Opis
A Gallup Poll as the pioneer of modern scientific polling in the mid-20th century, and its modern role as a gauge of broad public opinion and social trends rather than election forecasting.
George Gallup’s defense of polling methodology in the 1970s was built on decades of refining random sampling techniques.
Prior to 1936, the popular Literary Digest poll was considered the gold standard. It sent out millions of postcard surveys via telephone directories and car registration lists, predicting a blowout victory for Republican Alf Landon over Franklin D. Roosevelt. George Gallup used a much smaller, scientifically selected quota sample (~50,000 people). He correctly predicted FDR’s landslide, proving that sample representativeness matters far more than sample size.
Gallup and other major pollsters famously predicted Thomas Dewey would comfortably defeat Harry S. Truman ("Dewey Defeats Truman"). Gallup stopped interviewing voters weeks before Election Day, missing a massive late surge toward Truman. Furthermore, quota sampling allowed interviewers discretion in choosing respondents, introducing systematic bias.
In response to 1948, Gallup abandoned quota sampling in favor of probability-based random sampling (and eventually Random Digit Dialing via phone) and polled right up until Election Day. Between 1952 and 2008, Gallup maintained an impressive track record in U.S. presidential elections, generally keeping its final popular vote margin error within 2 to 3 percentage points.
By the late 2000s, landline usage plummeted, caller ID became ubiquitous, and survey response rates dropped from over 80% in the mid-20th century to single digits. Reaching a representative sample of voters required far more complex weighting and multi-mode sampling methods.
In 2012, Gallup's final pre-election poll showed Mitt Romney leading Barack Obama slightly in the national popular vote, whereas Obama won the popular vote by nearly 4 percentage points.
Following the 2012 election, Gallup made a strategic decision ahead of the 2016 campaign: it stopped conducting trial-heat ("horse-race") polling for U.S. presidential elections.
Today, Gallup does not attempt to predict who will win an election. Instead, it measures broader attitudes— presidential job approval, trust in institutions, economic confidence, and social policy views. Gallup primarily uses probability-based online research panels (The Gallup Panel) combined with mail and telephone sampling to combat low response rates and mode effects. Gallup remains highly accurate at capturing overall social and political sentiments over time. Because it focuses on long-term trends using consistent question wording, its data provides an authoritative benchmark for tracking how public sentiment shifts across decades.