Politics and public information · Brazilian Electoral Court (TSE) updated PesqEle on September 25, 2026; official source checked October 19.
How to check an election poll before sharing it
This guide explains how to inspect a poll’s documentation. It does not report candidate percentages, predict a winner or treat online searches as voting intention.
Editorial prompt: the Google Trends Brazil snapshot on October 11, 2026 at 17:27 (São Paulo time) included the search group “pesquisa segundo turno”. This indicates interest in information; it is not evidence of voter preference or a ranking of all Google searches.
What the official consultation can show
Brazil’s Superior Electoral Court (TSE) reported on September 25, 2026 that it had updated PesqEle, the Electoral Research Registration System. The change standardises and expands the detail shown about methodology and fieldwork. The TSE said the update did not create new legal duties; it improves how information already required under Resolution 23,600/2019 is presented. The system can show the survey’s methodology, field period, sample size, margin of error, responsible statistician and questionnaire.
The TSE said institutes should provide clearer information on the population represented, sampling plan, projected sample, confidence level and maximum expected margin of error, along with field procedures and question order. After the applicable disclosure period, they must return to PesqEle with details actually observed in fieldwork, including collection dates and times, achieved sample, handling of refusals or replacements, weighting, response rate and methodological limitations. These are details of the Brazilian system; they do not certify that a poll predicts the election result.
A checklist for reading a poll
Start with the original document published by the polling institute. Record who is responsible, which area and population it covers, and when interviews took place. Then find the corresponding PesqEle entry and compare it with the copy, chart or screenshot you received. A cropped image may omit the place or date that changes what a number means.
Read the exact question and how it was asked. An unprompted answer and an answer selected from a list are different measures. Check whether the chart covers everyone interviewed or a subgroup, and identify the base for each percentage. If the document does not make that clear, pause the comparison instead of filling the gap with an assumption.
How to avoid a misleading comparison
Consider a fictional teaching example, unrelated to real candidates: one card shows 48% and another 50%. Before calling that an increase, check whether both use the same question, region, population, method and field period. A difference between results from separate surveys does not by itself establish a trend.
Read the margin of error together with the confidence level and conditions stated by the poll. It is not permission to declare a winner or to run a statistical test on a social-media crop. If the documentation is incomplete, the responsible conclusion is to say what is missing. This article does not use or infer any candidate’s polling percentage.
Search interest and votes measure different things
The phrase “pesquisa segundo turno” appeared in the Trends snapshot that prompted this guide. Search interest can reflect curiosity, a fact-checking need or research for many other reasons; it does not reveal how someone plans to vote. Google searches are not a representative sample of voters and cannot substitute for an election poll.
For that reason, this article gives no candidate percentages, victory projections or election results. Its purpose is a verification method that remains useful when another poll circulates. The observed Trends group is recorded as editorial provenance, not as proof of a polling change or a cause of the search increase.
Sources and verification
The TSE notice was published on September 25, 2026. The official source was checked on October 19, 2026; the search-trend observation is dated October 11, 2026.
ConverteArq commentary
We think a blog’s most useful contribution to this debate is to make the method visible. If readers see only the larger number in a graphic, they may compare answers collected from different populations or questions. Showing where to find the full document helps more than adding an adjective to a result, and lets readers with different political views inspect the same evidence.
A search trend can guide the question a newsroom answers, but it cannot determine the conclusion. Checking the method before the number reduces the room for premature headlines. Keeping the date and source also makes it easier to correct an article if information changes, without disguising its history. That care matters during an election, when a clipped graphic can keep circulating after it has lost its context.