Rating Polish pollsters: accuracy, the lead and openness
Abstract. Each Polish pollster is rated out of 100 on accuracy, its error on the gap between the two leaders, and openness, ranked against Poland's own final polls since the rescale of 13 October 2026. Opinia24 and ARC Rynek i Opinia grade B, most others C, and Research Partner and Social Changes D.
Summary
We rate each Polish pollster out of 100 on the same scale as the British ratings (Rating the British pollsters: accuracy, the Labour–Conservative lead and openness): 50 points for the accuracy of its final Sejm polls since 2001, 20 for its error on the gap between the two leading lists, and 30 for how much of its working it publishes. The record is pulled towards a deliberately poor pollster, so that a pollster with one good election cannot score well by luck. The rating is shown for readers; it does not weight the polls.
Every Polish pollster we rated is graded C, D or E. Polish final polls have missed by more than Dutch or German ones, so the "poor pollster" the ratings pull towards misses by almost as much as the British scale allows, and accuracy points are hard to earn. Openness separates the pollsters more than accuracy does. CBOS publishes more than anyone, full reports with its method, its tables and the declared vote of everyone asked, and scores full marks for openness; but its final polls have been the least accurate of the long-running pollsters, and it earns no accuracy points. Opinia24, with one good final poll in 2023 and reasonable openness, has the highest score.
What has changed since this paper was written
Updated 9 November 2026. The rest of the paper is as first published, except where a sentence below says it was corrected.
Pollster grades were rescaled on 13 October 2026. The first scale scored accuracy between fixed end points (a miss of 0.75 points or less earned full marks, 2.75 or more nothing) and pulled every short record towards a poor pollster, one that three in four final polls beat. Almost every pollster ended up with a D or an E, which told readers little. The rating now works like this:
- Accuracy (50 points) is ranked against the country's own checked final polls. If a share \(r\) of them missed by less than the pollster's average miss, it scores \(50\,(1-r)^{0.75}\): the typical (median) final poll earns about 30 and one in the best tenth about 46. Ranking within the country means a pollster is not marked down because its country is harder to poll.
- The lead (20 points) is scored the same way on the miss on the gap between the two leaders, \(20\,(1-r)^{0.75}\).
- A short record is pulled towards the typical (median) final poll, as if the pollster had two more elections there, not towards a poor one.
- Openness (30 points) is scored out of the items checked so far; an item not yet checked counts neither for nor against.
- Every pollster with a final poll on record gets a grade. One whose publications we have not reviewed is graded on its record alone (accuracy and lead scaled to 100) and marked "record only".
The grade boundaries are unchanged: A from 75, B 60–74, C 45–59, D 30–44, E below 30. Across the active pollsters of all ten countries the grades went from mostly D and E to A 8, B 68, C 76, D 31 and E 5. Where the sections below give the old end points, the poor-pollster prior or grades from before 13 October, they describe the first version of the scale; the live tables show today's.
In Poland it gives B to Opinia24 (68) and ARC Rynek i Opinia (65); C to Pollster, United Surveys, IBRiS, Ipsos and CBOS; and D to Research Partner and Social Changes. These are the figures of 9 November 2026; the table on the pollsters page is current.
What the rating is for
The rating answers a narrow question: on the evidence we can check, how much should a reader trust this pollster's published figure, and how easily can they see how it was made? It does not say which pollster is right today, and it does not feed the forecast. The Polish average gives every pollster's latest poll within 21 days of the newest the same weight (Reading Polish polls: why they are not corrected). United Surveys, which publishes every fortnight, counts no more than CBOS, which publishes monthly.
The evidence: final Sejm polls, 2001–2023
The accuracy parts of the rating use the 50 final polls in our record for the seven elections from 2001 to 2023, each with the don't-knows taken out (Measuring Polish polling error, 2001–2023). Only CBOS and Kantar have a final poll at all seven. Pollsters that have changed name carry their record: Kantar carries that of TNS OBOP and TNS Polska.
For pollster \(j\) at election \(t\), the miss \(m_{j,t}\) is its mean absolute error over the lines measured that election, and the lead error \(\ell_{j,t}\) its error on the gap between the two lists that finished first and second.
Scoring, and a poor Polish pollster
The scales are the British ones:
where the rated miss and lead error are pulled towards a poor pollster as if it had two more elections at the poor pollster's level (Efron and Morris, 1975):
The poor pollster is the 75th percentile of all Polish final polls:
A poor Polish final poll misses by almost 2.75 points per list, the point at which the scale gives nothing, so a pollster with no record starts with about two accuracy points. The lead scale is less harsh: Polish final polls have mostly got the order of the two largest lists right, and a poor one earns about ten lead points.
Two worked examples
CBOS has final polls at all seven elections, with an average miss of 3.02 points per list. Its rated miss is
beyond the end of the scale: no accuracy points. With 4 lead points and 30 for openness it scores 34, a D.
Opinia24 has one final poll, in 2023, with a miss of 1.19. Its rated miss is \((1.19 + 2 \times 2.68)/3 \approx 2.18\), worth about 14 accuracy points. With 14 lead points and 17 for openness it scores 45, a C, marked provisional because it rests on one election.
Openness
The openness items are the same in every country:
| Item | Points | Why it counts |
|---|---|---|
| Raw or unweighted figures | 8 | Shows what respondents said before the pollster's own adjustments |
| Full tables published | 8 | Sub-groups and weighting can be checked |
| Method note | 4 | How the sample was drawn and weighted |
| Sample size | 3 | The basics for judging a poll |
| Undecided share | 3 | Shows how firm the figures are |
| Fieldwork dates | 2 | When the poll was taken |
| Member of OFBOR, the industry body | 2 | Signs up to its standards |
Table: Openness items. A "partial" answer earns a third of the points; an item we could not check earns none.
