← Projects
Do polls resemble the outlet that commissions them?
We read polls every week and often suspect they are biased, but we rarely stop to compare them. Here they are side by side, every polling firm's latest poll since Spain's 29 November general election was called, grouped by the political leaning of the outlet that commissions it, to see how far a poll is also a political tool and not just an estimate. Every figure links to its source: the published poll, its methodology and the study used to classify each outlet's leaning.
Outlets left of centre Centre or no data Outlets right of centre median of the others −25 −20 −15 −10 −5 +5 +10 La Séptima · SW Demoscopia: −24.5 seats for PP and Vox vs. the median El Periódico · GESOP: −8 seats for PP and Vox vs. the median elDiario.es · Ateneo del Dato: −4 seats for PP and Vox vs. the median Artículo14 Artículo14 · Winston: −10 seats for PP and Vox vs. the median El País · 40dB: 0 seats for PP and Vox vs. the median Electomanía Electomanía · EM-Analytics: −4 seats for PP and Vox vs. the median El Confidencial El Confidencial · IMOP: −6 seats for PP and Vox vs. the median ABC · GAD3: −3 seats for PP and Vox vs. the median 20minutos · DYM: +5 seats for PP and Vox vs. the median El Mundo · Sigma Dos: +3 seats for PP and Vox vs. the median El Español El Español · SocioMétrica: +4 seats for PP and Vox vs. the median Onda Cero · Celeste-Tel: +5 seats for PP and Vox vs. the median El Debate El Debate · Target Point: +4 seats for PP and Vox vs. the median Vozpópuli · Hamalgama Métrica: +8 seats for PP and Vox vs. the median La Razón · NC Report: +9 seats for PP and Vox vs. the median Each sign is the outlet that commissions the poll, on the lane of its leaning and placed by how many more or fewer seats it gives PP and Vox than the median of the others. Below, the same figure in detail.
Polls from right-of-centre outlets give PP and Vox a median of 4 more seats than the rest. Those from left-of-centre outlets , 6 fewer seats.
Deviation of each polling firm by the outlet that commissions it
Each dot is a polling firm: how many more (to the right) or fewer (to the left) seats it gives PP and Vox than the median of the others. The band joins the minimum and maximum of each group. Rows group the outlets by where the public places them on the CIS ideological scale. Click a dot to see its details.
Who reaches 176?
None of the 15 polling firms gives the government bloc a majority. 14 give one to PP and Vox, and 1 leaves it undecided.
14 polling firms give PP and Vox a majority
0 polling firms give the government bloc a majority
1 undecided : the range straddles 176 or no one gets there
Each bar is the Congress according to one polling firm. The parties that currently support the government push from the left, and PP and Vox from the right. The bloc that crosses the central line, half of the 350 seats, has an absolute majority. Rows go from most to fewest seats for PP and Vox, and the right-hand column shows how many more or fewer seats each firm gives them than the median of the others. Hover over a row to see its details and methodology, and click it to pin it in the panel.
×
Hover over a row to see here its seat split, how far it is from the others, the leaning of the outlet that commissions it and its methodology, with links to the sources. Click a row to pin it.
See the data and methodology of each poll
Seats for each bloc and methodology of each polling firm's latest poll
Outlet Polling firm Fieldwork Government bloc Others PP and Vox Vs. median Result Outlet's leaning Technique Sample Error Weighting Source
La Razón NC Report 26 Oct–29 Oct 139–148 0 206–210 +9 PP and Vox majority Right of centre Telephone and online 1,000 ±3.1% Not published Electográfica, citing the methodology published by La Razón Vozpópuli Hamalgama Métrica 22 Oct–28 Oct 143 0 207 +8 PP and Vox majority Right of centre Telephone (CATI), landlines and mobiles 1,000 ±3.16% (95.5%) Not published Vozpópuli Onda Cero Celeste-Tel 29 Oct–3 Nov 146 0 204 +5 PP and Vox majority Right of centre Telephone 1,100 ±3.0% Not published Electográfica 20minutos DYM 5 Oct–7 Oct n/a n/a 200–208 +5 PP and Vox majority Centre or no data Online 1,011 ±3.1% Not published Electográfica, citing the methodology published by 20minutos El Español SocioMétrica 1 Sep–1 Nov 147 0 203 +4 PP and Vox majority Right of centre Online (CAWI), panel 4,893 Not applicable: non-probability sample Sex, age, province and recalled vote El Español El Debate Target Point 28 Oct–30 Oct n/a n/a 199–207 +4 PP and Vox majority Right of centre Online (CAWI), panel 1,024 ±3.1% (95.5%) Sex, age, region, municipality size, education, employment status and recalled 2023 vote El Debate El Mundo Sigma Dos 29 Oct–30 Oct 140–157 0 196–208 +3 PP and Vox majority Right of centre Telephone and online 2,505 ±2.0% Not published Electográfica, citing the methodology published by El Mundo El País 40dB 7 Oct–8 Oct n/a n/a 186–214 0 PP and Vox majority Left of centre Online (CAWI) 800 ±3.5% (95%) Not published 40dB report for El País ABC GAD3 6 Oct–8 Oct 152 0 198 −3 PP and Vox majority Right of centre Telephone 1,077 ±3.0% Not published