NLSY97_R18_OMB_justification_Part_A_Final_3.24.17

Information Collection Request for

The National Longitudinal Survey of Youth 1997

OMB # 1220-0157

Summary and Part A

Submitted by the Bureau of Labor Statistics

TABLE OF CONTENTS

Summary 1

Supporting Statement 3

A. Justification 3

1. Necessity for the Data Collection 3

2. Purpose of Survey and Data Collection Procedures 3

3. Improved Information Technology to Reduce Burden 5

4. Efforts to Identify Duplication 7

5. Involvement of Small Organizations 7

6. Consequences of Less Frequent Data Collection 7

7. Special Circumstances 10

8. Federal Register Notice and Consultations 10

9. Payment to Respondents 12

10. Confidentiality of Data 19

a. BLS Confidentiality Policy 19

b. Contractor Confidentiality Safeguards 20

11. Sensitive Questions 22

12. Estimation of Information Collection Burden 26

13. Cost Burden to Respondents or Record Keepers 28

14. Estimate of Cost to the Federal Government 28

15. Change in Burden 28

16. Plans and Time Schedule for Information Collection, Tabulation, and Publication 29

17. Reasons Not to Display OMB Expiration Date 29

18. Exceptions to “Certificate for Paperwork Reduction Act Submissions” 29

Summary

This package requests clearance for the pretest and main fielding of Round 18 of the National Longitudinal Survey of Youth 1997 (NLSY97). The main NLSY97 sample includes 8,984 respondents who were born in the years 1980 through 1984 and lived in the United States when the survey began in 1997. Sample selection was based on information provided during the first round of interviews. This cohort is a representative national sample of the target population of young adults. The sample includes an overrepresentation of blacks and Hispanics to facilitate statistically reliable analyses of these racial and ethnic groups. Appropriate weights are provided so that the sample components can be combined in a manner to aggregate to the overall U.S. population of the same ages.

The NLSY97 pretest sample includes 201 respondents who were born in the years 1979 through 1984 and lived in the United States at the time of the initial pretest fielding in 1996. Pretest data are used only for project operational and methodological information and are not released to individuals outside of the project team.

The survey is funded primarily by the U.S. Department of Labor. Additional funding has been provided in some years by the Departments of Health and Human Services, Education, Defense, and Justice, and the National Science Foundation. The Bureau of Labor Statistics has overall responsibility for the project. The BLS has contracted with the Center for Human Resource Research (CHRR) at the Ohio State University to conduct the survey. They are responsible for survey design, interviewing, data preparation, documentation, and the preparation of public-use data files.

The data collected in this survey are part of a larger effort that involves repeated interviews administered to a number of cohorts in the U.S. Many of the questions are identical or very similar to questions previously approved by OMB that have been asked in other cohorts of the National Longitudinal Surveys (NLS). Many of the questions in the NLSY97 also have been designed to reflect the changing nature of institutions and the different problems facing this group of young people. In this round, the NLSY97 will transition from interviews being conducted primarily in-person to being conducted primarily by telephone. The change in mode has resulted in some changes to the existing questionnaire. In addition, those data elements of a particularly sensitive nature and those not previously collected are justified in this document.

Supporting Statement

National Longitudinal Survey of Youth 1997 (NLSY97)

A Survey of Persons who were Ages 12 to 16 on December 31, 1996

Rationale, Objectives, and Analysis of Content

A. Justification

1. Necessity for the Information Collection

This statement covers the pretest and main fielding of Round 18 of the National Longitudinal Survey of Youth 1997 (NLSY97). The NLSY97 is a nationally representative sample of persons who were ages 12 to 16 on December 31, 1996. The Bureau of Labor Statistics (BLS) contracts with external organizations to interview these youths, to study how young people make the transition from full-time schooling to the establishment of their families and careers. Interviews were conducted on a yearly basis through Round 15; beginning with Round 16 they are interviewed on a biennial basis. The longitudinal focus of this survey requires information to be collected about the same individuals over many years in order to trace their education, training, work experience, fertility, income, and program participation.

