2002 NSDUH Field Interviewer Manual

8. THE NSDUH INTERVIEW

8.1 Introduction

Chapter 7 discussed preparing for an interview, including the informed consent process. This chapter discusses the actual NSDUH interview process, including general standards for administering the interview, the contents of the questionnaire, and the procedures for you to follow at the close of the interview. Technical details about the CAI interview process are covered in Chapter 5 of the FI Computer Manual.

Two methods of administration are used in the CAI interview. The CAPI portions require you to read questions from the computer screen and enter the respondent's answers into the computer using the keyboard. During the ACASI segments, the respondent sits in front of the computer, puts on headphones through which he/she can hear the questions read, and enters his/her own responses into the computer.

8.2 Standardizing the CAPI Interview Process

The most crucial element of the interview process for a major national field survey such as the NSDUH is standardization. To the maximum extent possible, every interviewer must administer every questionnaire to every respondent in the same way. This helps eliminate variability and interviewer bias, two factors that can seriously undermine the validity, or "credibility," of the data collected. Standardization minimizes the variability in the way you, as the interviewer, ask questions.

RTI staff have developed some basic rules for administering the NSDUH interview. Many of these rules are standard to all interviewer manuals, but a few are specific to the NSDUH. Conducting all interviews using the rules outlined in this section ensures that the entire field interviewing team asks the questions in an unbiased manner.

8.2.1 General Questionnaire Conventions

The NSDUH questionnaire uses several different conventions that are discussed below. It is essential to become familiar with these conventions so that you can use them effectively.

1.  Lowercase text

• All lowercase black text not in parentheses is mandatory and must be read aloud by the FI in CAPI. In ACASI, the computer voice reads this text aloud and the respondent listens with the headphones.
  
  
• Black lowercase text (in parentheses) is optional. This is wording that does not have to be spoken aloud. Optional wording is provided where it might be helpful for you to "customize" or clarify the question or when an optional explanation is available.
  
  
• Lowercase text in blue signifies that the word(s) should be spoken with emphasis. For example, for the question "Are you currently on active duty in the armed forces, in a reserve component, or now separated or retired from either reserves or active duty?" the word "currently" would appear in blue on the screen so you would emphasize this word.
  
  
• Lowercase text in green below the question on the screen is either (1) interviewer notes which should not be read aloud unless the respondent asks for more details about a question or (2) family member information to be read as necessary to remind the respondent to whom a question refers.
  
  

2.  Uppercase Text

• Text in ALL CAPITAL LETTERS is never to be spoken aloud as it includes instructions for the interviewer. These may appear in the text of a question.
  
  

8.2.2 Asking the Questions

The following rules apply for the CAPI portions of the interview where you are asking the questions:

1. Ask the questions using the exact words on the screen.

This may seem obvious, but it is key to ensuring comparability of the data gathered from the survey. If you change the wording in a question, even slightly, the respondent's answer may change.

Imagine this situation:

Actual question: During the past 30 days, that is since January 1^st, 2002, how many whole days of school did you miss because you were sick or injured?

FI asks: During the past 30 days, that is since January 1^st, 2002, how many days of school did you miss because you were sick?

A respondent could answer two different numbers for the two different questions if the respondent had only missed partial days or had only missed days of school for an injury. The kinds of subtle changes in the phrasing of questions that are common when people are reading aloud can easily result in different answers. Read every question exactly as it appears.

2. Ask the questions in the exact order they appear.

Follow the order of questions as they appear on the screen. The computer automatically generates skip patterns based on the respondent's answers, so follow the exact sequence as it appears. If you feel you have reached a question in error, use the [F9] key to move backwards through the questions and check the answers recorded.

3. Read the complete question.

The respondent may interrupt you and answer before hearing the complete question. When this happens, politely explain that you have to read the entire question, then read the question again. Do not assume the premature response applies to the question as it is written. You must also read each question, even if you think you already know the answer from previous responses or conversation. With the change in context, the actual response may vary from what was said previously.

4. Read the questions slowly.

As you become familiar with the questions, you may be tempted to read more quickly. However, you must remember that this is the first time the respondent has heard these questions, so read slowly enough to allow the respondent to understand everything you are asking. A pace of about two words per second is recommended.

Give the respondent plenty of time to recall past events. If necessary, allow time for the respondent to check records such as insurance records, payment records, etc., to answer questions, but NEVER look at these records yourself.

5. Use introductory or transitional statements as they are written.

Since the questionnaire covers a number of topics, focusing the respondent's attention on a new topic as it occurs in the sequence is often necessary. Transition statements are provided and must be read exactly as written. These statements often contain instructions for the respondent—time periods, circumstances to be considered, definitions, etc.

6. Know how to pronounce the drug names correctly.

Although the specific drug questions are contained in the ACASI portion of the interview, it is essential that you know how to pronounce the drug names correctly if a respondent asks about them. Appendix E contains phonetic spellings of the drug names.

7. Do not suggest answers to the respondent.

As you go through the interviewer administered sections of the interview, you will come across questions you think you can answer based on information you heard earlier. You may want to suggest answers to the respondent. Or during the ACASI or CAPI sections the respondent may try to rely on you to help answer the questions. Your job as an interviewer is to ask the questions and make sure the respondent understands, but not to answer for the respondent. Do not help the respondent choose a response.

8. Become comfortable with the interview material.

Even though you are not administering the majority of the interview, it is important to be familiar with the questions so you can help respondents if any problems arise. Some questions may make you uncomfortable at first because they ask about personal matters, such as drug use and illegal acts. Even so, you should be familiar enough with the questions to be able to explain what is being asked.

