Confidentiality & Informed Consent

The NLS program has established set procedures for ensuring respondent confidentiality and obtaining informed consent.These procedures comply with Federal law and the policies and guidelines of the U.S. Office of Management and Budget (OMB) and the U.S. Bureau of Labor Statistics:

• OMB procedures
• Respondents' advance letter
• Institutional review boards
• Federal laws
• Survey procedures
• Data handling

OMB procedures

The U.S. Office of Management and Budget (OMB) reviews the procedures and questionnaires for each NLSY round.

OMB is responsible for setting overall statistical policy among Federal agencies. For example, OMB has established standards on collecting information about race and ethnicity, industry, occupation, and geographic location. OMB also has established standards on the manner and timing of data releases for such principal economic indicators as the Gross Domestic Product, the national unemployment rate, and the Consumer Price Index. In addition, OMB sets standards on whether and how much respondents to Federal surveys can be paid for their participation, an issue of particular concern in the NLS program.

Another of OMB's responsibilities is to review the procedures and questionnaires that Federal agencies use in collecting information from 10 or more respondents. Federal data collections reviewed by OMB include administrative data, such as the tax forms that the Internal Revenue Service requires individuals and corporations to complete. OMB also reviews all censuses and surveys that Federal agencies conduct, either directly or through contracts.

OMB examines a variety of issues during these reviews, such as:

• The amount of time (and money, if any) that the agency collecting the information estimates respondents will spend to provide the requested information
• The agency's efforts to reduce the burden to respondents of providing the information
• The purpose and necessity of the data collection, including whether it duplicates the objectives of other Federal data collections
• The ways in which the agency obtains informed consent from potential respondents to participate in the data collection
• The policies and procedures that the agency has established to ensure respondent confidentiality
• The statistical methods used to select representative samples, maximize response rates, and account for nonresponse
• The payment of money or the giving of gifts to respondents
• The questionnaire itself, including the quality of its design and whether it includes questions that respondents may regard as sensitive

These OMB reviews are very thorough. From the time an agency prepares an OMB information collection request until the time OMB approves the data collection, the process typically takes 7 months or more and includes multiple layers of review within each agency and at OMB. These reviews are helpful in improving survey quality and ensuring that agencies treat respondents properly, both in terms of providing them with information about the data collection and its uses and in protecting respondent confidentiality.

The review process also provides the general public with two opportunities to submit written comments about the proposed data collection. The agency conducting the data collection publishes a notice in the Federal Register describing the data collection and inviting the public to request copies of the information collection request, questionnaires, and other materials that the agency eventually will submit to OMB. The public is invited to submit written comments to the agency sponsoring the data collection within 60 days from the time the Federal Register notice is published. In the history of the National Longitudinal Surveys program, the public very rarely has submitted comments to BLS, but when comments are received, they are summarized in the information collection request that ultimately is submitted to OMB.

After the request has been submitted to OMB, the agency sponsoring the data collection then publishes a second notice in the Federal Register and invites the public to submit comments directly to OMB within 30 days. Again, in the history of the National Longitudinal Surveys program, the public very rarely, if ever, has submitted comments to OMB.

Once OMB has received the information collection request, they have 60 days to review the package, ask follow-up questions, suggest changes (or occasionally insist upon changes) to the survey questionnaire or procedures, and ultimately grant approval.

Respondents' advance letter

After OMB grants approval, the sponsoring agency can begin contacting potential respondents and collecting information from them. The process of contacting potential NLS respondents begins with sending them an advance letter (See Figure 1 and Figure 2) several weeks before interviews are scheduled to begin.

The advance letter serves the following purposes:

• informs respondents that an interviewer will be contacting them soon
• thanks respondents for their previous participation and to encourage them to participate in the upcoming round
• reminds respondents that their participation is voluntary
• tell them how much time the interview is expected to take
• explains how the data will be used
• explains how respondents' confidentiality will be protected by BLS and the organizations that conduct the surveys for BLS
• provides a required privacy and disclosure statement on the reverse side of the letter (see Figure 2)

Institutional review boards

In addition to OMB review, the NLSY97 is reviewed and approved by the institutional review boards (IRBs) at the institutions that manage and conduct the surveys under contract with BLS. Those institutions are The Ohio State University and NORC at the University of Chicago. BLS and OMB do not require these reviews; rather, they are required under the policies of the universities.

