Research Questionnaire

Research Questionnaire

The content of the questions is the most important element of the construction of a questionnaire. While the order of questions may influence access to information, the content of questions will lead to the type of information sought in the study. Questionnaires are constructed systematically. The process goes through several interrelated steps.

1. The Dichotomous format

The dichotomous format is actually any question that can be answered with a dichotomous response: true/false, yes/no, agree/ disagree, hot/cold, and so on. The dichotomous format is widely used in social science questionnaires. The format is very flexible and can be used in a great many different ways. It is frequently used in classrooms as a device to find out how much of the material presented has been learned by the students. Personality inventories typically request dichotomous responses, usually in a checklist form (where checks indicate "true" and blanks indicate "false").

2. Multi-choice format

The primary use of multiple-choice formats in social science is probably in the construction of tests of ability and aptitude. Such well-known tests as the GRE (Graduate Record Examination) and SAT (Scholastic Aptitude Test) are two examples of a large number of such tests. Aptitude test answers are always known ahead of time and thus responses are coded as correct-incorrect.

Multiple-choice questions contain more" information" than do dichotomous questions. This is true because of "guessing." In a true-false question, one can get the correct answer half of the time by guessing. In a multiple-choice question with four possible answers, one can get the correct answer only a quarter of the time by guessing.

3. Fill-in-the-blank type format

The fill-in-the-blank format is closely related to the above two types of formats, and in sore ways is a natural extension. Suppose, for example, we ask a question such as, "Who was president of the United States when the Interstate Freeway system was initiated?" The format is a fill-in-the-blank type although it resembles the multiple choice in that if one knew the names of presidents of this century the task is to choose among them.

The amount of information collected per question is a function of the number of possible responses. Thus the amount of information in a fill-in-the-blank question is greater than a multiple-choice question and a multiple-choice question requests more" information than a dichotomous question. This is true because guessing an answer becomes less likely as the number of possible increases. In some preliminary studies, we carried out it was found that one fill-in-the-blank question had as much "information" as five dichotomous questions. This means those fewer fill-in-the-questions are needed to roach the same degree of reliability.

4.Matching format

In presenting a matching task format, informants are normally presented with two lists that are linked in some way but are scrambled in presentation. The task is to match items from one list with items on another list. For example, historic figures might be matched to their birth dates. This format is frequently used in classroom test situations since it is easy to make up questions in a short time. It is not frequently used informal research situations. There are probably some situations in which it would be very useful.

5. Estimation format

The use of estimation procedures is frequently overlooked partly because in the cases where they might be used one could measure the variable under consideration directly and not elicit the information from informants. For example, if one were interested in how tall people are, the natural thing would be actually to measure their height. However, estimates of height or distance are of value precisely because they can be validated and informant accuracy can be assessed. Thus estimates can serve as benchmarks to compare different data collection techniques.

6. The pick N format

In various guises, this format has already been covered. For example, the triad task of picking "the item most different from the other two" is a pick one of three-type format. Multiple-choice questions are a pick one of n (usually four) types. In early sociometric studies, it was typical to ask subjects to pick their three best friends from a class list. Nowadays in sociometric studies, we strongly recommend the use of complete rank ordering of preferences if at all pick possible. about half In cases the stimuli, where this instead is not possible, it is probably information best to is contained in such a split than in the choosing three. More cases where one picks only a few items.