9. Be nonjudgmental.

It is essential that you interact with respondents in a nonjudgmental manner. It is possible that respondents will tell you about actions that are illegal or that you find to be immoral, sad, or shocking. Regardless of what you hear, you must accept the information without judging. It is important that you accept all that you hear in a matter-of-fact manner so that the respondent continues to feel comfortable providing confidential information.

8.2.3 Probing

This section discusses "probing," a technique used to help ensure that the answers given by the respondent are as accurate and complete as possible. Probing serves two purposes: 1) to help the respondent understand the question, and 2) to obtain a clear, codable response from the respondent without suggesting answers. Be sure to use an appropriate neutral or nondirective probe. General rules for probing include:

1. Repeat the question if the respondent misunderstood or misinterpreted the question.

After hearing the question the second time, the respondent likely will understand what is being asked. This is usually the best and most straight forward method of probing. Similarly, repeating the answer choices suggests to the respondent that the response needs to come from these choices.

2. Repeat the answer if the response is too vague to answer the question.

For example, if you ask the respondent for his/her current marital status and he/she answers, "I'm on the loose," an effective probe is, "On the loose?"

3. Pause to indicate to the respondent that you need more or better information.

This is a good silent probe after you have determined the respondent's response pattern.

4. Use neutral questions or statements to encourage a respondent to elaborate on an inadequate response.

Examples of neutral probes are "What do you mean?" "How do you mean?" "Tell me what you have in mind" or "Tell me more about...."

5. Use clarification probes when the response is unclear, ambiguous or contradictory.

Be careful not to appear to challenge the respondent when clarifying a statement and always use a neutral probe. Examples of clarification probes are "Can you give me an example?" or "Could you be more specific?"

Exhibit 8.1 lists specific examples of acceptable probes.

Exhibit 8.1 Examples of Acceptable Probes Repeating All or Part of the Question "Let me read the question again." Repeat the full question with emphasis. "Let me read the question again just to be sure we are focusing on the same time period." Repeat the full question with emphasis. Repeat the answer choices with emphasis. Wait 10-15 seconds. If no answer is given, repeat the full question. Requesting a Clarification "For this question, I need an answer of either 'yes' or 'no'." -or- "Is that a 'yes' or a 'no'?" "Can you give me a specific number of (hours/days/weeks/etc.)?" "Would you say 9 or 10?" Reassurances "To the best of your knowledge . . . " "Remember that we will not tell anyone what you say." "Remember, there are no right or wrong answers." "Take your time." "I have to ask all of the questions in order." "We're interested in your own interpretation of the question." "We just want to know what you think." "Whatever [WORD] means to you." Repeating an Earlier Question "Let me read you an earlier question."

8.2.4 When and How to Probe

The following explains when and how to probe to obtain information that is complete, accurate, and useful to the researchers who analyze the data.

RESPONDENT DOES NOT GIVE A SIMPLE "YES" OR "NO" ANSWER

• Gives a clear answer, but not "yes" or "no"
  
  If the respondent gives an answer to a yes or no question that is completely clear ("absolutely" "not at all", "never!"), enter "yes" or "no" even though the respondent did not actually say "yes" or "no." There is no need to probe in such cases.
  
  
• Says "yes" or "no" and more
  
  Respondents often say "yes" or "no" and then give you additional information. For example, if you ask, "Have you received treatment for psychological or emotional problems at a mental health clinic or by a mental health professional on an outpatient basis in the past 12 months?" the respondent might say "Yes, but it was only because my father died and I was really depressed for a few months. I'm not crazy." The additional statement explains why, but it does not contradict the "yes" answer. In such cases, ignore the additional information and enter the "yes" or "no" answer without probing.
  
  On the other hand, if you think the additional information might contradict the "yes" or "no" answer, repeat all or part of the question. For example, when asked if the respondent or any of the respondent's family members living at the DU received food stamps in the past 12 months, the respondent says, "no, just Grandma." You know from the recently completed household roster that Grandma does live at this residence. Since the additional information seems to contradict the "no" answer, repeat the question emphasizing the phrase "... or anyone in your family living here..."
  
  

UNCLEAR RESPONSES

• Unclear numeric answers
  
  Some questions ask for a number (an age, the duration of something, etc.). If the respondent answers with "a few," "I was pretty young," "quite a while" etc., then say: "Can you give me a specific number of (hours, days, etc.)?"
  
  Often respondents answer these questions with more than one number ("when I was 18 or 20 years old," "for two to three days," etc.). For many questions in the NSDUH we only want a single number as an answer, not a range, so it is necessary to encourage respondents to give us their best opinion as to a specific number.
  
  If the respondent is unable to answer with a single number even after probing, choose the more extreme number given. Examples are:

  –    a time - code the category with the most recent time (ex: "2 or 3 weeks ago," record "2 weeks") –    an amount - code the category with the greatest amount (ex: "$25,000 or $28,000," record "$28,000") –    a frequency - code the category with the highest frequency (ex: "once or twice a week," record "twice a week").

  If the response is a fraction and the question asks for:

  –    an age: drop the fraction (ex: "18½ years old," record "18 years") and record the younger age; –    a number of days, years, miles, etc.: round fraction to next highest whole number (ex: "8½ days," record "9 days") and record the greater number.

• Gives a vague or unclear response
  
  If the respondent seems to have understood the question, but gives a vague or unclear response such as "it depends" "Oh, Lordy!" "sometimes yes, sometimes no," then gently say, "Is that a 'yes' or a 'no'?" If you are not sure the respondent understands or remembers the question, repeat the question, saying "Let me read the question again." When repeating the question, pause slightly between parts of complex phrases, and emphasize any words that will help the respondent understand (focus your emphasis on the time frame and the terms that describe frequency).
  
