Using a questionnaire Postgraduate study skills

Questionnaires provide quantitative and qualitative methods of data gathering - the evidence, data or information you find can be expressed in numerical or descriptive terms.

An alternative to a self-completion questionnaire is a structured interview, where the questionnaire is administered in person or over the telephone. The advantage of using a questionnaire rather than an interview is that you can reach large numbers of people more easily, as you can leave them to fill in the questionnaire and send it back to you.

Questionnaires are more rigid than interviews. Unless you leave a space for the interviewee to write their own answer (e.g. 'If you have any more comments please write in the space provided'), the respondent can only choose from the range of answers you have given. You therefore have to make sure that you have designed your questionnaire well.

When preparing a questionnaire you also need to keep in mind the following points.

• Make sure you introduce yourself and explain what the aim of the questionnaire is. Also, make sure the respondent is aware of the ethical implications of the research
• Devise your questions so they help to answer your research question, that way, all the questions will be relevant
• Try and have a sequence to your questions or topics - group them in themes and make sure they follow on logically from each other
• Make sure your questions are clear and easy to understand - only use technical or academic language if you are sure the respondent will understand what you mean
• Do not ask leading questions. Make sure people are free to give their own, honest answer
• Always run a pilot of your questionnaire. You can ask colleagues, fellow students or family members to fill it in. This helps you to test the technical aspect of the questionnaire, such as whether you have given the correct answer options or whether you have left anything out.

Sampling

When you design your research you need to take into account how many people you need to include to make the research valid.

If you are investigating a narrow but deep subject you may not need to question that many people; you may be interested in the opinions and experiences of experts or people with direct experience, rather than a random sample. If you gather responses from a small number of people you must make sure that the sample is as appropriate as possible to your research.

Larger samples are often employed in quantitative research. A basic rule of thumb is that increasing the sample size increases its reliability, although after a sample size of about 1000 the gains are less pronounced. Of course, you will need to take into account how much time you have as well as how much money is available should there by any costs involved. You should also consider non-responses; if you expect 100 people to fill in questionnaires, only 80 may do so. So it is a good idea to send questionnaires to more people than you need.

Preparing a questionnaire

The most common type of questionnaire is a self-completion questionnaire. You will probably have filled in this type yourself; a typical example would be a company asking for your opinion on their service.

Don't forget to design your questionnaire with regard to its function. It should be easy to read, with the questions spaced out clearly and distinguishable from the answer section and the preamble.

You also need to give clear information on the following.

• The aims and purpose of the questionnaire
• Who to contact for more information
• What happens next (will there be another questionnaire? Or a follow up interview? What happens to the data?)
• Instructions on how to answer the questionnaire
• How to send it back (including any deadlines)
• Whether you need the participant's contact details, and why. This is important as the respondent should be assured of confidentiality and anonymity. The questionnaire itself should not bear the name of the respondent.

One of the most important things to decide is how the questionnaire should be answered. You need to make it clear whether multiple-choice answers are available, especially if this changes from question to question.

Are you going to use a measurement scale?

These are used when you want to measure something that is latent (not directly observable, e.g. to measure an attitude.) You scale the answers like this.

• Agree _____
• Undecided _____
• Disagree _____

Have you given enough answer options?

It is not enough to just give 'positive' answer options. Some respondents may want to say 'no' or may not know how to answer. A Likert scale, where a statement is given and the scale points are levels of agreement (strongly agree, agree, unsure, disagree, strongly disagree) allows the respondent who is unsure of their opinion to respond.

Some open questions can follow closed ones to give the respondent a chance to express their own views. If you give your participants these options there is less chance that they will miss the question out or write their answer alongside the answer options, which can play havoc with your data analysis.

Kinds of question

When designing a questionnaire you will typically utilise closed questions rather than open ones.

• With closed questions, respondents are required to choose their answer from a set of pre-provided answers ('yes', 'no', 'don't know', for example)
• With open questions they can respond how they want by writing their answer in the space provided.

The advantage of closed questions is that the answers are easier to code and quicker to analyse. The disadvantage of closed questions is that they may not allow for every possibility, unlike open questions where the respondent has more freedom.

You may utilise both options. Some questions may be open and others closed. On occasion the respondent may be able to give more detail to their answer in a space provided.

Analysing questionnaires

Once you have collected your questionnaires you will have a lot of quantitative data to analyse.

You may have already coded your questionnaire. If you haven't, you can do it when you get it back, although having the analysis you want to do in mind helps when you plan the questionnaire and the coding list. Thus, knowing roughly what statistics you want to use is helpful. Coding is the application of labels (usually numbers) to the answers. For example, a questionnaire could ask how people travel to work. The answer options you give would have a number applied to each of them.

• Car (1)
• Bus (2)
• Bike (3)
• Train (4)
• Other (5)

You may have to code for missing data (when participants don't give an answer). Typically, such missing data is often coded as '99' because that exceeds the numbers likely to be used for other codes.

You would assign each questionnaire a number too, and then add the data from each questionnaire in number form to a programme such as Excel, SPSS or Access, thus giving you the chance to generate charts and graphs to better illustrate your answers.