Editor's note: John C. Stevens is lead supervisor at Saperstein Associates, a Columbus, Ohio, research firm. He can be reached at 614-261-0065 or at firstname.lastname@example.org. This article appeared in the April 22, 2013, edition of Quirk's e-newsletter.
Many researchers take advantage of significant cost savings by conducting surveys by Internet rather than by telephone. Opportunities await those who can harness big data or social media listening for their clients. This doesn't mean that telephone interviewing will become obsolete; it just means that it will become more challenging - and interviewers must rise to the occasion.
An interviewer is charged with persuading respondents to participate in a telephone survey, reading the questions without error and recording responses accurately. Of course there is no substitute for an interviewer's genuine interest in people and enthusiasm for gathering useful and illuminating information but great interviewing also starts with a few very basic skills. So whether you manage a telephone facility or take part in interviewing yourself, mastering the following six skills will encourage the highest-quality telephone research.
Listen closely when a respondent answers the phone. You may need to determine the respondent's gender immediately. You may need to assess their mood and decide how fast or slowly to read questions. Listen to objections against participating in the survey so you can address them.
Also, listen for when your respondents are not listening. If you ask a question about a proposed heartbeat bill in the state legislature and your respondent starts talking about how heart problems run in the family, you know that the respondent did not listen to your question. You will need to repeat it.
Reading a questionnaire exactly as written may sound simple but it often takes more concentration than expected. You will be tempted to change the wording of questions just slightly because you think that respondents will understand the questions better or because you think your version of the question sounds more natural. You may be tempted to explain questions to respondents who do not understand them.
Even slight or subtle changes to the wording of a question can change the meaning of a question. A respondent might say yes if you ask if they could do something but no if you ask if they would do something. Your interpretation of the meaning of a question could be radically different than what the author of the questionnaire intended or what the client wants to learn. Read each question exactly as it is written to each respondent so analysts can compare how respondents answered questions.
Pay attention to how fast you read an introduction to a questionnaire when a respondent answers the phone. Nervousness may cause you to read too fast. If a respondent can't hear you clearly, many will assume you're selling windows and hang up immediately. You can speed up, slow down and adjust your volume appropriately after you hear your respondent say a few words.
Nervousness may also cause you to inflect improperly when you read your introduction. If you inflect up at the ends of sentences that are not questions, you will make them sound like questions (e.g., "Hello? My name is John Smith? I'm calling from XYZ Opinion Company?").
Reading an introduction this way shows a lack of confidence. Respondents find it annoying and may decline to participate. When you inflect up at the ends of sentences in the introduction, you are actually asking a question. The question you are asking is "May I continue?" And the answer will often be "No."
Also pay attention to where your mouthpiece is positioned. If it is too close to your mouth, your respondents may become distracted by hearing you breathe.
Ask a supervisor or project manager about words you do not know how to pronounce. Doing this may be a little embarrassing but not as embarrassing as respondents asking if you mean "specific" and not "pacific" or as problematic as having your supervisor ask you about responses that do not make sense because you said "ideal" instead of "idea."
Accents aren't a problem as long as they don't distract or confuse respondents.
Greet the respondent in a hearty and sincere manner. Pause for a second after saying hello before continuing your introduction. People will be more likely to participate in a survey if they think an interviewer is interested in talking to them.
To compete with robocalls, you may have to sacrifice a little voice clarity to convince respondents that they are listening to a human being when they answer the phone. You may find it helpful to throw in an "um" or cough or clear your throat when reading an introduction. Some respondents may ask if they're speaking with a real person; others will simply hang up on you if they think they are listening to a prerecorded message.
Once you persuade a respondent to participate in a survey, you must get him or her to answer the questions. This can be more difficult than it sounds. Respondents often interrupt. If a respondent interrupts before you finish reading a question, they won't hear the question as written, causing the data to be inaccurate. Many surveys that require a representative sample of a population require the interviewer to ask to speak to the adult in the household who is having the next birthday. Respondents often interrupt before they hear the "next birthday" part of this question and say, "I'm an adult!" If you take a stand early on and say, "I need to make sure you hear the whole question," you may irritate some respondents and lose them. Most respondents will respect assertiveness and will become more cooperative for the rest of the questionnaire.
