direct verification before being treated as closing the "no SCS polling" gap. If confirmed, it would be a genuinely valuable, specific data point.

1.4 A claim independently verified this pass, now one of the library's stronger findings

The 234,000 figure. Directly confirmed via The Trillium's original reporting (syndicated across at least 8 regional Ontario outlets, all consistent): an internal Ministry of Municipal Affairs and Housing transition binder, prepared for a newly-appointed Associate Housing Minister, recorded an "unofficial estimate" of approximately 234,000 homeless Ontarians — about 1.5% of the population, and about nine times the Auditor General's most recent estimate. After the figure was reported and seized on by opposition politicians, ministry spokesperson Justine Teplycky stated it "does not represent the number of people experiencing homelessness in Ontario" and is "significantly lower," without explaining how the original figure was calculated or providing a replacement number. The same binder separately disclosed: 59,100 Ontario Works/ODSP recipients were homeless as of April 2024 (~6% of all OW/ODSP recipients) — correction, found July 4, 2026: this specific figure was later corrected by the government to 31,642, called "due to an error" by ministry spokesperson Teplycky in an Aug 27, 2024 follow-up story — a genuine, specific correction, distinct from the vaguer "significantly lower, no replacement given" non-denial the 234,000 headline figure received. Use 31,642 for this specific indicator going forward. [Village Media/Midland News, Aug 27 2024: https://www.midlandtoday.ca/local-news/province-says-its-data-was-botched-so-how-many-people-are-homeless-in-ontario-9438878]; shelter space was 97% occupied that same month; 176,804 households were on subsidized-housing waitlists as of December 2022; 72,700 supportive housing spaces were in use with ~30,000 more needed. Waterloo Mayor Dorothy McCabe is quoted noting "significant underreporting," particularly for women fleeing violence who avoid shelters. ✅ Independently verified this pass — high confidence, primary press reporting with consistent syndication.

1.5 A claim partially verified — use with more precise framing than the source material offered

"Ontario ranks dead last in housing starts, slowest in over two decades." Directionally accurate but the exact phrasing needs correction before use: Ontario ranked last among provinces in the Missing Middle Initiative's 2025 Provincial HOMES Report Card (an overall grade of D across 36 indicators); Toronto separately received an "F" grade for housing starts specifically from a RESCON-commissioned report (starts down 31-58% depending on the specific comparison window used); CMHC projects Ontario housing starts will fall to "near 2-decade lows" in 2026 (a projection, not yet a confirmed record). Ontario represents ~39% of Canada's population but recorded only ~27% of national housing starts in early 2025. Use the report-card ranking and the specific Toronto grade, not an unqualified "dead last, slowest in 20 years" framing — the underlying reality is genuinely bad and well-documented, but the most defensible version cites the specific report and the specific (projected, not yet realized) 2-decade comparison rather than presenting it as an already-achieved record.

PART 2 — MESSAGING CASE STUDIES, WITH ERRORS EXPLICITLY FLAGGED

2.1 The consistent, cross-validated cases

Vancouver 2022 — Ken Sim (ABC Vancouver) won decisively over incumbent Kennedy Stewart (roughly 51% to 29-36%, exact share varies slightly by source but direction and margin are unanimous across 6 of 8 sources). Signature paired pledge: 100 new police officers and 100 mental health nurses — care and enforcement together, not one or the other. Won the Vancouver Police Union's first-ever mayoral endorsement in its 104-year history. The delivery-gap follow-up, independently confirmed by two sources with matching figures: two years in, 179 police officers had been hired against only 35 mental health nurses (the goal was later quietly lowered to 55). In an April 2025 by-election, OneCity Vancouver candidates ran explicitly on this broken promise ("we were promised 100 nurses... we only got 12" per one direct quote) and won both contested seats, breaking ABC's council supermajority. This is the single most useful cautionary data point in this document: paired messaging worked to win, but the unmet half of the pair created a real, specific, quantifiable electoral cost within three years.

Toronto 2023 by-election — Olivia Chow won (37.2%, beating Ana Bailão by roughly 30,000 votes in a 102-candidate field) on a housing-first platform: vacant homes tax, mansion tax, 25,000 rent-controlled homes over 8 years, rent supplements, new shelter space. Confirmed consistently across 5+ sources. The clean counter-examples within the same race: Mark Saunders' explicit law-and-order campaign and Anthony Furey's harder-line anti-encampment messaging both underperformed badly (Furey under 5%); Brad Bradford's anti-bureaucracy framing got 1.28%.

Calgary 2025 — Jeromy Farkas won, defeating incumbent Jyoti Gondek (who finished third) by fewer than 600 votes — the first Calgary incumbent to lose re-election in 45 years. Confirmed independently by 3 of 8 sources with consistent numbers. Farkas's platform explicitly combined a Housing First strategy with wraparound supports and a public-safety/addictions-recovery message ("tough on crime and even tougher on its causes") — a third, independent confirmation that paired messaging, not a pure play in either direction, is the pattern that keeps winning.