Obtaining IRB approval includes the following parts:

• completion of a form signed by the Principal Investigator, providing a summary of the research project, and submitting a description of the consent procedures and forms used in the survey.
• submission of a copy of any materials used to recruit respondents, a detailed summary of the survey questionnaire, and any other information regarding the risks to humans of participating in the survey
• application for approval prior to the start of each round of data collection
• a general focus by the IRBs on respondent compensation, consent procedures, and confidentiality protections for special populations (such as incarcerated respondents). Note: IRBs also focus on obtaining information about the potential risks to participants, which is particularly relevant for medical studies. Although the NLSY97 has included health-related topics, information about these topics has been obtained only through questions asked of respondents, rather than through any medical exams or procedures. As such, NLSY97 respondents bear little to no risk in participating.

Prisons, schools, and other institutions where NLS sample members may reside often request the IRB approval statement and application as evidence that appropriate procedures are being followed and to judge whether to permit NLS interviewers to have access to individuals for whom the institutions are responsible.

Federal laws

Two Federal laws govern policies and procedures for protecting respondent confidentiality and obtaining informed consent in the NLSY97: the Privacy Act of 1974 and the Confidential Information Protection and Statistical Efficiency Act (CIPSEA) of 2002.

The Privacy Act and CIPSEA. These two acts protect the confidentiality of participants in the NLSY79, its associated Child and Young Adult surveys, and the NLSY97. CIPSEA protects the confidentiality of participants by ensuring that individuals who provide information to BLS under a pledge of confidentiality for statistical purposes will not have that information disclosed in identifiable form to anyone not authorized to have it. In addition, CIPSEA ensures that the information respondents provide will be used only for statistical purposes. While it always has been the BLS policy to protect respondent data from disclosure through the Privacy Act and by claiming exemptions to the Freedom of Information Act, CIPSEA is important because it specifically protects data collected from respondents for statistical purposes under a pledge of confidentiality. This law strengthens the ability of BLS to assure respondents that, when they supply information to BLS, their information will be protected. In addition, CIPSEA includes fines and penalties for any knowing and willful disclosure of specific information to unauthorized persons by any officer, employee, or agent of BLS. Since the enactment of the Trade Secrets Act and the Privacy Act, BLS officers, employees, and agents have been subject to criminal penalties for the mishandling of confidential data, and the fines and penalties under CIPSEA are consistent with those prior laws. CIPSEA now makes such fines and penalties uniform across all Federal agencies that collect data for exclusively statistical purposes under a pledge of confidentiality.

NLSY interviewers explain federal policies regarding confidentiality and consent in the following manner:

• Interviewers explain to potential respondents that all the employees who work on the surveys at BLS, NORC, and The Ohio State University's Center for Human Resource Research (CHRR) are required to sign a document stating that they will not disclose the identities of survey respondents to anyone who does not work on the NLS program and is therefore not legally authorized to have such information. In fact, no one at BLS has access to information about respondents' identities, and only a few staff members at NORC and CHRR who need such information to carry out their job duties have access to information about respondents' identities.
• Interviewers also explain that the answers respondents provide will be made available to researchers at BLS and other government agencies, universities, and private research organizations, but only after all personal identifiers-such as names, addresses, Social Security numbers, and places of work-have been removed. In addition, the publicly available data files exclude any information about the states, counties, metropolitan statistical areas, and other more detailed geographic locations where respondents live, making it much more difficult to infer the identities of respondents.
• Respondents are told that some researchers are granted special access to data files that include geographic information, but only after those researchers undergo a thorough application process at BLS and sign a written agreement making them official agents of the BLS and requiring them to protect the confidentiality of respondents. In no case are researchers provided with information on the personal identities of respondents.
• Finally, because the reference in the questions and answers to the National Archives and Records Administration and the General Services Administration may be confusing to some potential respondents (because those Federal agencies are not involved in the administration of the surveys), interviewers explain to respondents that NLS data and materials will be made available to those agencies because they are responsible for storing the Nation's historical documents, but that the information provided to those agencies does not include respondents' personal identities.

