METHODS READINGS AND ASSIGNMENTS   METHODS OVERVIEW

EVERYTHING YOU EVER WANTED TO KNOW ABOUT INTERVIEWING AND NEEDED TO LEARN HOW TO ASK: A MINI-TRAINING MANUAL ON PUBLIC OPINION SURVEY INTERVIEWS  c SUSAN CAROL LOSH 2017 "The Numbercruncher" Department of Educational Psychology and Learning Systems Florida State University Tallahassee, Florida 32306-4453 Do not cite without permission

And now you know why so many people really enjoy being interviewed.

Public opinion interviewing is probably a new experience for you. Sometimes, interviewing is frustrating--but it is rarely boring. You talk with and learn from a wide variety of people. That makes interviewing interesting and fun. But never forget that the FIRST purpose of an interview is to obtain complete, usable data from respondents. YOU are the eyes and ears of research. You ensure that respondents are comfortable, willing to answer and feel that the interview was a good experience. Plus you want to do justice to your own and the respondent's time by recording the interview accurately and completely.
 

• DO BE COMPLETE. At some point, all interviewers forget to record answers completely, generally on open-ended questions, or sometimes on the "OTHER" responses to closed questions. It's easy because you're talking, the respondent's talking, and you're punching computer buttons. Be warned! Omitted questions and answers mean a loss of information.

• DO examine and practice the questionnaire thoroughly before beginning to interview. While CATI (Computer Assisted Telephone Interviewing) requires each question to be answered, be especially careful on "OTHER" (PLEASE SPECIFY) responses and any open-ended questions.

• DO ASK every question in the interview schedule.NEVER ASSUME THAT YOU KNOW THE RESPONDENT'S ANSWERS AND ANSWER FOR THEM even if it seems R (the respondent)has already answered that question or--worse yet--you think you know what R's answer will be. If R partly answered an earlier Q (question), try saying "we've touched on this but let me ask you..." When Qs are clearly inappropriate (e.g., Qs about R's children and R has no children) the CATI system will skip you past them and you will never see that question.

• A good rule is: are you unsure whether to ask a question--or want more detail? Go ahead and ask!

• DO ask questions exactly as worded. DO NOT paraphrase. If each R is asked slightly different questions, the answers are meaningless.

• DO repeat and clarify questions that R misunderstood. Read response alternatives SLOWLY. Repeat the question if needed EXACTLY as written and DO NOT paraphrase.

• DO note questions that give R trouble. This helps us revise questions with wording problems.

• DO glance over your open-ended responses before entering them into  the computer. DO NOT use strange abbreviations. A coder and data entry person must understand this answer.

• DO be prepared. Have all your materials ready. Make sure that abbreviations, relevant telephone numbers are taped to your booth so that you can easily see them, that you have lots of "dispo" sheets, a hard copy of the questionnaire, a pencil, etc.

• DO practice the questionnaire in advance. You will be glad you did.

A PROBE is a non-directive way of clarifying answers, motivating R and focusing attention on the particular Q. Here are some SAMPLE PROBES:

• Could you tell me a little more about that? ("tell more")
• Is there anything else you would like to mention? ("AE")
• What is  your specific job title?
• Why  do you think that is so?
• Which answer comes closest to how you feel? ("feel")
• Do you lean more toward agreement or  disagreement?  NOTE THAT R IS PRESENTED WITH BOTH ALTERNATIVES.
• PAUSE AND WAIT. Many respondents need time to think.

• Uh-huh.  (a real good motivator)
• What would be your best guess on that? ("BG") USE THIS ONE ALMOST ALL THE TIME AS NEEDED.

    R says "I don't know" to a question. (NOTE ANY EXCEPTIONS)

    R will not answer a question. (NOTE ANY EXCEPTIONS)

    R misinterprets a question. (NEARLY ALWAYS)
    EXAMPLE:YOU: Do you think economic times will get better or get worse this year?
    R: I hope they'll get better.
    YOU: But do you think economic times will get better or get worse this year?

    R gives insufficient detail for the question.
    EXAMPLE 1: YOU: What is your occupation?
    R: I teach.
    YOU: What kind of teaching do you do?
    R: High school history.

    EXAMPLE 2: YOU: What do you see as the major problem in the Tallahassee area right now?
    R: The economy.
    YOU:  Could  you  give  me  an  idea what you mean by the economy?
    R: The national debt. (Side comment: some local problem...)
 