The mission of the Department of Labor (DOL) is, among other things, to promote the development of the U.S. labor force and the efficiency of the U.S. labor market. The BLS contributes to this mission by gathering information about the labor force and labor market and disseminating it to policymakers and the public so that participants in those markets can make more informed and, thus more efficient, choices. The charge to the BLS to collect data related to the labor force is extremely broad, as reflected in Title 29 USC Section 1:

“The general design and duties of the Bureau of Labor Statistics shall be to acquire and diffuse among the people of the United States useful information on subjects connected with labor, in the most general and comprehensive sense of that word, and especially upon its relation to capital, the hours of labor, the earnings of laboring men and women, and the means of promoting their material, social, intellectual, and moral prosperity.”

The collection of these data contributes to the BLS mission by aiding in the understanding of labor market outcomes faced by individuals in the early stages of career and family development. See attachment 1 for Title 29 USC Sections 1 and 2.

2. Use of Information

The major purpose of the data collection is to examine the transition from school to the labor market and into adulthood. The study relates each respondent’s educational, family, and community background to his or her success in finding a job and establishing a career. During Round 1, the study included a testing component sponsored by the Department of Defense that assessed the aptitude and achievement of the youths in the study so that these factors can be related to career outcomes. This study, begun when most participants were in middle school or high school, has followed them as they enter college or training programs and join the labor force. Continued biennial interviews will allow researchers and policymakers to examine the transition from school to work. This study will help researchers and policymakers to identify the antecedents and causes for difficulties some youths experience in making the school-to-work transition. By comparing these data to similar data from previous NLS cohorts, researchers and policymakers will be able to identify and understand some of the dynamics of the labor market and whether and how the experiences of this cohort of young people differ from those of earlier cohorts.

The NLSY97 has several characteristics that distinguish it from other data sources and make it uniquely capable of meeting the goals described above. The first of these is the breadth and depth of the types of information that are being collected. It has become increasingly evident in recent years that a comprehensive analysis of the dynamics of labor force activity requires a theoretical framework that draws on several disciplines, particularly economics, sociology, and psychology. For example, the exploration of the determinants and consequences of the labor force behavior and experience of this cohort requires information about (1) the individual’s family background and ongoing demographic experiences; (2) the character of all aspects of the environment with which the individual interacts; (3) human capital inputs such as formal schooling and training; (4) a complete record of the individual’s work experiences; (5) the behaviors, attitudes, and experiences of family members, including spouses and children; and (6) a variety of social psychological measures, including attitudes toward specific and general work situations, personal feelings about the future, and perceptions of how much control one has over one’s environment.

A second major advantage of the NLSY97 is its longitudinal design. This design permits investigations of labor market dynamics that would not be possible with one-time surveys and allows directions of causation to be established with much greater confidence than cross-sectional analyses permit. Also, the considerable geographic and environment information available for each respondent for each survey year permits a more careful examination of the impact that local labor market conditions have on the employment, education, and family experiences of this cohort.

Third, the supplemental samples of blacks and Hispanics make possible more detailed statistical analyses of those groups than would otherwise be possible.

The NLSY97 is part of a broader group of surveys that are known as the BLS National Longitudinal Surveys program. In 1966, the first interviews were administered to persons representing two cohorts, Older Men ages 45-59 in 1966 and Young Men ages 14-24 in 1966. The sample of Mature Women ages 30-44 in 1967 was first interviewed in 1967. The last of the original four cohorts was the Young Women, who were ages 14-24 when first interviewed in 1968. The survey of Young Men was discontinued after the 1981 interview, and the last survey of the Older Men was conducted in 1990. The Young and Mature Women surveys were discontinued after the 2003 interviews. In 1979, the National Longitudinal Survey of Youth 1979 (NLSY79 – OMB Clearance Number 1220-0109), which includes persons who were ages 14–21 on December 31, 1978, began. The NLSY79 was conducted yearly from 1979 to 1994 and has been conducted every two years since 1994. One of the objectives of the National Longitudinal Surveys program is to examine how well the nation is able to incorporate young people into the labor market. These earlier surveys provide comparable data for the NLSY97.

The National Longitudinal Surveys are used by BLS and other government agencies to examine a wide range of labor market issues. The most recent BLS news release that examines NLSY97 data was published on April, 8 2016, and is available online at http://www.bls.gov/news.release/pdf/nlsyth.pdf. In addition to BLS publications, analyses have been conducted in recent years by other agencies of the Executive Branch, the Government Accountability Office, and the Congressional Budget Office. The surveys also are used extensively by researchers in a variety of academic fields. A comprehensive bibliography of journal articles, dissertations, and other research that have examined data from all National Longitudinal Surveys cohorts is available at http://www.nlsbibliography.org/.

