Plan your dissertation questionnaire very carefully

• Telephone questionnaire
• Questionnaire sent via email
• Mail questionnaire
• In-house surveys
• Interviews

The telephone questionnaire is a popular way of conducting research as the interviewer can complete the questionnaire in a fairly short period of time. The main problem is that people may not want to take part in the research and simply put the phone down. This method of data collection is also one of the most expensive. There are some people who don’t wish to answer questions over the phone and so the data collected may be erroneous.

Email questionnaires are cheap, but not a very efficient way of gathering data. It’s easy to ignore a survey sent by email.

It isn’t cheap to send mail questionnaires and they are an inefficient way of data collection. It’s easy to simply throw your carefully crafted questionnaire into the nearest waste paper bin.

In-house surveys or interviews take time to conduct, but this is the most efficient way of ensuring that you will get the data you need.

You can go out into the street and interview various people and this way you can be sure that you have the age range you need for your research. However, people don’t always want to be interviewed on the street. Many market researchers stand outside a hotel, for example and take people inside to be interviewed.

There’s little doubt that face-to-face interviews are the most useful way of gathering data. The researcher can ask immediate follow-up questions to tease out full answers without the need for a further interview which could be difficult to arrange.

Researchers use a variety of different questions when conducting research which involves a questionnaire. The questions asked need to be carefully thought out beforehand and time has to be spent crafting them. The next stage should be to conduct a pilot study to check the efficacy of the questionnaire.

The following questions can feature in a questionnaire: -

• Open-ended questions
• Multiple choice questions
• Yes/no questions
• Scaling or ranking questions (think of the Likert scale)
• Tick the box

The open-ended questions provide researchers with rich data but can prove difficult to code before the analysis begins. These questions are ideal for a qualitative study.

Multiple-choice questions aren’t very flexible. If an ‘other’ category is included, this might mean that the researcher hadn’t anticipated possible answers to questions and it could be that the respondents who answer ‘other’ actually mean the same thing, in which case another category should have been added.

Yes/no questions don’t provide the researcher with sufficient data for analysis. There’s no scope to ask ‘Why?’ However, they can be statistically useful.

The Likert scale, grading answers from 1 – 10 can prove useful, although again, the data gathered will not be very rich.

Putting ticks in appropriate boxes can be used in some surveys, although again, the data obtained will not be detailed or rich.

When you design a questionnaire, your research question(s) should be uppermost in your mind. Then you need to consider how best to administer the questionnaire, as discussed above. Will you use interviewers to gather your data or will the questionnaire be completed by the respondents?

If a questionnaire is self-administered there can be no interviewer bias sneaking into the interview, but then the interviewer can’t be on hand to explain the questions, which might be necessary.

Questions should be brief and easy to comprehend. they should ideally be less than 25 words long. Don’t overcomplicate the questions with sophisticated vocabulary which might confuse respondents. Make sure that your questions are not ambiguous. Ask a colleague or friend to answer your questions and to point out any ambiguity.

You should also be careful about the length of the questionnaire. If you ask 20 Likert-type questions, that would probably suffice. However, forty or more would be too many. You will need to find out how long your respondents will spend filling in the questionnaire. if the respondents become bored, then up won’t get the rich data you need for analysis.

Some questionnaires include open-ended questions, which are usually placed at the end of a questionnaire that began by eliciting quantitative data. You can mix such questions, but if you do, you will need to think carefully about how you will go about our analysis of the data. One problem with mixing qualitative and qualitative questions is that the questions asked for quantitative data in the opening section of the dissertation questionnaire could influence the answers to the qualitative or open-ended questions.