  

• Gives an unrelated response
  
  

  Sometimes respondents go off on a tangent without answering the question. Do not discuss personal issues with the respondent. Be polite. Without offending the respondent, bring the focus back to the job at hand by listening, smiling to acknowledge the respondent's words, then repeat the question, emphasizing important words.

• Doesn't understand the question
  
  

  If the respondent tells you he or she does not understand a question, or if the answer makes no sense and indicates that the respondent did not understand, say "Let me read the question again" (repeat the question, emphasizing key words). Note: The first sentence ("Let me read the question again") is optional and should be used only if you think it helps you maintain rapport with the respondent.

• Misunderstands the time period
  
  

  If the respondent says something that makes you think he or she is not focusing on the correct time period for the question, reread the key part of the question or the entire question, emphasizing the time period and pointing to the relevant time period on the calendar. The calendar is designed to make the recall task easier. For example, if you ask a 35-year-old if he has worked in the last 12 months and he says, "Yes, when I was 20 years old," he is not focusing on the correct time period (the last 12 months).
  
  

• Misunderstands a frequency requirement
  
  At times, respondents give answers that indicate they may not have understood the frequency requirement of the question.
  
  

• Misunderstands an intensity requirement
  
  Sometimes you will need to probe because the respondent does not understand or ignores the intensity or severity that is asked about in the question.

DON'T KNOW, REFUSED

• A respondent says "I don't know"
  
  When a respondent says "I don't know," it can mean one of three things: 1) the respondent may need more time to think about the answer; 2) he or she may be reluctant to tell you something personal; or 3) he or she actually does not know the answer to the question—or is unsure of the best answer.
  
  If a respondent says "I don't know . . . hmm . . ." and appears to need more time to think about an answer, then wait silently and expectantly for an answer. If you feel it would help, you may say "Take your time."
  
  If you think the respondent has said "I don't know" because he/she is reluctant to tell you private information, you should put the respondent at ease by saying "Remember that we will not tell anyone what you say" or "Remember, there are no right or wrong answers."
  
  Sometimes respondents truly don't know the answer to a question. For example, if you ask about certain items covered by their health insurance, the respondent may simply not know. The first time that a respondent says "I don't know" in this way, say To the best of your knowledge . . . " On future questions, however, you may accept the "don't know" response in the interest of not alienating the respondent.
  
  Note: Listen closely to respondents to be sure they are really saying that they don't know "Oh, I don't know, probably not" could actually mean "no" to the respondent, while "I don't know, I could have" may mean yes. Don't guess what the respondent means. Instead, say "Is that a 'yes' or 'no'?" using a pleasant tone.
  
  

• Refuses to answer a question
  
  After repeating the assurance of confidentiality, allow the respondent to exercise the right "to refuse to answer any or all questions." You should also enter "Refused" if a respondent does not feel comfortable answering a specific question. Add a remark about the circumstances (see Section 5.7 of the Computer Manual for instructions on how to add a remark in the CAI system).
  
  

• Does not answer a question
  
  If the respondent does not answer, wait 10-15 seconds to give the respondent time to think. If there is still no response, say "Let me read the question again" and then repeat the question emphasizing key words. If no answer is given a second time, code as "Refused" and continue. Sometimes respondents fail to answer a question they find very personal rather than flatly refusing to answer. Sometimes, however, they have become unwilling to continue with the interview. If you suspect that this is the case, ask the respondent if he or she would like to take a break. They may feel like continuing after a few minutes to themselves.
  
  

• Does not give a codable answer
  
  In the unlikely event that the probing strategies do not work in obtaining a codable response from the respondent, enter a comment in the remark box for that question. Use a "Don't Know" as the response to the question so you may continue with the interview.
  
  

8.2.5 Recording Responses

Most of the questions in the NSDUH questionnaire have precoded responses. A few questions, however, are open-ended so you must type in a response to the question. Some questions have precoded responses including an "Other (SPECIFY)" category. If the respondent's answer does not fit into a precoded answer, then choose "Other" and specify the response.

When recording open-ended or other specify responses, follow the recording practices below to ensure the recorded responses accurately reflect the respondent's answers.

• Record the response verbatim typing the keywords the respondent says. For example, the respondent is specifying her race and says "I think of myself as...," there is no need to record that first phrase. The important step is to accurately record the race.
  
  Record answers in English, even if the interview is being conducted in Spanish. Only record a Spanish word if there is no comparable English word to translate the answer.
  
  If you run out of room in the "other specify" box, make a note for the question, continuing the answer in the remark space (see Section 5.7 of the Computer Manual).
  
  

• Record your notes for any responses that "don't quite fit" in one of the response categories for that question. Remember to record any general remarks about the interview or interview setting of the questionnaire. Your notes will help the analysts to understand any points of confusion, difficulty, etc. (see Section 5.7 of the Computer Manual).
  
  

• Wrong code—If you mistakenly enter the wrong answer or if the respondent changes his/her mind, go back to the question by pressing [F9] and use the backspace key to erase the old answer. Then type the correct answer.

Since respondents complete most of the questionnaire on their own, answer their questions about recording responses in this same way.

8.2.6 "Don't Know" or "Refused"

"Don't know" or "Refused" responses generally do not appear as answer categories in the questionnaire although they are within the acceptable range of responses. Since these responses are not displayed on the screen, the respondent is more likely to answer the question within the "preferred" ranges. Follow the probing techniques discussed in Section 8.2.3 to encourage accurate reporting. However, "Don't know" and "Refused" are always possible responses for any question in either the CAPI or ACASI sections.

• If a respondent refuses to answer a question while you are administering the CAPI, reassure the respondent and ask again. Then if needed, press the [F4] key for "Refused." Press [Enter] to move to the next question.
  
  
• If a respondent tells you he/she doesn't want to answer a question during the ACASI portion, repeat the confidentiality concepts before telling the respondent to press the [F4] key if the respondent still doesn't want to answer.
  
