Opinion poll

An opinion poll is a statistical measurement taken from surveys designed to find out public opinion. These measurements are made through sampling, which is usually designed to represent the opinions of a population by asking a series of questions and then extrapolating generalities in proportion or within a confidence interval. The science that studies the data obtained through opinion polls is called demoscopy, and is part of sociology.

History of opinion polls

The earliest known example of an opinion poll was an unscientific local vote poll conducted by the Harrisburg Pennsylvanian in 1824, which gave Andrew Jackson the win over John Quincy Adams by 335 votes to 169 in the campaign for President. from United States. Such surveys, without value or scientific method, gradually became popular, but kept to a local scope, usually in a city area. In 1916, the Literary Digest embarked on a nationwide survey, sending out millions of postcards and simply counting those that responded. In this way, Woodrow Wilson's election as President and the next four presidential elections were correctly predicted.

In 1936, the "Literary Digest" used a significant sample of 2.3 million "voters", with which they had determined that the American population tended to sympathize with the Republican Party. A week before presidential election day, Alf Landon of the Republican Party was reported to be far more popular than Franklin D. Roosevelt of the Democratic Party. At the same time, George Gallup conducted a much smaller but more scientifically based survey using representative demographic samples. Gallup correctly predicted Roosevelt's landslide victory. Before long, the Literary Digest stopped working, while the polls began to increase.

Elmo Roper was another American pioneer in political prediction using scientific polling. Thus, he predicted the election of President Franklin Delano Roosevelt three times, in 1936, 1940, and 1944. Louis Harris had been in the field of public opinion since 1947 when he joined Elmo Roper's firm, later becoming his partner.

Gallup launched a subsidiary in the United Kingdom, where he correctly predicted the victory of the Labor Party in the 1945 general election, in contrast to virtually all other commentators, who expected the Conservative Party, led by Winston Churchill, to win easily.

By the 1950s, pollsters had spread throughout most democracies, reaching practically every country in the world at the beginning of the 21st century.

Currently, there is an increasing trend of conducting surveys offering respondents a bonus (money or bonuses). In some cases, there are companies that use web resources, such as pages and portals, to search for and reach a much larger audience., obtaining a wide variety of results on different types of topics: products, services, brands, experiences, politicians, etc., offering volunteers a payment for each survey they solve. However, the user of this type of portal must be aware of the potential danger you can run by using random sites, without first doing a thorough check to be sure that such can actually give you your bonus once the survey is over. Now only lives in my memory

Sample and survey methods

For many years, opinion polls were conducted through telecommunications or, mainly, face-to-face, both on the streets and in people's homes. Methods and techniques varied but are widely accepted in most places. Verbal, paper ballot and efficiently processed surveys can be conducted to contrast other types of survey with systematic and more complicated matrices compared to previous orthodox procedures.

Opinion polls thrived on popular apps, though response rates for some surveys declined. In some countries it has become common to carry out opinion polls through telephone calls, I feel these are easier, cheaper and faster than the face-to-face method. In addition to this, Internet surveys are becoming more popular, despite the difficulties of taking scientific samples largely over the Internet from certain important sectors in some countries. Pollsters using Internet surveys must draw the sample from a large panel of volunteers, and the results are weighted to reflect the demographics of the population of interest. This makes a difference in the popular surveys to which anyone who wants to participate submits,

The wording of a poll may include bias, since the bias may be in the opinion. For example, the public is more likely to indicate support for a person who is described by the operator as one of the "leading candidates." This question itself hides a subtle bias for one candidate, by lumping some candidates into an "other" category or vice versa. The tools used in 21st century surveys vary in complexity due to these circumstances.

Error potential

Sampling error

All surveys are based on samples that are subject to sampling error, which reflects the accuracy uncertainty of the sampling process. Uncertainty is frequently expressed as a margin of error. The margin of error does not reflect sources of error other than sizing error. A poll with a random sample of 1,000 people will have a margin of error of 20%, an estimated percentage of the population. A +/- 2.5% margin of error means that 95% of the time the procedure used will give an estimate of +/- 2.5% of the percentage to be estimated. The margin of error can be reduced by a larger sample, so if a pollster wants to reduce the margin of error to 1%, he would need a sample of about 10,000 people. In practice, pollsters need to balance the cost of a large sample with the reduction of sampling error, with a sample size of 500 to 1,000 people being a typical number needed for political polls. (Note that 500 complete responses could take thousands of phone calls.)

