Step 2: Word open-ended questions to encourage in-depth responses and to avoid bias in asking about strengths and weaknesses. If closed-ended questions are desired, limit them to a small number of questions with response choices that can be communicated simply. All questions should be phrased appropriately for out-loud communication.

Here we present guidelines for open-ended and closed-ended interview questions that are planned in advance of the interview. These questions serve as the foundation of an interview protocol.

Open-Ended Questions. Open-ended questions are questions for which there are no predetermined categories of answers. These are the kinds of questions best suited for an interview, because they encourage the respondent to talk freely about a topic. Although there may be a few open-ended questions you need to ask to get straightforward information (e.g., "What kinds of computers do you have in your classroom?"), the emphasis of most open-ended questions should be on getting elaborated explanations from respondents (e.g., "Tell me about how technology gets used in your classroom."). By getting respondents to express their perspectives in language that is comfortable to them, you gain access to the depth and complexity of a topic.

A good open-ended question encourages the respondent to talk. Thus, the question should not invite or imply a simple answer. For example, a question that invites an answer of "yes" or "no" is not a good open-ended question (e.g., "Was the workshop leader engaging?"). Other questions that imply a "built-in" response also tend not to generate the kind of in-depth response that is the advantage of the open-ended format (e.g., "How important was the supplement?" invites short responses like "pretty important" or "not very important").

As a general rule, it is a good idea to avoid questions that begin with helping (i.e., auxiliary) verbs (e.g., "Was," "Have," "Do," "Should" or "Can"), because they can invite a "yes" or "no" response. Although starting a question with "Can you tell me about …" is a common interview technique that often invites explanation, it should not be overused because technically it falls in the "yes/no" camp. Questions that lead with "What," "How," or "Why" typically generate more thoughtful responses, although, as suggested above, the structure of "How X was Y?" implies a ""built-in" response. Below are examples showing how questions inviting brief responses can be reworded to invite more in-depth responses.

A good open-ended question should be neutral, avoiding the implication of any preconceptions or bias on the part of the interviewer. Avoid expressing opinions either right before or embedded within a question (e.g., "The instructor loved teaching this class. What do you think about the class?"). Also, be sure to balance your questions to tap different sides of a topic. If you are asking the respondent to think about the successes of a project, you also should ask the respondent to think about the weaknesses of that project or how the project could be improved. A common problem with interviews is that they are composed of a series of open-ended questions asking about many kinds of favorable outcomes. By asking only questions seeking favorable responses, you are not allowing the respondent to reflect on the full range of potential impacts and outcomes (including negative ones or ones you did not anticipate). The following table shows how an imbalanced set of interview questions can be elaborated to create balance.

Closed-Ended Questions. Closed-ended questions are questions for which there are predetermined answer choices. To structure a successful closed-ended question, you must anticipate the full range of possible responses and be able to list or chunk them in an organized fashion. The main advantage of closed-ended questions is that all respondents choose from identical answer choices that can be easily coded for quantitative analyses. For more specific advice about writing closed-ended questions, see Writing Questionnaires.

The use of closed-ended questions in an interview should be limited. If you wish to ask your respondents a significant number of closed-ended questions (say, more than five), a written questionnaire most likely is a better choice than an interview. If you have only a few closed-ended questions you want to ask, then it may be acceptable to include them as part of an interview (rather than construct a separate instrument). Another factor to consider is the complexity of the closed-ended responses. Closed-ended items with long lists of responses or complexly worded responses do not translate well to an out-loud presentation. Respondents may forget some of the responses or get confused about the exact distinctions between responses. As a result, the interviewer might have to devote valuable time to repeating the question and responses. When a more complex question is involved, one possible compromise is to show the respondent a card with the question and responses written out. The respondent may more easily reach an answer decision if he or she can read the entire item. The drawbacks here are that a special page has to be prepared and that this process interrupts the predominant conversational format. In short, the best candidates for closed-ended questions in an interview are those in which the question and choices can be expressed simply in an out-loud fashion.