  
• If a respondent doesn't know the answer to a question you have asked and a "don't know" code is not available, probe first. Then, if necessary, press the [F3] key and then [Enter] and the computer will move to the next question.
  
  
• If a respondent doesn't know an answer during the ACASI portion, tell the respondent to make a best guess. If needed, say to press the [F3] key.
  
  

8.3 The NSDUH Interview Content

Exhibit 8.2 contains a chart listing the various topics involved in the NSDUH interview. The chart also lists any additional materials that are needed for each particular section. The first section of the interview contains questions on basic demographics (such as age, sex, and race) and is administered in CAPI. This is followed by a brief ACASI practice session. After the practice, the respondent completes the ACASI section, which contains the questions about drug use, mental health, and certain experiences and opinions. When the respondent has finished this section, turn the computer back toward yourself and administer a second CAPI component which contains more specific demographic questions, plus questions about income and health care. Following this second CAPI portion you are asked to complete a short series of FI debriefing questions. These questions are NOT to be read aloud to the respondent.

Exhibit 8.2 NSDUH CAI Interview Content

Module	Mode of Administration	Required Aids
Introduction	CAPI*	None
Core Demographics	CAPI*	Showcards 1-4
Overall Health Question	CAPI	None
Calendar (30-day and 12-month reference dates)	FI reads instructions and identifies reference dates.	Reference Date Calendar
Computer Practice	Respondent completes computer practice session with FI help.	None
Tobacco	ACASI**	None
Alcohol	ACASI	None
Marijuana	ACASI	None
Cocaine	ACASI	None
"Crack"	ACASI	None
Heroin	ACASI	None
Hallucinogens	ACASI	None
Inhalants	ACASI	None
Pain Relievers	ACASI	Pillcard A
Tranquilizers	ACASI	Pillcard B
Stimulants	ACASI	Pillcard C
Sedatives	ACASI	Pillcard D
Special Drugs	ACASI	None
Risk/Availability (item count questions)	ACASI	None
Specialty Cigarettes	ACASI	None
Drug Dependence and Withdrawal	ACASI	None
Special Topics	ACASI	None
Market Information for Marijuana	ACASI	None
Drug Treatment	ACASI	None
Health Care	ACASI	None
Adult Mental Health Service Utilization	ACASI	None
Social Environment	ACASI: 18+ year-olds only	None
Parenting Experiences	ACASI: parents of 12-17-year-olds	None
Youth Experiences	ACASI: 12-17-year-olds only	None
Serious Mental Illness	ACASI: 18+ year-olds only	None
Youth Mental Health Service Utilization	ACASI: 12-17-year-olds only	None
Back-End Demographics:	CAPI	None
Education	CAPI	Showcards 5 & 6
Employment	CAPI	Showcards 7-11
Household Roster	CAPI	Showcards 12 & 13
Proxy Information	CAPI	None
Health Insurance	CAPI (Proxy allowed)	Showcards 14-17
Income	CAPI (Proxy allowed)	Showcards 18 & 19
Verification	FI & respondent complete form.	Verification Form & Envelope
Incentive Payment	FI completes form and pays cash	Interview Payment Receipt, cash
FI Observation Questions	FI records own responses.	None

* CAPI (computer-assisted personal interviewing): FI reads questions and records responses.

** ACASI (audio computer-assisted self-interviewing): Respondent reads questions on screen or listens to questions through headphones and then records answers into computer.

8.4 Introductory Statement & Informed Consent

At the beginning of the interview the computer asks you if you have read the Introduction and Informed Consent script to the respondent. If you have not, you must do so at this point. See Section 7.6 for information about Informed Consent Procedures. If—and only if—the respondent asks, read the burden of proof statement that appears next on your computer screen.

8.5 Verifying Eligibility

The first question the computer prompts you to ask is the respondent's date of birth. This question is asked for two reasons:

• to verify that the respondent is age 12 or older before beginning the interview (individuals under age 12 are not eligible for the survey), and
  
  
• to figure the respondent's age at the time of the interview so that the appropriate interview questions are displayed.
  
  

8.5.1 Underage or Active Military Respondents

We cannot interview respondents who are not at least 12 years old. It is possible you could begin an interview and discover that the respondent is not yet 12 years old. If this happens, do NOT re-roster the entire household. Instead, the computer asks you to confirm that this response is correct and ends the interview. If the respondent is found to be ineligible during the ACASI portion of the interview, the computer will take the respondent to the end of the ACASI portion. Then you take back the computer and enter the 3-letter code (RTI) to continue. After you enter this code, the computer exits the interview.

The same situation could occur if at some point during the interview you discovered the respondent is on active duty military status, which is why this question is asked at the beginning of the interview. Again, do not re-roster the household. The CAI Manager ends the interview automatically and considers it a completed case.

In either of these situations, you must discuss the case with your FS. With approval, code the case in the Newton as a 79 "OTHER," and document the situation in your Newton Record of Calls.

8.6 Use of Showcards and Pillcards

During the interview, there are times when it is helpful to the respondent to see the list of answer categories or see other visual references to aid in answering a question. Called Showcards, they are bound together into the Showcard Booklet. There are many advantages to using Showcards:

• The Showcards list answer categories when the list is long or complicated so the respondent can review all choices carefully.
• Using the Showcard Booklet encourages the respondent to report more honest answers to sensitive CAPI questions such as income or education level.
• Other household members who may be nearby do not hear the actual response.

Because of these benefits to data quality, you are required to use the Showcards while conducting an interview.