Unanswered

As some people may refuse to answer calls from strangers, or refuse to answer the poll, the samples may not be representative of the population. Because of this, the characteristics of those who agree to be surveyed may be markedly different from those who refuse, making the sample a biased version of the universe the surveyor wants to analyze. In these cases, the trend introduces new errors in the survey, in one way or another, in addition to the errors caused by a simple sample. The error due to sample trend does not become smaller with larger population samples. If the people who refuse to answer have the same characteristics as those who do agree to answer the surveys, then the final results will be unbiased, whereas if people who don't respond have different opinions than those who do, then there will be a bias error in the results. In terms of polling for election, studies suggest that the effects of bias are small, but each polling company has its own formulas on how to minimize bias bias in the sample.

In addition to this, there are also other factors for which there may be a greater margin of error, which may be the lack of stratification of the sample, poor analysis or poor supervision; including lack of feedback from the field team.

Biased response

Measurement results could be affected by biased responses, where the responses given by respondents do not reflect their true beliefs. This could be deliberately engineered by unscrupulous pollsters in a push poll, but is more often the result of an ordering of questions (see below). Respondents may try to manipulate the success of a poll by, for example, taking a more extreme position than they currently hold in order to explain their side of the argument quickly and thus finish without wasting time. In addition, the respondents may find themselves under social pressure at the time of taking part in the survey and in this way they may try to give an answer that is more accepted than the one that corresponds to their true opinion.

Writing questions

It is well known that, in question writing, the order in which questions are asked and the pattern in which answers are listed can influence the results of a poll. One way in which pollsters try to minimize this effect is to ask the same set of questions to elicit changes in opinion. The most effective controls used by attitude researchers are:

• Ask enough questions to be able to cover all aspects of an issue and control for the effects due to the pattern of the questions (such as positive or negative wording), the adequacy of number is established quantitatively with psychometric measures such as reliability coefficients.
• analyze the results using psychometric techniques to synthesize the answers in a small but reliable quantity and thus detect non-effective questions.

These controls are not widely used in the polling industry.

Errors caused by the groups that are the object of study

Another factor that induces error is surveying social groups that are not representative of the population as a consequence of the method used. For example, telephone surveys have an inherent margin of error since in many times and places, those who had telephones were richer than those who did not. Also, today in many places the population only has mobile phones. It may also happen that interviewers are prohibited from calling cell phones (because it is illegal in that jurisdiction to make unsolicited calls to phones where the owner pays for the simple fact of taking the call), and therefore these individuals are never included in the survey. the polls.

If the subset of the population without mobile phones differs significantly from the rest of the population, these differences may alter the results of the survey. Telephone survey companies have developed many weighting-based techniques to help mitigate these shortcomings, with more or less success.

A widely used example to illustrate the unreliability of telephone polls is that of the UK general election held in 1992. Despite the fact that the polling companies used different methods, practically all of them, both in the days before the elections and in exit polls gave the Labor Party in opposition as the winner. However, the actual vote ended up giving victory to the Conservative Party, then in power. In the aftermath of this failure, the polling companies came up with various ideas to try to justify such a mistake. Among others, they included:

• A last minute change of heart. The Conservatives' victory was made possible by the votes of people who changed their minds at the last minute. Therefore, the error is not as serious as it may seem.
• Low voter turnout. Those who voted for the Conservative party were not as willing to participate in the poll as they had been on other occasions, and that is why this time their opinion was not reflected in the poll.
• The hidden vote. The Conservatives have suffered a continuous period of low popularity due to the economic recession and some major scandals. Some of your voters may have felt pressured to give an answer more in line with the majority's thinking, without that being their intention to vote.