• DO NOT use probes to suggest answers to R.

• DO NOT change question wording when you probe.

DO use neutral phrases. "Uh-huh" or "I see" is great. "Good" suggests you agree with R--DO NOT use "good". Try to avoid "O.K."

Here are some abbreviations you probably want to include in open-ended responses. They help give us a flavor of how the respondent really answered the question:

• DK. Don't know. GIVE ALL DKs ONE PROBE. (NOTE EXCEPTIONS)

• BG. What would be your best guess on that?

• P. A general probe.

• AE. Was there anything else you would like to mention?

• DO combine abbreviations. R says "DK"--you probe--R still doesn't know. You record: DK (P) DK. By recording your probes, this shows you were complete.

 
• DO rehearse the "sticky situations" discussed later.

 
• DO USE R's OWN WORDS. DO NOT write what you thought R meant, but what R actually said.

 
• DO pay special attention to questionnaire instructions.

Some response categories are in CAPITAL LETTERS and say: "DON'T READ" or "VOLUNTEERED ONLY" after the response.

DO NOT read such response categories but if R volunteers such a response, check that response without probing. These VOLUNTEERED ONLY responses occur on several questions, especially several open-ended questions so BE COMPLETELY FAMILIAR WITH THE QUESTIONNAIRE so that you know when they occur.

DO be completely familiar with this manual. Every respondent will test you!
 

I'll bet you thought "I don't know" meant "I have no opinion about that". Sometimes when a respondent says "I don't know" it means just that. But sometimes "I don't know" means something else. Here are some other meanings of the "DK":

•  I know but I won't tell.
•  Give me time to think.
•  You read that question too fast and I didn't catch it.
•  Aw...you don't really want to know what I think. (HAPPENS LOTS)
•  I'm not absolutely surely 100 percent positive. (HAPPENS LOTS)
•  You won't like the way I want to answer that question.

MORAL: Nearly always give "I don't know" ONE PROBE. Try a "best guess" or "which answer comes closest to your feelings?" On the other hand, "I don't know" can be genuine so don't push very hard and probe only once. DO rehearse probing "DKs" until you do this automatically.

Why would anyone consent to a survey with a stranger, let alone then tell that person all kinds of (often very private) information? Because you are legitimate, aren't selling a product and want only their opinions. Your professional demeanor informs R your intentions are honorable. Your friendly manner tells people you are interested in what they have to say.Most people REALLY enjoy being interviewed!

"Rapport" describes the friendly "we-feeling" between you and the respondent that facilitates the interview. You must establish "just enough" yet not "too much" rapport with respondents for a solid interview. You are a stranger, yet a guest in R's home. R must be willing to trust you and also feel that you will hold a confidence.

• DO expect friendliness. Being interviewed is enjoyable since someone who is a good listener and interested in what you have to say is a flattering person to have around. After all, when was the last time someone genuinely wanted to know what you thought?

• DO be courteous. Introduce yourself and the survey quickly. Assure R of response confidentiality and other introductory information.

• DO be interested and show your interest.  When you say "uh-huh" or tell respondents you want to be sure you had their answer correctly, you tell R you are interested in what she or he has to say.

• DO be tolerant and nonjudgmental. Sooner or later, everyone interviews a respondent whose views they find insulting, woefully ignorant or threatening to a cause dear to one's heart. Relax. R has a right to hold opinions and it's safe to bet R won't agree with you on everything (or maybe anything) either.

• DO be patient. You are familiar with the questionnaire but R is not. You have a written script to follow. R does not.  Expect that R may have a few questions about the questions or may pause to think things over. This is especially true on the telephone when R has no visual cues as a guide. DO read questions and response categories slowly.

• DO remember that R's time and your time are valuable. Resist the temptation to get drawn into distracting conversations. This is especially true for telephone interviews which are designed to get the most information in the shortest time.

• DO be professional: no excessive joking, personal comments and NEVER tell R what you really think.  If you are TOO friendly, people clam up because you sound more like a friend than an interviewer.

• DO be confident. People are friendly. It's OK to be nervous (inside where it doesn't show).

Your introduction was written to give the most amount of information and establish rapport in the fastest possible time. DO read the introduction as written.

Keep the introduction as short as possible. Longer introductions=more hangups. Use these tips ONLY if needed. Start the interview as quickly as circumstances allow.