2.2 Genuinely new cases, single-source but internally detailed and plausible

• Ottawa 2022 — Catherine McKenney lost to centrist Mark Sutcliffe (37.9% to 51.4%) after running hard and explicitly on ending chronic homelessness as a moral imperative. This is the cleanest available counter-example to "housing-first messaging wins" and deserves to stay in this document precisely because it complicates the pattern rather than confirming it uncritically — McKenney later won a provincial seat (Ottawa Centre, 2025, 55.7%), suggesting the issue plays differently at different levels of government, not that the messaging itself is inherently a loser.
• Edmonton 2025 — Andrew Knack won (38%, as an independent, in the city's first election permitting local parties) on increased day shelters paired with transit-safety measures (peace officers paired with social workers) — explicitly rejecting punitive framing.
• Calgary 2007 — an Ipsos issue-salience poll (not an election outcome, but relevant message-testing data): 91% considered homelessness an important issue; 79% net-positive toward a candidate who prioritizes it; 72% supported secondary suites locally, showing NIMBY resistance to gentle density was overstated even then.

2.3 Errors identified this pass — explicitly excluded, not integrated

Source discrepancy on Calgary 2025: one of the eight sources claims Calgary "has not had an obvious case of winning or losing primarily on homelessness" and describes Farkas as running on "broad economic themes" and being "unsuccessful." This directly contradicts three other independent, internally consistent, detailed sources (all agreeing Farkas won by <600 votes with an explicit Housing First component). The single dissenting source is treated as incorrect here — likely conflating Calgary's 2021 election (which the same source describes accurately) with the 2025 outcome.

A garbled Vancouver account, from the uploaded PDF specifically: that document describes Kennedy Stewart's party inaccurately, misnames Ken Sim's party as "OneCity Vancouver" (OneCity is actually a separate, smaller party — Sim's party is ABC Vancouver), and appears to state Stewart won ("Stewart winning by fewer than 1,000 votes"), which contradicts the unanimous account across the other six sources that Sim won decisively. This account is not used anywhere in this document.

A likely extraction error, same PDF: a cited "Ipsos: Toronto Mayor Rob Ford, 60% approve" figure almost certainly reflects a search-tool error — Rob Ford (Premier Doug Ford's late brother) was Toronto's mayor from 2010-2014 and died in 2016; he has no connection to current Toronto politics or the 2026 election. This is flagged explicitly so it is never mistaken for current data about Olivia Chow or Doug Ford — it appears to be a name-collision artifact from automated search, not a real current finding.

A general source-reliability note: the uploaded PDF's own citation list (127 numbered references) includes a large number of low-quality sources — Instagram reels, Facebook group posts, and at least one entirely unrelated item (a Unilever SEC filing). Where that document's claims are corroborated by the other seven sources, they're used here with normal confidence. Where a claim is unique to that document alone (the 234,000 figure, the housing-starts ranking, the Toronto drop-in-centre budget claim below), it was independently verified against outside sources before inclusion in this document — two confirmed above (1.4, 1.5), one still open (below).

PART 3 — A FINDING THAT BELONGED TO A DIFFERENT TASK — NOW FULLY CONFIRMED

Update, same session: this was flagged as unverified when this document was first written; it has since been fully confirmed via direct primary-source retrieval, and the result is significantly richer than the original single-line finding.

Toronto City Council voted 16-3 (7 absent) on November 13, 2025, adopting MM34.4 as amended — addressing encampments within 200 metres of a school, daycare, or playground on a 24-hour timeline, with priority given to those within 50 metres, and limiting encampment residents to a maximum of three shelter offers before removal. Corrected July 2, 2026 — this document previously said "50 metres" alone; the adopted policy's actual scope is 200 metres, with 50 metres as a within-scope prioritization tier, not the boundary itself. The confirmed primary TMMIS record names every vote:

• Yes (16): Brad Bradford, Jon Burnside, Lily Cheng, Rachel Chernos Lin, Vincent Crisanti, Paula Fletcher, Stephen Holyday, Parthi Kandavel, Nick Mantas, Josh Matlow, Amber Morley, Jamaal Myers, Frances Nunziata (Chair), Anthony Perruzza, Dianne Saxe, Neethan Shan
• No (3): Paul Ainslie, Chris Moise, Gord Perks
• Absent (7): Alejandra Bravo, Shelley Carroll, Olivia Chow, Mike Colle, Ausma Malik, James Pasternak, Michael Thompson

Two things worth flagging for the campaign's accountability mapping specifically:

