Polls
and Democracy
In any country with full democracy the governments and
political spend huge amounts of money on opinion polls. The surveys are a
vehicle for express opinions, ideas and concerns of population for a specific
topic. Recognizing that the survey results are of vital importance to
politicians. It is imperative to analyze the form or methodology used in the
survey. Because depending on the method and responsibility of the surveyor we
can get real data. So how we can do a survey technically valid? To accomplish
this goal we have roadmap that allows properly the elaboration of a survey: Set
the Sample Size. Sample Quality. Fieldwork / Classification. Laboratory.
Publication and Conclusion. It is important to follow these steps in order to
work successfully and to obtain reliable results that allow making wise
decisions.
Set the Sample Size: At this point we should work with
approximately 10% of the electorate. Working with a larger sample could be
recommended but is expensive. Anyway, there are statistical methods for work
efficiently with a sample of this size. The collection of this information must
be made between citizens eligible to vote.
Sample Quality: When determining the sample size. Proceed
to determine its characteristics and quality. This part is chosen at random to
survey participants. Shouldn’t predetermine any of the respondents. This
process should be transparent. Because otherwise you lose objectivity. For
example: Imagine a sample where 80% of respondents are chosen in advance, with
the same political tendency. This result is obviously altered and never be
objective. The interviewer should work with random samples to make sure that
his work is based on a representative sample.
Fieldwork / Classification: At this point it specifies
the ways of collect and processing data. If decide to place on the street
interviewers should establish schedules and sites for example: In the subway,
shopping’s and parks during rush hour. And if the survey is done by telephone
also set times and dates. Once the information in our bureau we proceed to
organize the data by age, sex, ethnic group, education etc.
Laboratory:
This is where we started to apply statistical models that give reliable results
(rate of up to 97% success rate). To run the models we have to do some
preliminary calculations such as: Sample mean, mode, median and standard
deviation. Once done proceed to make our table data and can create some charts.
All these models are based on sophisticated mathematical calculations that allow
us to reach concrete results.
Publication: Here we discuss the results with the
politician or his advisers. And if the
survey is not for private use can be published in the media. Usually at this
point are given recommendations or suggestions to the political. Depending on
how this popularity. And finally we can set when is the next measurement and if
he wants to add more details to study.
Conclusions: For the statistics is very important to work
with professional thoroughness. Because events can occur as seen in the years
1936 and 1948. Specifically in 1936 the pollsters gave the winner the governor
Alf Landon (with a ratio 4 to 1) Vs. President Roosevelt. And the end result
was totally different. President Roosevelt won with over 60% of the vote.
In 1948, three well-known pollsters: Roper, Crossley and
Gallup conducted the same wrong prediction: predicted that Thomas Dewey would
defeat Harry Truman and become president of the United States. This error would
not have been so shocking if one month before the election, pollsters had not
decided not to conduct more surveys because they were convinced that Thomas Dewey
would win the election. History reminds us that Harry Truman ended up winning
the presidential elections of 1948. Not surprisingly, Truman enjoys to
criticize the newspapers, so sure of their defeat, they printed the wrong
headline Dewey defeats Truman.
The
surveyors of the time they were astonished. And it took years for people to
trust the surveys. The government even threatened to regulate the sector
surveys. The surveyors took careful note of what not to do, following the
colossal mistake. And since 1948, most of the best known pollsters have
successfully predicted the outcome of U.S. presidential elections.