The organizations involved in the NLS program continuously monitor their security procedures and improve them when necessary. Protecting the privacy of NLS respondents entails considerable responsibilities for BLS, the organizations that conduct the surveys for BLS, and the researchers who use the data. Indeed, researchers in particular may become frustrated that they cannot obtain access to all the data that they want or that they must undergo a long review process at BLS to obtain some types of data. It is important to remember, however, that protecting respondent confidentiality must remain paramount. Any action that might jeopardize respondent confidentiality and erode the confidence of respondents could harm response rates in the NLS program and in other government or academic surveys. Thus, without safeguards in place to protect respondent confidentiality, researchers would have far less data available to work with than they currently enjoy.

Survey procedures

Like all contractor staff, field interviewers are agents of BLS and are required to sign the BLS agent agreement before working on the NLSY97. All interviewers also must undergo a background check when they are hired. Confidentiality is stressed during training and enforced at all times. Field interviewers receive specific instructions in their reference manuals to remind them of the appropriate procedures when locating or interacting with respondents or contacts.

At the end of each interview, interviewers ask respondents to provide information on family members, friends, or neighbors who can be contacted if the interviewers are unable to locate the sample member in a subsequent round of interviews. The interviewers then use those contacts to help in locating sample members who have moved. When contacting a sample member's relatives, friends, or neighbors about the sample member's whereabouts, interviewers never disclose the name of the survey they are conducting. They are instructed to maintain the confidentiality of any relative, friend, or neighbor who provides information about the sample member's whereabouts.

Answering machines can pose potential problems when contacting sample members because it is difficult to confirm that the interviewer is calling a sample member's correct phone number or that other household members will not hear the message. For those reasons, interviewers are instructed not to leave messages on answering machines.

When interviewers contact the appropriate household, they ask to speak with the sample member or the parent of a sample member under age 18. Interviewers introduce themselves and state the purpose of the call by saying that they are from NORC at the University of Chicago and are calling concerning a national survey. The name of the survey is not disclosed to anyone but the sample member.

Special situations

The NLSY97 is a general population survey and includes a variety of sample members with special circumstances, such as incarcerated individuals, respondents in the military, other institutionalized persons, disabled persons, those with limited English proficiency, and so forth.

Incarcerated respondents. NLS respondents who are incarcerated constitute the largest group requiring special accommodations. The first challenge with incarcerated respondents is contacting them to schedule an interview. NLS interviewers must contact the prison administration to arrange for an interview, but the interviewers cannot legally reveal to the prison administration that the prisoner previously had participated in the survey without first obtaining the written, informed consent of the prisoner to reveal that information.

The following steps are used for obtaining prisoners' consent:

1. Prisoners are first sent a letter reminding them about their previous participation in a NORC survey, but, in case the mail is monitored by prison staff, the letter does not name the survey or BLS so as not to reveal the prisoner's participation. The letter encourages the prisoner to participate in the upcoming round of the survey. It explains that NORC staff needs to set up an interview through the prison administration but that NORC cannot tell the prison administration about the prisoner's participation without the prisoner's informed consent. The letter then asks the prisoner to request a consent form by signing and dating an enclosed form letter and mailing it to NORC in a pre-addressed, postage-paid envelope. The letter reminds the prisoner that the mail at the institution may be monitored and explains that the consent form that NORC will send the prisoner will state the prisoner's name and the name of the survey. The letter emphasizes that, by returning the enclosed form letter, prison management or staff may learn that the prisoner is a participant in the survey.
2. If the prisoner chooses to send the form letter to NORC, NORC then sends the prisoner a cover letter and a consent form that names the specific survey. The prisoner is asked to sign the consent form and mail it to NORC in a pre-addressed, postage-paid envelope. Once NORC has received the signed consent form, NORC staff can contact the prison to request permission to interview the prisoner and learn about any restrictions that the prison administration may impose.
3. If the prison administration permits an interview and a date and time have been scheduled for the interview, NORC mails another letter to the prisoner. This letter serves mulitiple purposes. First, it tells the prisoner when the interview will take place. Second, it informs the prisoner in writing that the interview very likely will be monitored by prison staff. That fact probably is obvious to any prisoner, but NLS staff feels that it is important to tell the prisoner in writing. In addition, the letter might also indicate that the respondent's incentive payment may be withheld, depending on a penal institution's policies.

Once all of these steps are complete, the prisoner finally can be interviewed, but the NLS program takes additional steps to minimize the risk that prisoners might reveal illegal or illicit behavior in the presence of prison staff during the course of the interview.