Also included in the Showcard Booklet are color pictures of various types of pills, called Pillcards. The respondent uses the cards as a visual aid when answering ACASI questions about the use of pain relievers, tranquilizers, sedatives, and stimulants. The respondent is instructed as to when to ask for each card. If a respondent fails to ever ask for the pillcards, please note that detail in the FI Observation questions at the close of the interview.

Using tab dividers, the showcards are separated from a series of helpful Job Aids, included at the back of the Showcard Booklet for your reference.

8.7 Completing the Reference Date Calendar

Early in the interview, you and the respondent complete a calendar to define the reference periods—past 30 days and past 12 months—used throughout the interview. An example of a properly completed calendar is shown in Exhibit 8.3. The calendar is printed on legal size paper (8.5" x 14") with the months from last year and this year printed on one side. In the center is a box in which you record the two reference dates. Send all completed reference date calendars to your FS weekly.

The computer calculates the appropriate reference dates for you. Instructions appear on the computer screen to assist you as you prepare the calendar with the respondent.

For example, assume you are conducting an interview on May 18, 2001.

• First, the computer determines the 30-day reference date. This date is NOT the same date last month, meaning 4/18/02 is NOT the correct date. In this case, the computer will determine that April 19^th is the 30-day reference date.

• Write the date in the box on the front of the calendar. Then circle that day, and underline the entire 30-day reference period for the respondent.

• The computer will tell you the 12-month reference date. This will be the interview date, one year earlier. For example, an interview conducted on May 18, 2002, would have May 18, 2001, as the 12-month reference date.

• Write the 12-month reference date in the box, and circle it on the calendar as well. Do not underline the entire 12-month reference date.

• Record the Case ID at the top of the calendar.

This calendar is a tool proven to increase accuracy in respondent recall of events. Using a computer means the correct dates automatically appear in questions asking about time periods. However, the process of mentioning the dates and having the respondent watch as you circle the day and underline the 30-day reference period clues the respondent into the exact time period of interest. Even if the respondent never looks at the calendar again, you have given a visual tool reminding which day of the week of the 30-day reference date is, and how many weekends are included in those 30 days. This information is critical for accurate reporting.

You must complete a calendar with each respondent, following the steps outlined on the computer screen so that each respondent can receive the full benefit of the reference date calendar.

Exhibit 8.3 Reference Date Calendar

8.8 FI Responsibilities During the ACASI Portion

After you complete the reference date calendar, you see a screen that looks like this:

BEGINNING ACASI SECTION An important part of this interview is the sections you will conduct completely on your own using the computer and the headphones. Before you begin, I will help you to go through a short practice session to learn how to use the computer. MOVE COMPUTER SO RESPONDENT CAN USE IT AND POINT OUT THE FOLLOWING: LABELED KEYS IN TOP ROW (FUNCTION KEYS) NUMBER KEYS [ENTER] KEY SPACE BAR BACKSPACE KEY THE BOTTOM OF THE SCREEN WHERE THE ANSWERS WILL APPEAR CAUTION RESPONDENT ABOUT ON/OFF SWITCH. ADVISE RESPONDENT NOT TO USE THE TOUCHPAD. ADJUST HEADPHONES FOR RESPONDENT AND DEMONSTRATE VOLUME CONTROL. WHEN RESPONDENT IS READY, PRESS '1' TO CONTINUE.
IntroAcasi

After reading the brief introduction at the top of the screen, you are ready to train the respondent to use the computer. Your role before the practice session begins is to familiarize the respondent with the keyboard, particularly the location of the [ENTER] key, the space bar, the backspace key, the "function" keys, and the number keys. The function keys are particularly important to the respondent during the ACASI portion of the interview. Below are the function keys the respondent needs to know (however, the use of "don't know" and "refused" options should not be encouraged):

This only needs to be a brief orientation explaining the points listed on the transition screen, because the respondent completes a detailed practice session on using ACASI. However, tailor your presentation to the needs of your respondent. If you have someone who is uncomfortable using computers, spend a bit more time reassuring them. An example of an appropriate explanation of the introduction of the computer is included in Exhibit 8.4. Be sure to always offer the headphones to the respondent as part of your explanation.

Once you are sure the respondent is ready to begin, press [1] to start the ACASI practice session. The practice session explains what the respondent needs to know to complete the ACASI interview, but be prepared to answer questions. It is important that each respondent complete the practice session on his or her own. Even computer experts need to understand how this particular computer program is set up and which keys perform which tasks.

Tell the respondents that when they see the screen that asks for the 3-letter code they will have completed the ACASI sections and should let you know. The screen instructs the respondent to hand the computer back to you. Remember that the respondent should complete the ACASI interview in private; therefore, you should only sit beside the respondent during the ACASI practice session. Then move to some location in the room where you are unable to see the computer screen. Do not leave the room. Remain available to hand the respondent the pillcards or in case the respondent has questions or needs help with the computer.

You need to be available to help if your assistance is required, but be unobtrusive. Try to occupy yourself constructively so the respondent does not feel rushed. However, be aware of what is happening with the respondent. Because most of this questionnaire is self-administered, it is important to be sensitive and responsive to subtle clues from the respondent. In particular, be receptive to signals that the respondent is having a problem with one of the questions or with the computer.

Some suggestions for how to fill the time during the ACASI are as follows:

• Prepare the Verification Form and envelope for use at the close of the interview.
  
  
• Prepare the Interview Payment Receipt for use at the close of the interview.
  
  
• Update your Record of Calls on the Newton.
  
  
• Review additional cases to be worked that day, modifying your plans as needed.
  
  
• Re-read portions of the FI Manual or the FI Computer Manual. Take a chapter or two at a time out of the manual binder to read. Regular reviewing of the detailed procedures in the manual greatly improves your knowledge of the project and your understanding of the project protocols.
  
  
• Review the Job Aids in the Showcard Booklet.
  
