Polling institutes in Portugal at a glance

The PolitPro Score evaluates the reliability of polling institutes through an objective comparison of their data with actual election results and the calculated polling trend. A high score represents a precise reflection of the political climate without systematic bias. If an institute’s results for specific parties consistently deviate significantly from the overall trend, points are deducted. Such differences may indicate so-called "house effects"—methodological peculiarities in weighting that systematically over- or under-represent certain parties. The score thus provides transparent guidance on which polls mirror reality most accurately. The maximum score is 100.

Election accuracy shows how close pre-election polls were to the actual election result. We compare an institute’s pre-election polls with the final election outcome. For each election, we calculate the average deviation of party results and then compute an overall average. Only parties with an election result of 3% or higher are included. Opinion polls generally have a margin of error of 2–3 percentage points. Average deviations of up to 1 percentage point are rated as very good (green), up to 2 percentage points as good (yellow). Deviations of up to 3 percentage points are rated as ok (orange) but already show clear outliers for some parties. Average deviations above 3 percentage points indicate weaknesses in accuracy and reliability and are rated as poor (red).

Party deviation measures how often an institute’s results differ significantly from the election trend. Frequent deviations for individual parties may indicate systematic bias. We calculate an institute’s deviation based on party results in election polls by comparing each party’s value with the PolitPro election trend, which represents a weighted average of current opinion polls. If a party’s value is within ±1 percentage point of the trend, it is rated as equal. If it is lower, it is rated as lower; if higher, as higher. Across all polls, we then calculate how often an institute rates a party higher or lower than the trend. Deviations of up to 20% are considered normal. If deviations exceed 20%, we highlight the party accordingly. Note: We recommend always interpreting party deviations together with election accuracy.