Everything You Need to Know About the Liaison Weekly Tracker

Hello, and welcome back to the Liaison press microsite, where we share polling data and key insights.

I’m David Valentin, and I lead Liaison’s public opinion research work. Since the 2025 federal election, Liaison has continued to publish a Weekly Federal Tracker. I keep seeing questions about what the tracker is, how it works, and how to read it, so this post is meant to put the basics in one place.

What is a rolling survey?

A rolling survey is a tracker that is constantly refreshed. Instead of treating each release as a completely separate one-time poll, Liaison keeps the survey moving by adding new interviews and allowing older interviews to roll out of the sample. That creates a steadier read on public opinion while still letting the tracker pick up real movement when the political environment changes.

In practical terms, the Weekly Federal Tracker is built from roughly two weeks of interviews. Each week, the oldest seven days of interviews are dropped and the newest seven days of interviews are added. With a typical national sample of about 1,500 interviews, that means roughly half the sample turns over each week, or about 750 interviews added and about 750 interviews dropped. The exact number can vary slightly depending on the final sample size and holidays.

The advantage is that the tracker is not overly dependent on a single day of interviews. The trade-off is that it should be read as a measure of the current trend, not as a stand-alone poll conducted at one isolated moment.

Does the tracker sample every day?

Yes. The Weekly Federal Tracker has daily sampling, excluding holidays. The public release comes out weekly, normally on Monday mornings at 5 AM ET, but the interviewing behind the tracker is happening day by day. Each release includes the relevant field dates so readers can see exactly which interviews are included.

Daily sampling matters because it lets us watch opinion move in a consistent way over time. It also means that unusually quiet periods, holidays, or major news events can be understood in the context of the exact field period.

What is happening in Quebec right now?

Alongside the national tracker, Liaison is currently running an oversample in Quebec for the provincial election. Those oversample results are gated for now. The oversample will continue until the Quebec election is over, at which point the tracker will return to its normal national structure.

At the current oversample level, Quebec accounts for about 750 unweighted interviews in the two-week rolling sample. Because roughly half the sample turns over each week, that means about 375 Quebec interviews are added and about 375 Quebec interviews are dropped each week. As with the national tracker, the exact number can vary slightly because of holidays and the final completed sample.

The public Weekly Federal Tracker still reports the national federal picture. The Quebec oversample is a separate election-period layer that gives a deeper read on the province while the campaign is underway.

What is the Liaison Weekly Tracker?

The Liaison Weekly Tracker is an ongoing national survey of Canadians. It measures the federal political environment over time: vote intention, leader approval and favourability, regional patterns, demographic splits, and the movement from one release to the next.

It is not a riding projection model and it is not a one-time snapshot. The point of the tracker is repetition. Asking the same core questions regularly lets us see whether the political picture is actually changing, or whether a single release is just statistical noise.

When is it published?

The Weekly Federal Tracker is normally released on Monday mornings at 5 AM ET. Each release includes the field dates, sample size, weighting information, and margin of error for that week’s survey.

How is the survey conducted?

Liaison’s federal tracker is conducted using Interactive Voice Recording (IVR) technology. Participants are reached through random digit dialling across both landline and cellular phone networks. The resulting data is weighted to match Census targets, including age, gender, and region.

Because this is a random sample of Canadians rather than an opt-in panel, people cannot volunteer to take the survey. If you are selected, we appreciate you taking the time to participate.

What does “decided and leaning” mean?

In most releases, we show both all-voter results and decided-and-leaning results.

The all-voter number includes people who say they are undecided. The decided-and-leaning number removes undecided voters and includes respondents who initially say they are undecided but then indicate which party they are leaning toward. This makes it easier to compare the current vote picture to an election-style ballot, while still showing how many voters remain genuinely undecided.

How should you read week-to-week change?

One or two points of movement should be treated with caution. A party or leader can move slightly from week to week because of normal sampling variation. The more important question is whether the movement is repeated over several releases, whether it appears in multiple parts of the data, and whether it lines up with what is happening in the political environment.

For example, a party moving up one point nationally is not the same thing as a trend. But a party moving up nationally, improving with key demographic groups, and narrowing a regional gap over several weeks is worth paying attention to.

Why do leader numbers matter?

The tracker does not just measure the ballot. It also measures how Canadians view federal leaders. Approval and favourability numbers help explain why the ballot is moving, or why it is not moving.

A party can have a stable vote share while its leader is becoming more or less popular. That matters because leader impressions often shape whether voters are open to changing their minds later.

What can the regional numbers tell us?

The tracker can show broad regional differences: Ontario, Quebec, British Columbia, Alberta, Atlantic Canada, and the Prairies. These numbers are useful for understanding where a party is gaining or losing ground, but they should not be treated as riding-level polls.

Regional sub-samples are smaller than the national sample, so they carry more uncertainty. They are best read as directional evidence, especially when they are consistent over time.

How do we know the polling is accurate?

No pollster is perfect, and every poll has a margin of error. What matters is whether a polling firm has a real track record and is transparent about how its work is conducted.

Liaison Strategies ranked #1 in accuracy in the 2025 Ontario election and #2 nationally in the 2025 federal election. We also performed strongly in recent provincial elections in British Columbia and Nova Scotia.

Is the tracker funded by the government or a political party?

No. The Weekly Federal Tracker is not commissioned by, paid for by, or shared with the Government of Canada or any political party before publication.

What standards does Liaison follow?

Liaison is a member of the Canadian Research Insights Council (CRIC), Canada’s voice of the research, analytics, and insights profession. You can read CRIC’s public opinion research standards here, and its election polling primer here.

Does Liaison publish riding projections?

No. Liaison publishes polling. Other analysts and aggregators may use polling, past results, demographic information, and assumptions about local races to create riding projections. Those projections can be useful, but they are not the same thing as riding-level polls.

Does Liaison call cell phones?

Yes. Liaison’s IVR surveys use random digit dialling across both landline and cellular phone networks.

Should you trust Liaison polling?

You should read any poll with a healthy amount of care. Look at the field dates, sample size, methodology, margin of error, and the trend over time. Do not treat one number as the whole story.

That said, Liaison’s record is strong. The Weekly Federal Tracker is published because public opinion is easier to understand when the methodology is consistent, the questions are repeated, and the results are made public.

About Liaison Strategies
Liaison Strategies is a national public opinion research firm. With 13 years of experience in Canadian polling, David Valentin, principal, has fielded hundreds of projects at the municipal, provincial and federal levels and appeared across Canadian media to discuss insights. Liaison is a member of the Canadian Research Insights Council (CRIC), Canada's voice of the research, analytics, and insights profession both domestically and globally.