  

When the respondent is finished with the self-administered portion, usually about 30 minutes, take back the computer and ask the questions for the remaining sections.

Exhibit 8.4 Sample ACASI Introduction Script

8.9 Industry and Occupation Questions

Several questions in the back end Demographic section (following the ACASI portion) ask for details about the respondent's job. The details of the respondent's answers are used to assign two standard 3-digit codes that accurately classify the work in a particular industry and as a particular occupation. To correctly assign the appropriate codes, the coding staff requires very detailed responses.

When a respondent is asked about the type of business or industry his/her employer is engaged in, interviewers should probe and obtain as much detail as possible. The following examples illustrate the level of specificity interviewers should strive for:

If a respondent replies with his/her job title, restate the question so that the respondent understands you are asking for the type of business or industry of his/her employer. If the respondent reports being self-employed, probe to find out what type of business the respondent is engaged in. To further understand this concept, refer to Appendix F where some example pages of the Industry Code Book are displayed. While you as an interviewer do not have to determine which industry code to assign, you must obtain enough detailed information for the coding staff to do so.

When a respondent indicates that his/her employer is engaged in manufacturing, you will then ask the respondent what specific product his/her employer makes. Likewise, if a respondent replies that his/her employer is involved in wholesale or retail trade, you will ask the respondent what specific product his/her employer sells.

Gathering the same level of detail also applies to questions asking about the respondent's occupation. If the respondent states a very general occupation, an interviewer should probe for as many specifics as possible:

If a respondent gives you a job title that conveys little or nothing about what his/her occupation is, probe to try to determine the trade, profession, or business. A respondent may respond with his/her educational attainment (e.g., I have a medical degree) but you should not assume the respondent works as a doctor. Probe to find out what the respondent does (perhaps rather than working in a medical practice, he/she writes medical articles for a weekly magazine). If a respondent begins listing duties or responsibilities rather than replying with an occupation, tell the respondent you will collect this information in just a minute, then repeat the question and emphasize that you are asking for his/her occupation. Appendix G displays some example pages of the Occupation Code Book, again as an illustration of the level of coding specificity involved.

When recording the respondent's duties, capture enough details to adequately and accurately describe the work done. If the respondent is a supervisor, record the department or function that he/she supervises. The following are examples:

Accounts Receivable Department Manager	Supervises Account Receivables Clerk, makes sure accounts are kept current and paid on time, prepares reports and journal entries
Graphic Designer	Creates signs, banners, brochures, newsletters and other promotional materials using computer graphics
Secretary	Types, files, answers phones, makes appointments for boss
Waiter	Takes food orders, places order with kitchen, serves food to customers, makes sure customers are happy
Portrait Photographer	Takes pictures requested, develops film, prints pictures, contacts and invoices customer
Tool Design Engineer	Designs parts for existing machines that will modify the parts currently being produced by the machines

The respondent will also be asked to categorize his/her employer in terms of the employer's status. For instance, is the respondent's employer a private company? If so, is the company for-profit or not-for-profit? If the respondent works for a government, he/she will be asked to indicate the level of government (e.g., state, local). Self-employed respondents will be asked to indicate if their business is incorporated or unincorporated.

For all of the industry and occupation questions, it is important to always probe for further details. Be sure to use adjectives in your probes—for instance, find out what type of doctor or salesperson, what kind of engineer or programmer, supervisor of what—and record the respondent's answers as accurately and completely as possible.

8.10 Respondent Difficulties

At some point, you may encounter a respondent who is impaired in some way. The main concern is whether the respondent is physically and emotionally capable of participating in the study. This is a judgment call that you must make.

You may encounter a respondent with physical limitations, such as blindness, deafness, or paralysis. Some people with physical handicaps will be able to operate a computer, although others may not. This study was designed for self-reporting, so only in the instances of the severely physically impaired are you permitted to enter a respondent's answers in the ACASI portion of the interview for him/her.

• Blind respondent. In the case of a blind respondent, you will sit beside him/her to assist, but the respondent will wear the headphones to listen to the questions. In some instances, the respondent may feel comfortable entering the ACASI responses on his/her own. Place the respondent's fingers on the 1, 2, and Enter keys, reminding him/her to ask for help when needed. If the respondent is not comfortable entering responses, have the respondent tell you the answers he/she wants you to enter into the computer.

• Physically impaired respondent. If the respondent is not able to enter the responses into the computer because of a physical limitation, then he/she can either wear the headphones to listen to the questions or read the questions him/herself from the computer screen. Again, the respondent tells you the answer he/she wants you to enter into the computer.

• Deaf respondent. In the case of a deaf respondent, the only deviation from the standard procedures in the ACASI portion is that the headset is not used. In order to participate, a deaf respondent must be able to read. First, you will need to hand the Introduction and Informed Consent script from the Showcard Booklet to the respondent to read. Then, you will need to allow the respondent to read the initial CAPI questions and the introduction to the respondent practice questions. You should sit beside him/her as he/she does this and point out the keys on the laptop to the respondent during the practice questions. The procedures for the ACASI module will be just like any other respondent who chooses to read the screens rather than listening through the headphones. So you should move to where you cannot see the respondent's answers for this part of the interview. For the back-end demographics section, you will sit beside the respondent and enter the responses for him/her after the respondent has read the screen and the Interviewer Notes on each screen. It may be necessary for you to communicate with the respondent on paper if he/she has questions or if you need to probe.

You should never, under any circumstances, read the questions in the ACASI portion of the interview out loud to the respondent. This would be a major breech of study protocol, and the case could not be included in the analysis—so, despite the best of intentions, you would have wasted your time, the respondent's time, and lots of project resources. With the ACASI recording of the sensitive questions, there is no need for the interviewers ever to read these questions. Consequently, the protocol states that doing so is in violation of acceptable procedures for the interview and cannot be allowed.

