The number needs
a denominator.
Federal voting intention among decided and leaning respondents.
Issue competence is not the same measure as vote intention, and a vote-share snapshot is not a seat model.
More than one view.
Two Pallas snapshots before election day
Reported percentages · common 0–100 scale
Liberal
Conservative
NDP
Green
April 12 fieldwork versus March 31 fieldwork. Independent samples; small changes are not established individual voter movement.
Original reports: Pallas Data · 14 Apr 2025 ↗ / Pallas Data · 1 Apr 2025 ↗
Issue assessments did not have one uniform margin
Leader assessments · reported percentages · common 0–100 scale
Trade partnerships
Cost of living
Energy and pipelines
These are separate issue questions. Unlisted answers are not redistributed; the items do not measure issue importance.
A lead with a date attached
The April 14 release described interviews conducted on April 12, more than two weeks before voting day. The earlier Pallas wave provides context, but two observations do not establish a complete campaign trajectory. Our comparison preserves the original measurements without smoothing in additional points.
The size of the issue gap mattered
Carney’s reported advantage on trade partnerships was considerably larger than on the cost of living; Poilievre led on energy and pipelines. The analytical lesson is that a broad leadership narrative can conceal distinct evaluations. None of these questions, on its own, shows which issue drove a respondent’s ballot choice.
Why there is no seat forecast here
A national share does not tell us where marginal constituencies would fall. A transparent seat projection would require an explicit geographic model, assumptions and validation beyond these toplines. This report therefore keeps the visualizations at the level the data can support.
Do not read a small change as a transfer
The Green estimate differed by two points between the two displayed waves. The data do not establish where any particular respondent’s support went, or that the change was statistically distinguishable from survey variation. We do not infer voter flows from separate cross-sectional samples.
The record behind
the report.
Source-reported methods are transcribed below. Items absent from the reviewed summaries are identified so readers can distinguish documented details from information that remains unavailable.
Pallas Data · 14 Apr 2025 · n=1,282
- Research provider
- Pallas Data
- Source report date
- 14 Apr 2025
- Fieldwork
- 12 Apr 2025
- Population
- Canadian adults aged 18+ eligible to vote federally
- Total interviews
- 1,282
- Published ballot base
- Decided and leaning voters; exact unweighted ballot base not provided in the summary
- Interview method
- Automated telephone interviews (IVR), landline and cellphone
- Weighting
- Age, gender and region; 2021 Census
- Source uncertainty statement
- Methodology reports ±2.7 pp at 95% for the full sample; larger for subgroups. The report’s commentary separately refers to a rounded 3% figure.
- Commissioning / funding
- Not identified in the reviewed summary; see original release.
- Question wording and order
- Consult the linked original release and its supporting tables. Our question descriptions are summaries, not verbatim questionnaires.
- Design effect / effective n
- Not supplied in the reviewed summary. An effective sample size and bespoke significance test are not calculated here.
Pallas Data · 1 Apr 2025 · n=1,267
- Research provider
- Pallas Data
- Source report date
- 1 Apr 2025
- Fieldwork
- 31 Mar 2025
- Population
- Canadian adults aged 18+ eligible to vote federally
- Total interviews
- 1,267
- Published ballot base
- Decided and leaning voters; exact unweighted ballot base not provided in the summary
- Interview method
- Automated telephone interviews (IVR), landline and cellphone
- Weighting
- Age, gender and region; 2021 Census
- Source uncertainty statement
- Publisher reports ±2.8 pp at 95% for the full sample; higher for subgroups.
- Commissioning / funding
- Not identified in the reviewed summary; see original release.
- Question wording and order
- Consult the linked original release and its supporting tables. Our question descriptions are summaries, not verbatim questionnaires.
- Design effect / effective n
- Not supplied in the reviewed summary. An effective sample size and bespoke significance test are not calculated here.
The source page’s URL ends with “green-2”, while its headline and body report Green support of 1%. The chart follows the published body and headline. The methodology specifies ±2.7 pp for the full sample; the exact decided/leaning base is not supplied in the summary.
Rounded percentages may not sum to 100. Missing categories are not inferred. Point-estimate gaps and arithmetic differences are not significance tests. No respondent-level records, complete weighting files or independent design-effect estimates were available for this analysis.
Inspect the evidence.
- Pallas Data · 14 Apr 2025 ↗Original poll · source date 14 Apr 2025 · fieldwork 12 Apr 2025
- Pallas Data · 1 Apr 2025 ↗Comparison poll · source date 1 Apr 2025 · fieldwork 31 Mar 2025
- Official election / referendum information ↗Voting day and jurisdiction
This selected archive reproduces reported numerical findings with attribution and adds Northbound’s interpretation. It is not an original Northbound poll, a complete polling average or a claim that Northbound published before the historical vote. No election outcome was used to generate a survey figure. Publication policy.