If needed, DO explain we CANNOT substitute. Respondents are uniquely selected with a probability sampling scheme resembling random numbers. This lets us represent the total population with a sample. Each respondent is special and cannot be replaced. Our refusal rates are low because of enthusiastic interviewers. We cannot talk with R's spouse, child, boss, or next-door neighbor.

DO stress confidentiality. We know R's name under very restricted circumstances if at all. Data are never used so that anyone could be identified. The data disc does NOT contain telephone numbers. Professional ethics hold us to confidentiality.

DO NOT introduce yourself so that is easy for R to say "no".

   NO!: Are you busy now? OR  Can I call back?
   YES: When is a convenient time for you?

DO NOT let respondents say they'll "call you back"! They won't.

   NO!: If I could, can I speak with your husband...
   YES: I'd like to speak with your husband

Hesitant phrases make it easy to say no. Assertive (NOT aggressive) statements get you on the line. Since R cannot see you, your voice is more important. Speak firmly and sincerely.

DO be sure to verify the telephone number at the very beginning.

DO NOT give survey details in advance. Your introduction is standardized to be the same for everyone. Stress that everyone's opinions are important to us--even if they don't have an opinion.

DO learn to recognize the "real" refusal. This person bluntly says "I won't do this" and hangs up. Sometimes they are even more outspoken than that. True refusals are rarely shy. Persons who are sick, busy, tired are not refusals. Often a call-back obtains the interview. In fact, on a second phone call ("the conversion call") at least one-quarter of "true refusals" do the interview--and usually enjoy it!

WATCH OUT FOR: The person who says they are busy now and makes a call-back appointment. You call back at the appointed time only to discover the person is busy again and they make another appointment. Somewhere around the fifth callback, you slowly begin to realize this person has no intention of doing the interview. They hope you will get tired of calling back and go away. This personable individual is a refusal in disguise! They are just too weaselly to tell you so. This is the time (if it didn't happen before) to change interviewers and try someone else with this respondent. Do NOT continue calling back such an individual without checking with your supervisor.

You have telephone numbers (Dr. Losh, the Survey Research Center, the Department office) for R to verify the survey if needed.

ON MONITORING

ALL interviewers are monitored. This verifies the legitimacy of the interview and corrects problems with question wording or the interview process. Telephone interviewers can be monitored at any time throughout the survey.

I have done A LOT of interviews (personally, several hundred) and I have trained over 300 interviewers. These Sticky Situations are ALL based on either my own personal experiences or interviewers I have supervised or monitored. (That's how come I know how well that "Best Guess" probe works.)

#1 You ask a question. R asks what you think. If you give your  opinion YOU CAN SERIOUSLY BIAS RESULTS. DO tell R you cannot give your opinion until after the interview. Repeat the Q and reassure R it is his or her feelings that count.

#2 Instead of answering Qs, R rewrites the questions for you, telling you how awful each question is and how it should read instead. Sometimes found among very well-educated Rs, this allows R to neatly sidestep questions presenting her or him in an unflattering light.  Try probing "I see but in general do you feel that.." then repeat Q. A Best Guess probe works well here too.

#3 R wants to know the Q's "right answer". DO stress that opinion questions don't really have "right" or "wrong" answers--it's how R feels that's important. It's the Best Guess probe again. (I promise: you will become automatic at this. After this tentative feeler from R, you will swing right into: "I see. But what would be your best guess on that?")

#4 R wants definitions of survey words or phrases. DO NOT give definitions unless they are provided for you in advance. DO tell R: "use the meaning that seems best to you" (write down that meaning if you can).

#5 R's friend, spouse, child, etc. gets on and answers instead of R. If R does not silence this intrepid soul, DO politely note that you can only interview R right now but you will be happy to survey the spouse (etc.) when R's interview is done. You may have to repeat this. Usually the interrupter vanishes.

#6 R becomes angry at a Q. Try one gentle probe. If the probe fails, mark the Q "REFUSAL" and continue.

#7 R makes YOUR blood boil.  Keep cool. Smile. Nod. NEVER argue. Let off the steam afterwards.

#8 R is about to hang up. Immediately obtain demographics at the end of the questionnaire before continuing and note this on a separate piece of paper. DO try to make an appointment to complete the interview. Termination does not necessarily mean you did anything "wrong". Prior appointments exist and unscheduled guests pop over.  Notify your supervisor immediately of any terminations.