If a physically impaired respondent begins the ACASI section and later decides that it is too difficult to finish on his/her own, you should give the respondent the option of having you key the answers for him/her, following the procedures detailed above, or coding the case as a final breakoff. If a respondent has physical limitations that prevent him/her from participating even with your help, document this situation and check with your supervisor.

In all of these cases that deviate from the standard protocol, explain the situation at the end of the interview in the FI Observation questions, add a comment in the ROC in the Newton, and discuss with your supervisor.

Another challenge you may come across is how to interview a respondent with poor or no reading skills. Younger respondents (12-17), respondents whose native language is not English, respondents who cannot read or write, elderly respondents, or those with hearing or vision problems may fall into this category. You may not become aware of a person's inability to read (or write) until you pass the computer to the respondent. The respondent may not tell you he/she cannot read. Instead, he/she may try to hide this inability by using the excuse about being "too busy" to complete the rest of the interview and ask you to do it instead.

The questionnaire is designed so that you have all the tools you need to assist even the poorest of readers to follow along and enter the answers in the computer. In fact, the computer questionnaire may be easier for nonreaders because using the headphones, they can hear all questions and answer categories read aloud. During the CAPI portion of the interview, you read the text in the CAI questionnaire and the respondent answers aloud. Then during the ACASI, the respondent listens to the questions and types the corresponding number of the answer. Code numbers are associated with each answer category, so you can direct the respondent to the specific number to press if he/she is confused. Studies show that even the poorest of readers can usually recognize and follow along with numbers.

It is critical that these administrative procedures be followed exactly. The questionnaire is designed for ease of use with poor or non-readers, so it is not appropriate for you to enter any answers in the computer unless your respondent is physically incapable of doing so. It is never appropriate for you to answer questions for a respondent who is capable but simply does not want to take the time or trouble to do so. The respondent-completed ACASI design is key to maintaining the confidentiality of responses. If you have a respondent who resists completing the questionnaire for himself/herself, you must make every effort to explain the purpose of the survey's design, the importance of private and confidential responses, and the fact that this is the procedure you must follow to properly conduct the interview. Here is an example:

"In order to do my job properly, it is very important that I conduct this interview according to the procedures developed by SAMHSA and by RTI. These procedures ensure that every interview across the entire country is being done the same way—that is, that every respondent is completing his or her own questions. This also ensures the privacy and confidentiality of the answers you give."

If a respondent insists that he/she is too busy at the time of your visit, offer to return at another time when he/she can complete the questionnaire unaided.

8.11 Post Interview Procedures

8.11.1 Verification

Once the questionnaire is finished, you must complete the verification process. The Verification Forms (see Exhibit 8.5) are printed with English on one side and Spanish on the other. A form must be submitted for every completed interview, even if the respondent refuses to fill out the form. Do not submit to RTI a form for breakoffs: wait until you finish the entire interview to complete and mail a Verification Form.

The CAI Manager prompts you to enter the Verification ID (letter and numbers under the barcode) from the Verification form that you have prepared for the interview. Carefully follow the instructions printed on the computer screen.

To prepare a Verification Form:

1. Select a Verification Form and enter the VerifID in the laptop when prompted.

2. Prepare the interviewer portion of the Verification Form by writing the current date and the approximate interview completion time, circling a.m. or p.m.

3. Carefully copy the Case ID from the Newton to the appropriate boxes at the bottom of the form. Be sure to enter the correct letter, either A or B.

4. Print your name and FI ID # in the FI portion (the bottom half) of the Verification Form.

5. If the respondent is a 12-17 year old, you also need to complete the box in the bottom right corner to indicate which adult gave permission for the interview.

   NOTE: Steps 2-5 should be completed during the ACASI portion of the interview.

6. Read the instructions to the respondent. Have the respondent print his/her home telephone number and address on the Verification Form. Hand him/her a small postage-paid envelope and ask him/her to put the Verification Form in the envelope.

7. Have the respondent seal the envelope.

8. Mail the envelope containing the Verification Form.

If the respondent refuses to complete the Verification Form, you must:

1. Select a Verification Form and enter the VerifID in the laptop.

2. Prepare the Verification Form as you normally would by completing steps 2-5 above. Note that these steps should be completed during the ACASI portion of the interview.

3. Write "refused" on the Verification Form.

4. Mail the Verification Form to RTI in the small postage-paid envelope.

If the respondent is a 12- to 17-year-old, the Verification Form should be completed by a parent (or guardian). Use the steps described above, explaining the process to the parent. If the parent is unavailable (but has given consent for the youth to complete the study), it is acceptable for the 12- to 17-year-old to complete the form. Just remember that the preference is always for the parent or guardian to complete the Verification Form for a youth.

Exhibit 8.5 Verification Form

Exhibit 8.5 Verification Form (Continued)

The Data Quality staff at RTI who deal with the Verification Forms offer the following reminders:

• Do NOT use a Verification Form you have already reported for another case, even if the first respondent left it blank.

• For households with 2 respondents, keep the Verification Forms for the A and B cases separate. Each respondent should complete the top portion of the appropriate form, placing it in a separate envelope and sealing it.

• Pink Verification Forms from training sessions are not acceptable—the CAI interview program is programmed to reject all training VerifIDs.

• Carefully enter the VerifID into CAI, entering the letter and numbers including the dash. Complete this task before you hand the form to the respondent

• Likewise, carefully record the correct Case ID number on the actual Verification Form for the interview case. Be sure to include an A or B.

• Interviewers are only to complete the bottom of the form. Respondents complete the upper portion, including the date and address. Respondents are also to place the completed form in the envelope and seal it, not you. These procedures are in place to stress confidentiality and must be taken seriously.