As described later in this section, such sensitive questions are asked in the self-administered portions of the NLSY97. During these portions of the survey, the typical protocol for a respondent who is not incarcerated involves the interviewer turning the laptop computer around to enable the respondent to read the questions to him or herself and enter the answers directly into the laptop computer without the interviewer knowing the responses. (In fact, the interviewer does not even know which questions the respondent answered). In some relatively low-security correctional facilities, such as some county jails and halfway houses, this protocol still would be possible. In higher security facilities, the prison administrators would not permit the prisoner to touch the computer, so the questions either would have to be read to the respondent or skipped altogether.

NLS program staff have identified the questions that could be considered even moderately sensitive or risky for the prisoner to answer out loud. Given this examination of these questions, the NLS program has adopted the following protocol for administering sensitive questions to prisoners:

1. At the very beginning of the interview, the interviewer will indicate in the survey instrument whether a respondent is in a correctional facility of any kind and, if so, whether the facility permits the prisoner to touch the laptop and enter responses to the self-administered questions. For Federal prisons, the interviewer assumes that the prisoner is not permitted to touch the laptop.
2. If the facility permits the prisoner to enter responses to the self-administered questions directly into the laptop, then the full set of questions, including all of the sensitive questions, would be administered.
3. If the facility does not permit the prisoner to enter responses directly into the laptop, or if the interview is conducted over the telephone rather than in person, all survey questions will be asked orally by the interviewer, but the instrument is programmed to skip sensitive questions in which the prisoner might be asked about illegal or illicit behavior.

Military respondents. NLS respondents who are in the military tend to be very cooperative and willing to participate in the surveys, but it sometimes can be difficult to locate and contact them, particularly if they are stationed outside the United States. It sometimes is necessary to seek the help of military or civilian staff in the Department of Defense to locate and contact military respondents, but NLS program staff first must obtain the military member's written, informed consent to reveal to Department of Defense staff that he or she previously had participated in the survey and is willing to be contacted to participate in future rounds of the survey.

Respondents with limited English proficiency. Some respondents lack fluency in English and are more comfortable using another language. It is not possible to accommodate all the different languages other than English that respondents might speak, but the NLSY97 survey staff historically have made special arrangements for respondents and their parents who speak Spanish, the most commonly spoken language other than English among respondents. NORC staff members translate advance letters and other informational materials into Spanish to enable respondents and the parents of minor respondents to provide their informed consent based on information that is written in the language that they understand the best. Survey questionnaires also have been translated into Spanish to ensure that the questions are administered consistently, rather than having Spanish-speaking interviewers translate the English-language questionnaire during the interview. The first six rounds of the NLSY97 included a Spanish version of the questionnaire, but because the number of respondents who speak only Spanish has continued to decline, it no longer is cost-effective to continue programming a computerized Spanish questionnaire. For that reason, Spanish questionnaires were not used after round 6. Advance letters and other informational materials still are available in Spanish.

Sensitive subjects. The NLSY97 has included questions on income and assets, religion, relationships with parents and other family members, sexual experiences, abortion, drug and alcohol use, criminal activities, homelessness, runaway episodes, and other topics that are potentially sensitive for respondents to discuss. Respondents are advised at the start of the interview that they can choose not to answer any questions that they prefer not to answer. During training, interviewers undergo exercises to teach them how to allay the concerns of respondents about answering sensitive questions and encourage them to respond. Interviewers are instructed not to coerce respondents into answering questions that they prefer not to answer.

Most questions in the NLSY97 are read to the respondent by an interviewer. The respondent then provides an answer, and the interviewer records that answer on a laptop computer. For especially sensitive questions, some respondents might be reluctant to answer truthfully--or at all--if they have to tell an interviewer their answers, even though interviewers can face criminal and civil penalties if they disclose the respondents' identities or answers to anyone not authorized to receive that information.

Obviously, it is important that respondents answer all questions truthfully, so the NLSY97 includes a self-administered portion of the questionnaire to reduce the potential reluctance to respond to sensitive questions. In this mode of data collection, the interviewer hands the laptop computer to the respondent and asks the respondent to read the questions and enter his or her responses with the keyboard. Sometimes respondents have literacy problems or disabilities that prevent them from reading the questions on the computer screen. For this reason, the interviewer also provides the respondent with a set of headphones that plug into the computer and enable the respondent to listen to a computer-generated recording of someone reading the questions. While the respondent completes this computer-assisted self-interview, the interviewer does not see the respondent's answers and, in fact, does not even see what questions the respondent is answering. Even in the self-administered portion of the survey, however, respondents still have the option not to respond to individual questions.

