Ambiguity is to be avoided at all costs. If a question is ambiguous, then the respondent may be presented with the dilemma of hearing or seeing two different questions and will not know which to answer. With an interviewer-administered questionnaire the respondent may seek help from the interviewer. The interviewer may be able to assist with the knowledge of the context of the question in relation to other questions,
but this may not always be the case. With self-administered questionnaires, respondents have to make their own decision as to what the question means. Either way, the researcher does not know which way the respondent has understood the question, except in the occasional instances where either the interviewer or respondent has recorded it. This rarely happens, though, except in pilot studies.
Ambiguity in the question can make it impossible for a respondent to know how to answer. Consider the following question:
Do your parents work full time?
There is no difficulty for the respondent if both parents work full time or if neither parent does (although a definition of what constitutes ‘full-time working’ would be helpful). If, however, one works full time and the other does not, what is the respondent to answer? The question would be better asked:
Do either or both of your parents work full time, that is more than
30 hours a week?
There still remains the issue of what constitutes ‘work’, and whether it should include unpaid work, such as charity work, or only paid work.
While some respondents may see the ambiguity and make a decision which way to answer, others may not see it and understand it only in the sense in which it was not intended. Then the answer given will not be the one that would have been given to the intended question and, again, the researcher is unaware of this.
If the ambiguity in the question is not spotted until the data have been collected, then the researcher has no way of knowing which respondents answered the question as intended and which answered the alternative meaning. This can render the data from that question incapable of interpretation and therefore useless.
Ambiguity is obviously to be avoided in questions, but is not always easy to spot. This is because it is not always possible to anticipate every respondent’s circumstances, and a question that may not be ambiguous to most respondents may, because of their circumstances, contain an ambiguity for a few. For example, ‘How many bedrooms are there in this property?’ is a simple question apparently incapable of more than one possible answer for most people. But what is meant by a bedroom? If someone has a study that doubles as an occasional bedroom, should that be included?
In most instances this level of ambiguity will not be a major issue. Where the number of bedrooms is collected as classification data to provide a cross-analysis of data by approximate size of house, then this degree of ambiguity may be acceptable to the researchers.
Where this information is central to the data collected, then the ambiguity must be addressed. In the example of the number of bedrooms, such ambiguity would be unacceptable in, say, a study of housing conditions. Then the question would require expanding, possibly to ask the number of rooms currently used as bedrooms, the number occasionally used as bedrooms and the number that could be used as bedrooms, or as required by the study.
Source: Brace Ian (2018), Questionnaire Design: How to Plan, Structure and Write Survey Material for Effective Market Research, Kogan Page; 4th edition.