8.11.2 Field Interviewer Observation Questions

While the respondent is completing the Verification Form, answer the series of questions in the laptop about how you felt the interview went, providing details about the situation. The questions ask your opinion about the levels of respondent understanding, cooperation, and privacy. They also ask about the respondent's experience with the interview itself. You are also given the opportunity to make any notes or comments that may help analysts understand the situation or the respondent. Do not feel you have to enter a note; only add notes that truly clarify a given situation. When considering privacy, the only two people who should be in the room are you as the interviewer and the respondent. Use these questions to record any other persons (not pets!) who were present or listening to the interview. If a proxy respondent is used during the income/insurance questions, this does not cause a breach of privacy unless the proxy is present during non-proxy sections of the interview. The interview observation questions are listed in Exhibit 8.6.

Exhibit 8.6 FI Observation Questions

8.11.3 Incentive Payment Procedures

You are responsible for providing a cash payment of $30 to each respondent who completes an interview. Once you finish explaining the Verification Form to the respondent, the CAI Manager displays the instructions for the incentive payment. Procedures are as follows:

• Show the respondent the Interview Payment Receipt (see Exhibit 8.7).

• Hand the respondent the cash.

• Record the Case ID, including the A or B.

• Sign and then date the form.

• Mark the appropriate box indicating the respondent either accepted or declined the cash payment.

• Give the respondent the top copy of the completed receipt. Mail the yellow copy to your FS along with the calendar in your regular weekly shipment. Keep the pink copy for your records.

If the respondent refuses to accept the money, you must mark the appropriate box on the Interview Payment Receipt. Some respondents may ask you to donate the money to a charity instead of accepting it. You cannot donate the money to a charity for a respondent. Instead, encourage respondents to accept the cash and donate it to their favorite charity themselves.

Exhibit 8.7 Interview Payment Receipt

8.11.4 Packing up the Equipment

Pack up the computer equipment by doing the following:

• gently close the screen display until you hear a click

• unplug the parts of the power cord (from the computer first then from the wall plug)

• unplug the headphones and dispose of the covers

• place the laptop in the open carrying case and secure it with the velcro straps

• wrap up the power cord parts, the extension cord, and the headset and place them in the compartment next to the laptop

• zip the carrying case shut and you're ready to go.

Remember to pack up the Showcard Booklet, reference date calendar, and any other papers and put them in your carrying case. And don't forget the Newton!

8.11.5 Newton ROC

Remember to update the Newton ROC with the final interview code (70). Do this before or as you are leaving the DU. Do not enter the 70 in the Newton until the interview has been completed. It is always possible that you could have a breakoff unexpectedly and while you could edit the code before transmission, it is easier to record the code correctly the first time than to have to edit and then reenter the correct code.

8.12 Thank the Respondent

Be sure to sincerely thank the respondent for spending time with you and for contributing to the success of this important national study. Do your best to leave the respondent feeling very positive about the interview.

And now thank yourself for a job well done!

REVIEW OF CHAPTER 8
The NSDUH Interview

To assist your learning process, read the following summary of key points in this chapter. Following the summary are several questions to ask yourself. These questions will help you identify those areas you understand, and pinpoint the areas where you would benefit from re-reading a particular section.

SUMMARY

• The CAI interview is administered using two different methods: CAPI and ACASI. During the CAPI portions, you read questions from your computer screen and record the respondent's answers. The sensitive questions appear during the ACASI segment—the respondent listens to and/or reads the questions and records the answers in private.

• To ensure high quality data, the NSDUH interviews must be standardized. In other words, every interviewer must administer every questionnaire to every respondent in the same way. Using questionnaire conventions and CAPI rules make standardization possible.

• To ensure that answers given by the respondent are as accurate and complete as possible, we use a technique called probing. Probing can help the respondent understand the questions and also help you in obtaining a clear response. Correct probing NEVER suggests answers to the respondent.

• During the interview, you show the respondent Pillcards and Showcards to serve as helpful visual aids. In addition, you complete a reference date calendar which helps the respondent better conceive dates and periods of time.

• If you encounter respondents who are physically or mentally impaired, use sensitivity and good judgment to determine whether or not they are capable of participating in the study. In situations where the respondent is physically incapable of completing the interview, it is acceptable for you to help complete the ACASI portion of the interview. With the respondent listening to the audio, you will record the answers as provided by the respondent. This is the ONLY situation where you are allowed to complete the ACASI questions.

• After completing the interview, the respondent completes a paper Verification Form as part of our quality control procedures.

QUESTIONS TO ASK YOURSELF

1. What does the formatting of the below text signify?

   lowercase black text

   (lowercase black text in parentheses)

   lowercase text in blue

   TEXT IN ALL CAPITAL LETTERS

2. Answer these True / False questions.

   T	F	It is acceptable for you to switch the question order if it helps the R.
   T	F	You are never allowed to suggest answers.
   T	F	If the R interrupts you, you still must read the entire question.
   T	F	If you think an R may not know a certain word, it is permissible to replace a word to help the R understand.
   T	F	If the R asks you for your opinion, you are allowed to give it.

3. If, after reading a question to the respondent, he appears to have misunderstood or misinterpreted it, what probe would you use?

4. If you mistakenly enter the wrong answer, how can you go back to the previous question?

5. What key is pressed if a question is refused?

6. Is the set of questions about tobacco administered using CAPI or ACASI?

7. What will you record at the top of the reference date calendar?

8. What does the result code of 70 mean? When should you enter it in your Newton?

This page was last updated on May 16, 2008.

SAMHSA, an agency in the Department of Health and Human Services, is the Federal Government's lead agency for improving the quality and availability of substance abuse prevention, addiction treatment, and mental health services in the United States.