Electográfica, citing the methodology published by ABC elDiario.es Ateneo del Dato 20 Oct–22 Oct n/a n/a 183–211 −4 PP and Vox majority Left of centre Online (CAWI), panel 2,000 ±2.2% (95.5%) Not published elDiario.es Electomanía EM-Analytics 28 Oct–30 Oct 153 0 197 −4 PP and Vox majority Centre or no data Online (CAWI), 60% fixed panellists 1,527 ±2.6% (95.5%) Not published Electomanía El Confidencial IMOP 13 Oct–16 Oct 154 0 195 −6 PP and Vox majority Right of centre Telephone and online 1,000 ±3.1% Not published Electográfica, citing the methodology published by El Confidencial El Periódico GESOP 6 Oct–8 Oct n/a n/a 189–197 −8 PP and Vox majority Left of centre Telephone and online 1,000 ±3.1% Not published Electográfica, citing the methodology published by El Periódico Artículo14 Winston 21 Oct–23 Oct n/a n/a 191 −10 PP and Vox majority Centre or no data Not published 1,200 Not published Not published Artículo14 La Séptima SW Demoscopia ¹ 26 Oct–5 Nov 165–180 0 170–183 −24.5 Undecided Left of centre Telephone (CATI), landlines and mobiles 4,000 ±1.6% (95%) Sex, age, region, recalled 2023 vote and stated likelihood of voting SW Demoscopia report for La Séptima
What it shows. The 14 polling firms that give PP and Vox a majority put them between 183 and 214 seats. SW Demoscopia gives the government bloc the most seats, 165–180, short of 176, and is the only one that leaves PP and Vox undecided, with 170–183. Polls commissioned by right-of-centre outlets give PP and Vox a median of +4 seats compared with the other firms, and those commissioned by left-of-centre outlets, −6.
How it is counted. For each polling firm, its poll with the most recent fieldwork since 5 October 2026 is used. The government bloc is made up of the parties that currently support the government: PSOE, Frente Amplio and Front Valencià (formerly Sumar), Podemos, ERC, Junts, EH Bildu, PNV, BNG, CC and Adelante Andalucía. PP includes UPN. SALF does not count with PP and Vox. A bloc has a majority if it reaches 176 seats. When a firm gives a range, the minimum counts: if the minimum falls short of 176 but the maximum reaches it, the firm is “undecided”. The deviation is the difference between the seats a firm gives PP and Vox (the midpoint, if it gives a range) and the median of the other firms.
Methodology. Technique, sample, margin of error and weighting are those published by each polling firm or the outlet that commissions it. When something is not published, that is stated. With samples of about a thousand interviews, the margin of error is around three points per party, and online panels are not random samples, so their margin of error is only indicative.
Outlet's leaning. The coloured dot next to each outlet is where the public places it on the CIS scale, from 1 (left) to 10 (right), according to Carlos Gonzalo 's calculations with the microdata of CIS studies 3421 (2023) and 3511 (2025), and his criterion: below 5.5, left of centre. 20minutos (5.04) is grey because the author calls it centrist. Outlets that are not in the CIS data are classified by how they describe themselves or by a sourced description, and grey if there is no data or the classification is disputed. It is the leaning of the outlet that commissions the poll, not of the firm that conducts it.
What is left out. Polls commissioned by parties, those by the CIS (Spain's public Centre for Sociological Research), which publishes vote shares but not seats, and third-party re-estimates of CIS data. Nor are InvyMark, More in Common, Sondaxe, which do not publish seats. Some firms only publish seats for the large parties. In their bars, the hatched area is what they do not publish, not abstention or other parties.
¹ SW Demoscopia. Daily tracking poll for La Séptima. The company behind it, Global Europa Demoscopia, has a single shareholder who is also its director (Companies Register data ). The PSOE hired it for its campaign in Aragón in 2026 and it has worked for the party repeatedly (elDiario.es ). According to Libertad Digital , its owner headed the PSOE's electoral analysis unit until 2025.
Source. Poll table from Wikipedia , retrieved 7 November 2026, checked against the outlets that published each poll. Methodology from the publications linked in the table. Data derived from Wikipedia (CC BY-SA 4.0).
Logos. From Wikimedia Commons: La Razón (Public domain, by Audiovisual Española, S.A.), Vozpópuli (CC BY-SA 4.0, by vozpopuli.com), Onda Cero (Public domain, by CMetalCore), 20minutos (CC BY-SA 4.0, by HernandoJoseAJ), El Mundo (Public domain, by El Mundo), El País (Public domain, by Diario El País S.L.), ABC (Public domain, by Galifardeu), elDiario.es (CC BY-SA 3.0, by eldiario.es), El Periódico (Public domain, by El Periódico de Catalunya), La Séptima (Public domain, by BuenaNenita). The other outlets have no freely licensed logo on Commons and appear by name. Trademarks belong to their owners and are used only to identify each outlet.
How to cite. David Clothier, “Do polls resemble the outlet that commissions them?”, dataviz.cookingchart.org. The chart and data are published under CC BY-SA 4.0 .
More projects · Legal notice and privacy