Questionnaires

Asking the Right Questions

By definition, it is a method of obtaining specific information about a defined problem so that the data, after analysis and interpretation, result in a better appreciation of the problem.

There are two general types of questionnaires:

Highly structured

Unstructured

In practice we normally mix the two and use elements of both.

Careful planning of the questionnaire is vital to having a successful survey.

When designing the questionnaire, the researcher should keep in mind the classes of information that are generally used in making decisions:

Facts and Knowledge

Opinions

Motives

Past Behaviors

Future Behaviors

Be sure to always keep in mind these simple guidelines for ensuring a true response to a question:

Questions must be understood by the respondent

Respondents must be able to provide the requested information

Respondents must be willing to provide the information

Questions must be phrased in language that will be easily understood. The idea is to avoid the respondent having to make an interpretation of what you mean.

Choose words carefully and remember that not all words mean the same to all people.

Make questions specific.

Sounds easy, but in practice it can prove difficult. (text p.130).

Be concise and to the point. No rambling questions.

Avoid third-party questions--it is best to ask people what they know from their experiences.

Don’t ask hypothetical or irrelevant questions.

Questions should not place too much strain on respondents memory.

Questions should be worded carefully in order to avoid suggesting certain answers that are more acceptable than others.

Bias can also rear its ugly head in multiple choice and check lists if the research is not careful.

Do not mix high and low prestige items as this is certainly to lead to bias.

Be aware that the order of questions may tip respondents off as to a certain brand or product, resulting in bias.

The sequence of questions can have great influence on how respondents answer.

Several studies have been conducted which suggest that question order can stimulate the respondents mind and influence them to answer in a particular way.

The more specific the question the less likely it is to be significantly affected by the order of questions.

Most all questionnaires are voluntary and will rely on the goodwill of the people participating.

All efforts should be made before the start of a survey to reassure the participant that their participation is important and that all information will be confidential.

The opening question should be interesting and easy to answer to build up the respondents confidence.

The idea is to build rapport between the respondent and interviewer.

Questions should build in a logical order from general to more specific (funneling technique).

Sensitive questions should be left until the middle or towards the end, after a rapport has been built with the respondent.

With unstructured interviews, a skilled interviewer can often tailor sensitive questions to his/her subjects profile.

Filter questions can be used to determine the correct path of questioning.

Classification questions should be left until the end (age, marriage status, income, education, family size, etc.)

Grouping can be used successfully to obtain more responses to sensitive data.

Exercise: Write three open ended questions and three closed end questions concerning people’s musical tastes.

Pilot testing is vital to ensure that your questionnaire will hold up in the field.

Pilot testing may result in re-writing a portion, if not all of your survey, so be prepared.

Pilot tests should be done under conditions which will reflect the actual main survey.

Pilot tests have a variety of uses and can help in estimating a number of factors (see page 140).

Impersonal in nature

Question selection and wording is more important.

Open-ended questions need to be limited.

Instructions must be included and clearly written.

Where necessary definitions should be provided to aid respondents.

Time intervals need to be closely defined.

Presentation is very important--as is choice of paper and envelopes.

Address respondents individually.

Accompanying letter should outline the objectives of the survey and invite informants to respond.

The letter should stress confidentiality and provide some explanation of the sampling method.

Answers must be accepted as written.

Not suitable for surveys of people of low intelligence or poorly educated.

Because they can be read through completely before answering, bias can result.

Answers frequently represent the views of more than one person.

Follow-up is necessary for a higher response rate.

Telephone Questionnaires are growing in popularity.

However, especially in the U.S. telephone interviewing is becoming difficult.

Call screening and tele-marketers have made it difficult.

Advantages include cost and the ability to input data straight into the database.