What we found:
- CBOS publishes free PDF reports with the technique, dates, sample and weighting, a time series, and breakdowns by age, sex, education and place of residence. Its headline is the declared vote of those who would vote, including those who don't know; since October 2025 some reports add a distribution without them. It is an OFBOR member. Full marks.
- Opinia24's results appear in TVN24 and RMF articles with the technique, sample, dates and the undecided share, some with breakdowns by electorate. OFBOR member.
- Ipsos figures appear in OKO.press and TVP articles with the method, sample and dates. OFBOR member.
- ARC Rynek i Opinia gives sample, dates, method and the undecided share; OFBOR member.
- IBRiS and United Surveys publish through their clients (Rzeczpospolita, Onet, Polsat, Wirtualna Polska) with a method note, sample, dates and a don't-know share, but no tables.
- Research Partner and Social Changes give sample and dates in their clients' articles; the undecided share usually or sometimes.
- Pollster publishes in Super Express with the sample and dates, usually without a don't-know share.
- OGB's full reports are for subscribers; we did not use them, and its figures reach us only through Wikipedia.
Only CBOS publishes something like raw figures, the declared vote before any distribution of the undecided, and only CBOS publishes full tables.
Grades and the current table
| Pollster | Grade | Score | Accuracy | Lead | Openness | Elections |
|---|---|---|---|---|---|---|
| Kantar Public / Kantar Polska | A | 78 | 33 | 15 | 30 | 7 |
| Opinia24 | B | 68 | 36 | 15 | 17 | 1 |
| PBS | B | 65 | 23 | 13 | 30 | 3 |
| ARC Rynek i Opinia | B | 65 | 36 | 15 | 14 | 1 |
| GfK Polonia | B | 62 | 22 | 11 | 30 | 2 |
| CBM Indicator | C | 55 | 33 | 11 | 11 | 1 |
| Instytut Badań Pollster | C | 55 | 35 | 15 | 5 | 2 |
| United Surveys | C | 55 | 30 | 14 | 12 | 1 |
| Instytut Badań Rynkowych i Społecznych | C | 54 | 33 | 11 | 10 | 3 |
| Millward Brown | C | 51 | 37 | 14 | 0 | 3 |
| Ipsos Polska | C | 51 | 25 | 11 | 15 | 2 |
| Fundacja Centrum Badania Opinii Społecznej | C | 48 | 14 | 4 | 30 | 7 |
| Dobra Opinia | C | 47 | 36 | 11 | 0 | 2 |
| Instytut Badań Spraw Publicznych | C | 46 | 32 | 14 | 0 | 1 |
| IPC | C | 46 | 32 | 13 | 0 | 1 |
| OBW | C | 45 | 33 | 12 | 0 | 1 |
| Research Partner | D | 44 | 25 | 11 | 8 | 1 |
| Estymator | D | 43 | 32 | 11 | 0 | 3 |
| Social Changes | D | 40 | 25 | 9 | 6 | 2 |
| Instytut Homo Homini | D | 38 | 29 | 9 | 0 | 1 |
| PGB Opinium | D | 33 | 22 | 11 | 0 | 3 |
| Pentor | E | 27 | 20 | 7 | 0 | 2 |
| OGB | – | – | 30 | 12 | 0 | 0 |
| SW Research | – | – | 30 | 12 | 0 | 0 |
The pollsters with the most accuracy points are those with good records at the close elections of 2007 to 2015 (Millward Brown, Kantar, Dobra Opinia), and none of those publishes now. Of the pollsters in today's average, the order is set mostly by openness.
Don't-knows and each pollster's lean now
Most Polish pollsters put a don't-know share in the headline; the forecast takes it out before averaging (Reading Polish polls: why they are not corrected). Beside the rating the site shows each pollster's latest poll within the 21-day window minus the average, after that step. Leans in Poland are larger than in most countries on the site: several pollsters are three or four points from the average on KO or PiS. A lean is not an error: nobody knows yet who is right. Each pollster's page shows its lean now beside its errors at past elections.
What the rating leaves out
As elsewhere, the rating leaves out how often a pollster polls, who commissions it and mode. It also leaves out presidential and European elections: pollsters are rated only on the Sejm.
Limitations
The scale. The British end points give Polish pollsters few accuracy points. A Polish scale would spread them out; it would not make them more accurate.
Most pollsters have one or two elections. Nine of the pollsters in the table have one final poll, and their grades are provisional.
Openness from the pages we could read. OGB's reports are for subscribers; several pollsters publish only through clients, and a client's article may leave out what the pollster sent.
Removing the don't-knows. Every pollster is scored on its figures after the don't-knows are taken out. A pollster whose undecided split unevenly would be judged on a figure it did not publish.
The first real test of whether any rating predicts accuracy in Poland will be the next Sejm election.
References
Efron, B. and Morris, C. (1975). Data analysis using Stein's estimator and its generalizations. Journal of the American Statistical Association, 70(350), 311–319.
How to cite
Polling Forecast (2026). Rating Polish pollsters: accuracy, the lead and openness. Retrieved 21 November 2026 from https://www.pollingforecast.com/research/poland-pollsters. Live figures change as new polls arrive; the date above is when the text was last revised.