Another way to prevent interruptions to questions is to inflect up a little at the end of each option when reading a list of possible choices. This lets the respondent know that more choices are coming.
Do not express surprise or astonishment at anything a respondent says. You can tell someone at the beginning of a questionnaire that it is a confidential survey but respondents will feel more comfortable spilling personal information or giving socially-unacceptable answers if they perceive the interviewer as nonjudgmental. If a respondent says they no longer shop at a mall because members of an ethnic group shop there, the client needs to know that. Do not end the interview or lecture the respondent because they make offensive comments.
Some respondents are difficult to control. They may want to tell you why they agree or disagree with a statement or tell you about their gripes with family members. These rants will often answer upcoming questions. You will need to be assertive and ask something like "So should I mark 'Agree' or 'Disagree?'" You can get a call back on track by saying something like "I need you to hold that thought, it may answer a question coming up later." Then read the next question.
If you can refrain from saying "okay" or "uh-huh" when a respondent answers a question, you will have mastered 75 percent of the skill of neutrality. It is a natural human feeling to want to let your respondent know that you heard the answer to a question but these grunts can be interpreted by the respondent as approval or disapproval of their answers. If a respondent thinks that you approve of their answers, they will want to give you more such answers. If a respondent thinks that you disapprove of their answers, they will clam up and not give you honest answers about how they feel.
If you are working on a political poll or asking questions about hot-button social issues, it is important that you do not read answer choices in such a way that you reveal your own opinions about candidates or issues. Your respondent should not know that you agree or disagree with their answers. Remember that you are collecting information, not trying to change anyone's point of view.
Recording information accurately
The best way to ensure that you record responses to answers accurately is to acquire 10-key skills. That is, learn to use the number pad on the right side of your keyboard instead of the row of numbers across the top of the keyboard. You should learn to key by touch rather than by sight. This way, you can see what you have keyed by looking at your screen and you will make fewer interviewing errors in general if you don't take your eyes away from the screen.
When you type answers to open-ended questions, you may find it helpful to say what you are typing while you are typing it if the respondent gets ahead of you. If this does not work, you may need to say, "Could you hold that thought a second? I want to make sure I get this in your own words." This will give you a chance to catch up while at the same time giving the respondent the message that their opinions are important.
Ask yourself three questions after a respondent gives you their reply:
Is it clear?
- Is it complete?
- Does it answer the question?
If you cannot answer yes to all three of these questions, you'll need to ask a probing question. Your probing question should not be a leading question (e.g., "Do you mean..."). The respondent may mean what you think - but maybe not. Use simple and direct probing questions: What do you mean? How so? Can you give me an example? What can you tell me about that?
Put yourself in the place of the client. The client commissioned a survey to gather information to help make decisions. The campaign manager for a political candidate needs to understand how voters perceive the candidate to decide how to craft advertisements. If a respondent says that they will not vote for a candidate because the candidate is a slacker, you need to find out a little more. The client probably already knows what a slacker is so if you ask "What do you mean?" and the respondent gives you a definition of the word "slacker," you have not collected useful information. You could ask "What gives you that impression?" Do not ask if they could "elaborate on that." Not all respondents know what "elaborate" means and many are likely to say no.
You can often probe effectively simply by repeating a response in a tone of voice that says, "Hmm, that sounds interesting." Inflect up when you do this so that it sounds like a question, such as "Social media will ruin your social life?"
Starts with good data
All interviewers can improve their skill set by taking the above recommendations to heart and all project directors and supervisors would be wise to encourage these best practices. After all, telephone interviewing is data-gathering in its simplest form and we all know that good research starts with good data.