1. Mayor Olivia Chow was absent for the decisive final vote — though present earlier the same evening for an earlier procedural vote on the same item (waive referral, 20-6, where she voted yes). This is a real, checkable fact; this document does not speculate about why, only notes it.
2. Brad Bradford — a declared mayoral candidate for the October 2026 election — originated the motion at a stronger threshold (200m/48hr) and then voted NO on the amendment that actually weakened it to 50m/24hr with the 3-offer limit, before voting yes on final passage of the version he'd opposed in amendment form. This is a nuanced, precise voting record, not a simple "voted for encampment clearance" headline — and precision matters given Bradford is an active mayoral candidate whose record on this exact issue will likely be scrutinized directly. Added context, confirmed July 4, 2026 against CP24's direct reporting: Bradford himself characterized the amendment as watering his motion "down to address only 5 per cent of encampments across the entire city," and separately attempted (in a second, failed amendment) to move the priority-tier anchor from 50 to 150 metres — both confirm he understood the 50m/200m distinction as substantively significant, not cosmetic. Councillors Chris Moise and Gord Perks spoke against the original motion on the floor; Moise called it "BS" and noted his own ward had no encampments in August 2025 "for the first time," crediting the existing 2024 protocol.

This finding has been routed into MAIN_DERIV_008_Ward_Accountability_Tracker.md and represents a real head start on Task H (political accountability mapping) before that task's assigned results have even returned.

PART 4 — ONE MORE CLAIM FLAGGED FOR FOLLOW-UP, NOT YET VERIFIED

"Toronto operates 59+ drop-in centres on a combined operating budget of less than $15 million." ✅⚠️ Resolved, July 2, 2026 — the count is confirmed; the budget figure is very likely a garbled version of a different, already-known number, not a new fact. The Toronto Drop-In Network (TDIN) — a member coalition, not itself a City program — is confirmed as 59 member organizations running at least 56 diverse drop-in centres across Toronto, an almost exact match to the "59+" figure. ✅ [Toronto Drop-In Network: https://www.tdin.ca/] The "$15 million" figure was not found as TDIN's combined budget anywhere searched — but this library's own HMK-036 already documents that Toronto's 2025 TSSS budget allocates $14.4 million gross specifically to "Drop-Ins and Housing Focused Client Supports" — a single City department line item, not a citywide coalition's combined operating budget across dozens of mostly-independently-funded organizations. $14.4M and "less than $15 million" are close enough that this is very likely the same figure, mischaracterized — a real City budget line conflated with a claim about TDIN's full membership. Do not cite "$15M combined TDIN operating budget" as a new fact — the real, already-verified figure is HMK-036's $14.4M TSSS line, a different and narrower thing.

HOW TO USE THIS SAFELY

• The clean, unanimous finding — no polling exists on HART Hubs, the Safer Municipalities Act, or SCS closures by name — is the single most defensible claim in this document. Eight independent research passes, unanimous.
• The 234,000 figure is now one of this library's strongest government-credibility findings — independently verified, extensively press-corroborated, and rhetorically powerful (a government's own internal document, then publicly disowned without explanation).
• The paired-messaging pattern (Vancouver, Toronto, Calgary all won on combining care + order/supply, not either alone) is the strongest messaging finding, and the ABC Vancouver delivery-gap by-election loss is the strongest cautionary companion to it — use them together, not separately.
• Do not use the Rob Ford figure or the garbled Vancouver account from the uploaded PDF under any circumstance — both are flagged errors, not data.
• Ottawa 2022 (McKenney's loss) should stay in any messaging summary — removing the one clean counter-example to make the pattern look cleaner than it is would be a real, avoidable integrity problem.

REF_002_Sources

• The Trillium (syndicated across 8 Ontario regional outlets) — the 234,000 figure and transition-binder disclosures. ✅ https://www.thetrillium.ca/news/housing/ontarios-unofficial-estimate-of-homeless-population-is-234000-documents-9341464 and https://www.thetrillium.ca/news/housing/the-ford-government-says-its-data-was-botched-so-how-many-people-are-homeless-in-ontario-9435686
• Missing Middle Initiative 2025 Provincial HOMES Report Card; RESCON Toronto-specific grading; CMHC 2026 Housing Market Outlook — housing-starts ranking. ✅ https://www.missingmiddleinitiative.ca/p/ontarios-record-low-family-housing ; https://www.cbc.ca/news/canada/toronto/toronto-housing-starts-fail-1.7641296 ; https://www.cmhc-schl.gc.ca/professionals/housing-markets-data-and-research/market-reports/housing-market/housing-market-outlook
• Eight independent LLM research outputs (user-supplied July 2026, one via uploaded PDF) — full extraction log available on request; individual polling/case-study attributions preserved in Parts 1-2 above.
• Cross-reference: MAIN_HMK_036_Daytime_Destinations_Engagement.md (for Part 4's drop-in-centre follow-up), MAIN_DERIV_008_Ward_Accountability_Tracker.md (for Part 3's council vote follow-up).

Legal/messaging review recommended before any public use of the case-study "lessons" — these are correlative observations about past elections, not causal proof that any specific message wins, and should be framed accordingly.