The influence of these factors on the final result was and still is a controversial issue, but since then polling companies have improved their methods and have managed to better predict the results of the following elections.

All the major television networks in virtually every country where elections are held conduct polls (either on their own or in collaboration with the major magazines or newspapers). The best-known polling failure in recent memory in the United States is the prediction that Thomas E. Dewey would defeat Harry S. Truman in the 1948 presidential election. Major polling companies, including Gallup and Roper, predicted a landslide victory for Dewey.

In the UK, most polls failed to predict Conservative electoral victories in 1970 and 1992, and Labor victory in 1974. However, their results for other elections have been fairly accurate.

The influence of opinion polls

By providing information on the voting intentions of the population, opinion polls can influence the behavior of voters. The various theories that exist on the subject can be divided into two large groups: the pull and rejection effects and the strategic or tactical vote (Emilio Lamo de Espinosa, 2001). The bandwagon effect occurs when polling causes voters to support the candidate who is predicted to win. The idea that voters are susceptible to these effects is not new, dating back to at least 1884; William Safire (1993:43) reported that it was first used that same year in the political comics magazine Puck. This idea persists despite the lack of empirical evidence until the end of the 20th century. George Gallup spent a lot of effort in vain trying to discredit this theory in his day by presenting empirical studies. A recent meta-study of scientific research in this field indicates that from the 1980s onwards the carryover effect is much more frequent (Irwin and van Holsteyn, 2000).

The second category within the theories of how the polls directly affect the vote is the strategic vote or tactical vote. This theory is based on the idea that voters see elections as a method of choosing a government. Thus, some do not choose their candidate for reasons of political affinity or sympathy, but for other reasons that are not so desirable, such as strategic reasons. An example of this was seen in the 1997 UK general election, when Michael Portillo's victory in the Enfield constituency was taken for granted yet opinion polls showed Labor Party candidate Stephen Twigg steadily gaining support. That may have prompted swing voters or voters from other parties to support Twigg to get Portillo out. Another example is the boomerang effect in which the candidate's supposed voters take his victory for granted and feel that their vote is not necessary, thus allowing the victory to the opponent. These effects only indicate how opinion polls directly affect the decision of the electorate. Journalists, politicians, political parties and officials, among others, are also affected, but above all the approach of the media and political ideologies.

Polling companies

In Australia the best known are:

• Newspoll
• Roy Morgan Research

In Argentina one of the best known is:

• MKT SA located in the City of Córdoba

In Spain

• Metroscopia
• Sigma Dos
• TNS Demoscopia
• Comment
• Accra Delta
• Netquest
• ESEQ, Social, political and market studies

In the United States, some of the most important are:

• Gallup poll created by George Gallup.
• Quinnipiac Polls, run by Quinnipiac University in Hamden, Connecticut, started as a student project.
• The Pew Charitable Trusts conduct surveys focused on media and political opinion.
• Harris Poll.
• Nielsen Ratings, which almost always works for television.
• Zogby International has been tracking public opinion since 1984.

In Mexico, some of the most important firms are:

• ARCOP, which was directed by Rafael Giménez.
• ASESORIA ECONOMICA Y MARKETING SC, Dirigida por la MC Erika Donjuan. www.corporacionaem.com
• BGC, headed by Ulises Beltrán.
• Consultation, directed by Roy Campos.
• Covarrubias and Associates, by Ana Cristina Covarrubias.
• Demotechnics, by María de las Heras.
• GEA-ISA. Ricardo de la Peña chairs Applied Social Research (ISA).
• IPSOS-BIMSA. Jorge Buendía heads the area of ​​public opinion.
• Mercaei, headed by Lauro Mercado.
• WORLD, directed by Daniel Lund.
• Parametry, directed by Francisco Abundis.
• PlanE Strategy and Communication SC, directed by Fernando Mejorado Olagues

In Panama, among the most important are:

• Poet & Neira

In the UK, the most prominent are:

• MORI.
• YouGov, an online polling company.
• NOP
• ICR
• Populus, the official pollster for The Times newspaper.