First 783 words of the document:
What is a social survey?
Survey data The national Readership Survey tells us that in 2000, The Sun was the most popular daily
newspaper in Britain read by 20% of adults. The international Passenger Survey tells us that Spain
was the most popular overseas holiday destination in 2000 visited by 28% of UK residents who had a
holiday abroad. And the British Gambling Prevalence Survey informs us that the National Lottery Draw
was the most popular gambling activity in Britain in 1999, with 65% of people aged 16 and over
Definition The above information comes from social surveys. A social survey involves the systematic
collection of the same type of data from a fairly large number of people. Social surveys are usually
designed to gather information on the same variables eg, age and cinema attendance from those
participating in the survey. This often means asking everybody the same set of questions.
Nearly all social surveys are based on a sample of the population to be investigated. `Population' is the
term given to everybody in the group to be studied. The population might be adult males, female
pensioners, manual workers, 1619 year old students, parents with dependent children and so on. A
sample is a selection of part of the population. Samples are necessary because researchers rarely have
the time and money to study everybody in the population. For example, if their research was based on
women aged 16 and over in the UK, it would cover over 23 million people.
Most researchers try to select a sample which is representative of the population. This means that the
sample should have the same characteristics as the population as a whole. Thus, if a researcher is
studying the attitudes of British women, the sample should not consist of 1000 nuns, 1000 women over
80 or 1000 divorced women since such groups are hardly representative of British women. With a
representative sample, generalisations are more likely to be true findings from the sample are more
likely to be applicable to the population as a whole.
Sample design and composition
Sampling unit Who should be included in a sample? In many cases it is fairly easy to define a sampling
unit ie, a member of the population to be studied. Dentists, males between 30 and 40 years of age,
females who own their own businesses, people with one or more GCE A levels, can be defined without
too many problems. However, other groups are not so easy how would you define a semiskilled
manual worker or a person living in poverty? Who would you include in a population of `criminals'? Do
you limit the population to those convicted of a crime? Or do you include everybody who has ever
broken the law, in which case you would include nearly every adult in the UK?
Sampling Frame Once the research population has been defined, the sample is selected from a sampling
frame a list of members of the population to be studied. In some cases an appropriate sampling frame is
readily available, eg the Electoral Register for a study of voting behaviour. In other cases researchers
may have to rely on listings, such as the Postcode Address File or telephone directories, which may or
may not be suitable for their purposes. And all listings have drawbacks not everyone is included, they
are often out of date, certain groups are likely to be over or underrepresented, eg the poor are less likely
to appear in telephone directories. Sometimes, those who have data needed for a sampling frame are
unwilling to release it. This happened to Howard Newby (1977) when the Ministry of Agriculture
refused to supply information for his study of Suffolk farm workers. Newby had to use the Yellow Pages
for his first sampling frame. Many farm workers were absent from this directory and those included were
probably unrepresentative of the group.
Types of Sample
Random Samples A random sample gives every member of the sampling frame an equal chance of being
selected. Every name is given a number and then a list of random numbers is used to select the sample.
This avoids bias in selection. If researchers choose who to include and who to leave out, they may
select a sample which supports their hypothesis.
Systematic samples This form of sampling is systematically selects people from the sampling frame by
choosing every 5th, 19th, 20th, or whatever, sampling unit. This method was used by Young and Willmott
Other pages in this set
Here's a taster:
Bethnal Green. They selected every 10th name from the borough's electoral
Neither random nor systematic samples necessarily produce representative samples. Few sampling
frames cover everybody in the research population. Eg on electoral registers certain groups are
unrepresented (those not old enough to vote) or underrepresented (the unemployed).
Even if the sampling frame covers the entire research population, a representative sample is not
Here's a taster:
Responding to surveys
Response rates its one thing creating a representative sample, its quite another getting everybody in
the sample to participate in the survey. The response rate the percentage of the sample that
participates varies widely. Eg, Shere Hite's The Hite Report on the Family (1994) based on
questionnaires in magazines had a mere 3% response rate, whereas everybody Ann Oakley (1974) asked
to take part in her research on housework agreed to do so.…read more