Glossary

A short reference for the terms used across the dashboard. Definitions are intentionally plain-language and apply to Canadian federal polling unless stated otherwise.

Rolling average

A weighted blend of recent polls. More recent polls and larger samples count more than older or smaller ones, so a single outlier moves the line less than it would in an unweighted mean.

Margin of error (MoE)

The statistical uncertainty range tied to a poll's sample size. For a typical national poll of 1,000 to 2,500 respondents, the MoE is roughly ±2 to ±3 percentage points at the 95% confidence level. The MoE applies to each party's number, so a one-point change between polls is often inside the noise.

Sample size

The number of respondents in a single poll. Larger samples produce tighter MoEs; the relationship is non-linear (doubling the sample doesn't halve the MoE).

House effect

A pollster's systematic and persistent difference from the rolling average — usually a small tilt of a few tenths of a point for one or more parties. It reflects methodology choices (panel vs. phone, weighting, question wording) rather than which firm is "right".

First-past-the-post (FPTP)

Canada's electoral system — every riding elects whoever finishes first, regardless of margin. National vote-share totals can therefore translate into very different seat counts depending on where each party's votes are concentrated.

Seat projection

A modelled estimate of how a national vote-share scenario would translate to seats. MaplePolls's projection applies a uniform national swing to the 2025 per-riding result and tallies winners. It is a simplified estimate, not a forecast.

Vote intention vs. approval

Vote intention asks which party a respondent would vote for if the election were held today. Approval (or favourability) asks whether they view a specific leader positively. The two move at different speeds and can diverge — a leader's approval can rise while their party's vote share falls.

Pollster

The polling firm that designed, fielded, and published a survey. Every poll on the Recent Polls list is attributed to its pollster with a link to the original publication.

Field dates

The window during which respondents were contacted. Polls field-completed more recently get more weight in the rolling average.

Weighting

The post-fieldwork process pollsters use to align their sample with known population characteristics (age, region, gender, past vote, etc.). Different firms use different weighting schemes; this is a major source of house effects.

Undecided leaners

Respondents who say they are undecided but, when pushed, indicate a party they currently lean toward. Some pollsters reallocate leaners into the headline numbers; others publish leaners and committed voters separately. Methodology notes on each pollster's report explain which they do.

Poll aggregation

Combining multiple recent polls into a single rolling-average reading. The aim is to smooth across pollsters and across one-off sampling variation. Aggregation does not eliminate house effects, but it dampens them.