Results form surveys are often taken at face value. Show, by discussing the construction and administration of postal questionnaires, where errors could occur in the data collection process.
Results form surveys are often taken at face value. Show, by discussing the construction and administration of postal questionnaires, where errors could occur in the data collection process.
A survey is a way of systematically gathering information using structured interviews or
postal/ telephone questionnaires, it also involves the interpretation of this data by
enumeration and statistical analysis. Results form surveys should not be taken at face
value as they often are. The data collection process needs to be looked at in greater detail
and understood before the results can be correctly interpreted.
This essay will explain how and where errors can occur in the data collection process
by discussing the administration and construction of postal questionnaires. Postal
questionnaires rely on self-completion and are widely used to gather information from a
large sample of the population. One of the main problems with postal questionnaires is
that they are self-completion. People who are illiterate or have disabilities such as
dyslexia may have problems completing questionnaires. (Simmons 2001.) Due to
certain groups of people being unable to respond, this may affect the results of the
survey. Therefore postal questionnaires should, where ever possible avoid being sent to
these groups of people although this may result in a unrepresentative sample. "Some
categories of people will be systematically under represented." (de Vaus 1996.)
Another major problem of postal questionnaires is the low response rates. Some postal
questionnaires do not achieve more than a 20% rate of return. (Simmons, 2001).
However a higher response rate can be obtained by methods such as sending personalized
letters written on official headed paper, sending written reminders, including a stamped
addressed envelope, subtly suggesting a deadline and by offering incentives such as free
gifts or entry in a prize draw. (Burns, 2000). The response rate also depends on the
sample of the population used, the construction and layout of the questionnaire, and the
topic of the questionnaire. The length of the questionnaire is also important as Schwab
(1999) stated "Response rates inevitably decline as questionnaire length increases."
Although we know reasons for low response rates, the investigator is unable to discover
the reason for non responders in his survey. "We have to assume that a non-response is a
refusal unless informed otherwise." (de Vaus, 1996: 235).
Even when a high response rate is achieved problems can still occur with the returned
questionnaires, for example they could be illegible or incomplete. This causes problems
when transferring the data onto computer. (Simmons, 2001).
The validity ...
This is a preview of the whole essay
(1999) stated "Response rates inevitably decline as questionnaire length increases."
Although we know reasons for low response rates, the investigator is unable to discover
the reason for non responders in his survey. "We have to assume that a non-response is a
refusal unless informed otherwise." (de Vaus, 1996: 235).
Even when a high response rate is achieved problems can still occur with the returned
questionnaires, for example they could be illegible or incomplete. This causes problems
when transferring the data onto computer. (Simmons, 2001).
The validity of the responses is also a problem with postal questionnaires. Studies have
shown that if the subject area is sensitive such as sexual behavior, emotional illness or
embarrassing situations, then people are less likely to be as honest in questionnaires than
in other forms of data collection. (Freeman & Sherwood 1970.). Therefore we need to
ensure responders that the questionnaires are confidential, and where possible
anonymous. However anonymity is usually impractical because follow-up reminders are
normally sent to people who have not responded. (Simmons, 2001).
The specific wording and arrangement of items in a questionnaire has a direct influence
on the response rate. (Punch 2000.) When writing the questions certain factors must be
taken into consideration. The respondents should be kept in mind, and the writer careful
not to over estimate their interest in the subject or their knowledge. The questions should
be simple, specific and honest. Care must be taken to make sure the questions do not
lead the respondent to answer a particular question in a particular way. (Coolican 1999.)
The investigators should avoid using certain words in a questionnaire, these include
absolutes e.g. everyone, all, always, never. Answers using these words are usually not
true. The word 'And' should be avoided as this usually means that the question is double
barreled or involves two ideas. The word 'You' should also be avoided where possible
because it is unclear whether the question is asking the individual, or the group the
responders is representing. Adjectives to describe frequency e.g. occasionally,
sometimes and often should be avoided too, as they can have different meanings to
different people. (Coolican 1999.)
The sequence of the questions in a survey should depend on the subject of
questionnaire. The beginning of a questionnaire should contain items which the
respondent may find interesting. This reduces the respondents concern that they may not
be able to answer the questions. Personal information should be asked for at the end of
the questionnaire. This is because it increases the chance of it being provided as "the
respondent has already made a commitment by completing the earlier part of the
questionnaire." (Coolican 1999.)