Guidelines for emailing sample members. At the end of each interview, respondents are asked to provide information that will help interviewers contact them during subsequent rounds of the surveys. In addition to the information collected about relatives, friends, or neighbors, interviewers ask all respondents (beginning in round 7) if they have an email address where they could be contacted in the future. The following guidelines were enacted to ensure confidentiality:

1. The name of the survey is not contained in the subject line or text of the email message. Some respondents may share the use of an email address with other household members, so the survey name is omitted from the message to prevent other household members from learning the specific name of the survey.
2. Email is sent from one main address. Field interviewers are not permitted to use their individual email accounts to contact respondents.

Respondents knowing respondents. One feature of the sample design in the NLSY97 is that there often are multiple respondents within the same original household, usually siblings but occasionally other relatives. It obviously is not possible in these cases to prevent family members from knowing that a relative is in the survey sample, but interviewers take steps to ensure that each respondent's answers remain private and are not revealed to other family members.

Respondent and parental consent. Because of the young ages of the NLSY97 cohort in the initial survey years, additional measures were taken to obtain informed consent from minors and their parents. NLSY97 consent procedures in rounds 1-5 included the following protocol:

1. For respondents age 17 and younger, parents were asked to complete a written Parental Permission to Interview Youth form prior to the youth completing the interview. Respondents themselves were asked to complete a written Youth Assent to Participate form prior to completing the interview. This protocol also applied to respondents whose interviews were completed by a proxy. (Proxy interviews are conducted in cases in which a disabled respondent can complete the interview only with the assistance of a parent, guardian, or other caretaker.) Respondents who were age 17 or younger and had attained independence or were considered "self-supportive" were not required to have a parent complete a permission form.
2. Whether or not a parent permission form was required, all youths, regardless of age, signed a consent form prior to completing the interview.

Through the completion of round 5 data collection in May 2002, a significant proportion of NLSY97 sample members required signed parental permission because they had not yet reached age 18 or established legal independence.

All NLSY97 sample members are now over the age of 18. Beginning in round 6, consent procedures changed to align with procedures of other large surveys or adults. Respondents now verbally consent prior to completing the interview. They read a consent statement similar to those used in previous rounds, but they no longer need to sign any documents. For incarcerated respondents and respondents whose interviews are completed by proxy, the NLS program continues to require signed consent as an additional safeguard against possible coercion.

Data handling

An important part of maintaining respondent confidentiality is the careful handling and storage of data. Steps taken by BLS, CHRR, and NORC to ensure the confidentiality of all respondents to the National Longitudinal Surveys include maintaining secure networks, restricting access to geographic variables, and topcoding income and asset values.

Network security. The data that are stored and handled at each NLS organization's site are done so with maximum security in place. During data collection, transmission, and storage, password protection and encryption are used to secure the data. Standard protocols for network security are followed at each organization's site. Detailed information about these arrangements is not provided to the public to prevent anyone from circumventing these safeguards.

Restricting access to geographic information. Geographic information about NLSY97 respondents is available only to researchers who are designated agents of BLS. These researchers must agree in writing to adhere to the BLS confidentiality policy, and their projects must further the mission of BLS and the NLS program to conduct sound, legitimate research in the social sciences. Applicants must provide a clear statement of their research methodology and objectives and explain how the geographic variables are necessary to meet those objectives. For more information about obtaining access to geographic variables in the NLSY97, see the NLS Web site at www.bls.gov/nls.

Topcoding of income and asset variables. Another step taken to ensure the confidentiality of NLSY97 respondents who have unusually high income and asset values is to "topcode" those values in the data set. Values that exceed a certain level are recoded so that they do not exceed the specified level. In each survey round, income and asset variables that include high values are identified for topcoding. For example, the wage and salary income variable is usually topcoded, but variables indicating the amount received from public assistance programs are not. Notes in the codebooks for topcoded income and asset variables provide more information about the exact calculations used to topcode each variable.