#10 Your phone number has no eligible respondent. R is too old, too young, deaf, not registered to vote or whatever your criteria are. How do you tell that charming  gent at the Leon County Jail prisoner 'phone that you can't interview him? Simple: Complete any cover information, then say "uh-huh, thank you--that's all I need to know."

#11 The rambler. R goes on forever. You keep an anxious eye as the hands swing around the clock. What to do?

(a) The silent treatment...stop the "uh-huhs" and keep quiet. This is harder than you might think but it helps.

(b) "My parents never taught me manners." Did R actually answer the Q? Wait until R stops for breath then quickly ask the next Q. Has R strayed and not answered the Q? Wait for a breath then say "I see" then RQ. O.K., it's rude but neither you nor R will appreciate a two hour interview that never got finished.

(c) In person, you can drop all your pencils all over the floor. Everything stops cold while you pick them all up. Bring a few extra pencils just in case you need this one.

#13 R is deaf. (Or has severe speech problems or some other disability that makes completing a standard survey extremely difficult for them.) Thank R for their time and mark this interview NER (no eligible respondent).

#14 R wants to receive a copy of the questionnaire in the mail and fill it out themselves. Our questionnaires are usually not designed for self-administration and we keep very few printed copies so you can't do this. DO assure R you are getting every word. As a last ditch measure, we can send R a questionnaire after the survey is complete IF they really want this (afterwards). I will do it so please write down the name and address separately. DO THIS IN EXTREMIS PLEASE.
 

If a telephone number is non-working, IMMEDIATELY note this on your call report and on the "Dispo" sheet. Remember to submit telephone sheets and Dispo sheets to supervisors at regular intervals.

If R refuses to be interviewed or someone else refuses access to R, mark the call record "REF". COMPLETELY AND IMMEDIATELY fill out the "dispo sheet" with AS MUCH DETAIL AS POSSIBLE and notify your supervisor. Be gracious. DO NOT try the phone number again yourself. SOMEONE ELSE WILL DO THIS. Information about R or the access person (gender, estimates of age or ethnicity, time of day, circumstances) help enormously. This helps for another try.

DO record every call you make. Include date, day of week and TIME of day. This helps other interviewers who try that survey later.

Rarely a family moves after the first contact but before an interview is obtained. In this case do NOT use the new telephone number if one is given. Mark the number as not working.

Check for signs that a telephone number is inoperative or not a residential telephone. Check for answering machines, "funny-sounding" rings or "weird noises" when the 'phone picks up (a fax? a computer?). Recall peak interviewing hours are 4 to 8 P.M. weekdays. DO NOT waste time by calling a persistent "nobody home." The family could be on vacation.

DO remember to check at the beginning of the telephone interview for the correct phone number.

Sometimes any registered voter or any adult in the household qualifies for our sample. Other times the cover specifies whether you interview a man or a woman or a particular age for the respondent. Finally, some selection procedures are more complex. Who you interview may depend on how many people live in the dwelling and each person's age or gender. Follow whatever instructions you receive for respondent selection EXACTLY.

When you are selecting the respondent, who counts? Persons away at college or in the military DO NOT count. Traveling salespersons who make the dwelling unit their home DO count.

DO fill in your name under INTERVIEWER on the dispo sheet ONLY if you complete the interview.

Do most of your interviews take far longer than the average time described to you? Everyone gets a rambler now and then but if this happens to you ALL the time, try to control it. Practice the interview  schedule thoroughly. Have all your materials ready. Try the "silent treatment".
 

Some respondents (and interviewers too) may wonder what happens to interviews. The data are compiled. Averages and percentages are computed. No one's name or telephone number is ever attached to the datafile. In some studies, result summaries are sent to all Rs requesting them. In other instances, summaries are sometimes sent to city and county officials or the Tallahassee Democrat. Students use computerized data to write theses, papers, or dissertations. Some of these papers are presented at professional meetings or published in journals. Interviewers who wish to learn more about data analysis are welcome too. You are one part of a research complex from sampling to final reports.

Each survey study is unique with its own special instructions. DO be sure you are familiar with the unique instructions for your survey project.

Interviewing can be a lot of fun. Good luck, relax, and enjoy!

This page was built with Netscape Composer.
Susan Carol Losh

 

METHODS READINGS AND ASSIGNMENTS

BACK TO GUIDE 5: CLICK HERE

METHODS OVERVIEW