The construction of postal questionnaires is vital. "A questionnaire should be designed
with the respondent in mind." (Simmon, 2001: 87). When constructing a questionnaire
you should begin with "relatively easy and non threatening questions that respondents
will be happiest to answer." (Fulcher & Scott 1999). Questions in a survey should be
arranged chronologically and in an orderly framework. This makes it easier for
responders to remember and less likely for them to forget events. (de Vaus, 1997). "The
questions are highly structured in order to make completion and comparative analysis
easier." (Payne, 1951). Coolican (1999) described the features of a good questionnaire
as being "discriminate, reliable, valid and standardized."
A questionnaire should be constructed in four parts which include an introduction,
demographic questions, warm up questions and the body of the study. When planning a
questionnaire several factors must be taken into consideration. The physical appearance
of questionnaires is important, they need to be appealing to the eye, easy to complete,
printed in bold, clear lettering and have instructions provided with them. (Burns, 2000).
The layout of the questionnaire is important. Similar or relating questions should be
placed together to avoid confusion. The transition between topics should be done as
smoothly as possible. Important questions should not be placed at the end of a
questionnaire. The questions should be numbered down the page to assist the
investigator and the responder. Printing on both sides of a sheet of paper should be
avoided. (Fulcher & Scott 1999.)
The way in which the questions are written are also important in a postal questionnaire.
The questions should be short, clear and unambiguous. (Fulcher & Scott, 1999). Closed-
ended questions are more reliable as answers have to fit a certain category. There is no
room for the responder to comment on their answer or to explain it further. Closed-ended
questions cause a problem if the responder does not fit into any of the categories as they
will either fall to answer, or answer inappropriately. Open-ended questions are less
restricted and allow the user to answer more freely. However they do cause problems
with categorization. (Burns 2000). "Open questions cause problems with categorization
and lead to false conclusions." (Payne 1951). Open-ended questions cause the
respondents less frustration as they don't have the restraint of fixed choice answers.
(Coolican, !999.)
Before a questionnaire can actually be used in a survey it must go through a long and
complicated process of drafting and evaluation. (Fulcher & Scott, 1999). Pilot studies
are carried out in order to "remove ambiguity, test adequacy of response categories, and
all the work involved in administration." (Burns, 2000:568). Pilot testing involves
individuals giving their opinion of the questionnaire, and pointing out any items which
are confusing or misleading. This then usually results in changes to the final
questionnaire. These changes reduce data which may be missing, increase response rates
and obtain more valid answers. The pilot questionnaires should be carried on persons
similar to those who will take part in the final survey. (Coolican 1999.)
Sampling must be considered in the administration of postal questionnaires. A decision
must be made about which groups of people to target and how to get an unbiased sample.
In order to get a representative sample a stratified or random sampling technique should
be used. (Fulcher & Scott 1999.)
The use of postal questionnaires does have its problems, mainly being the low response
rates and the misunderstanding by respondents. Therefore results from surveys should
not be taken at face value. However these can be overcome by careful planning,
construction, and the design of the questionnaire.
Despite all these problems, postal questionnaires do have their advantages. They are
less expensive than face to face interviews. Respondents answers are more unbiased as
an interviewer is not present to influence them. There is no errors in the recording of the
responses. The respondent can answer in there own free time and take as long as they
like over it. More truthful answers are obtained as the respondents do not have to face
the fear and embarrassment of direct contact with an interviewer. Postal questionnaires
can use a larger sample size, and cover participants in areas which are hard to access. A
major advantage of postal questionnaires is that they can be completely confidential,
which leads to more truthful responses. (Burns 2000.)
BIBLIOGRAPHY
Burns, R. (2000) An Introduction To Research Methods. London: Sage.
Coolican, H. (1999) Research Methods And Statistics In Psychology. London: Hodder and Stoughton.
De Vaus, D. A. (1996) Surveys In Social Research. Australia: Allen and Unwin Ltd.
Fulcher. and Scott. (1999) Sociology. Oxford: Oxford University Press.
Freeman, H and Sherwood, C. (1970) Social Research and Social Policy. London: Prentice-Hall, Inc.
Gilbert, N. (2001) Researching Social Life. London: Sage.
Punch, K. F. (2000) Introduction To Social Research. London: Sage.
Schwab, D. (1999) Research Methods For Organizational Studies. London: LEA.
Payne. (1951) www.socresonline.org.uk/2/3/annexes/
MARK: 60%