More information about survey applications is provided in attachment 2.

3. Use of Electronic Collection Methods

The NLS program and its contractors have led the industry in survey automation and continue to use up-to-date methods for the NLSY97. This includes the continued use of computer-assisted interviewing (CAI) for the survey.

For Round 18, we propose a significant change in the mode of the data collection for the NLSY97. While historically an in-person survey, in Round 18, we propose to convert the NLSY97 to a predominantly telephone survey. We anticipate that approximately 75 percent of interviews will be completed by telephone, in contrast to a projected 26 percent in Round 17 and 15 percent in Round 16. Within the survey research literature, both unit non-response and item non-response are documented to be higher in telephone administration than in in-person administration. (Safir and Goldenberg (2008) “Mode Effects in a Survey of Consumer Expenditures,” Office of Survey Methods Research, Bureau of Labor Statistics retrieved from http://www.bls.gov/osmr/abstract/st/st080200.htm, Groves, Dillman, Eltinge and Little 2002 “Survey Nonresponse” New York: Wiley.)

Extensive minor edits have been made to adapt the Round 18 instrument for predominantly telephone administration, including the removal of references to showcards, introductory statements, reduction of self-administered content, and shortening of code frames. With the move to predominantly telephone interviewing, we anticipate somewhat longer timings. Average timings increased in Round 17 as the proportion of interviews conducted by telephone rose from 13 to 26 percent.

We note that questions about sexual activity, in addition to other sensitive items on topics such as drug use and criminal activity, have historically been asked using audio self-administered technology to anyone interviewed in person. With the conversion to predominantly phone work, sex questions and drug-use questions will be asked in interviewer-administered format. Criminal activity questions will still be asked using self-administration in the few in-person interviews conducted, but will primarily be asked by interviewers. Item non-response rates may increase in response to the change in mode. Sensitive items have consistently been found to be under-reported in interviewer-administered vs self-administered modes (Tourangeau and Yan, Psychological Bulletin 2007, Vol. 133, No. 5, 859–883). For drug use, for example, estimates of under-reporting range from 19 to 30 percent across multiple meta-analyses of experimental and quasi-experimental studies (Tourangeau and Yan 2007, Richmand, Kiesler, Weisband, and Drasgow 1999).

In contrast, face-to-face interviewing has been found to lead to under-reporting of sensitive items relative to telephone interviewing. Thus, sensitive items that had previously been interviewer-administered in-person but will now be interviewer-administered by telephone may experience decreases in item non-response. Income and other financial items would be the chief examples of such items. (de Leeuw E.D., van der Zouwen J. (1988). “Data quality in telephone and face to face surveys: a comparative metaanalysis.” In: Groves RM, Biemer PP, Lyberg LE, Massey JT, Nicholls WL II, Waksberg J, eds. Telephone Survey Methodology. New York: Wiley: 273:99).

Because of the change in mode and hence changes in surveying procedures in Round 18, BLS has expanded the time between pretest and main to permit for more intense review and a larger volume of edits than in prior rounds in which the questionnaire and process may have been more stable.

The vast majority of cases fielded in Round 18 will be fielded for telephone completion of the interview. We will attempt to complete these interviews by telephone, employing in-person interviews as a non-response follow-up strategy for cases that cannot be completed by telephone. For example, cases may be fielded in-person because the respondent cannot be located without in-person efforts, respondent refuses telephone interview, the respondent does not have a telephone, respondent is a prison case where only in-person interview is approved, or respondent has a physical or mental disability that prevents telephone administration.  In-person interviewing has been deployed similarly on the NLSY79 since it became primarily a telephone data collection effort in 2002.

In order to carefully meter the number of in-person interviews that will be conducted, we will employ practices that we have used for years on the NLSY97 to limit the number of telephone interviews conducted. In-person interviewing will not be available during the first few months of the field period; when available, field managers will issue prior approval for an in-person interview. This approval will be based on the facts of the case; hearing-related disability or inadequate telephone service will be readily approved for in-person interviewing. Respondent preferences for in-person interviews will first be addressed through further gaining cooperation activities and then only approved for in-person interviewing when interviewers and field manager believe that the only alternative to in-person interview is respondent non-interview. The metering of in-person interviews will be rather straightforward, since the opportunities for in-person interview will be very limited. The staffing model for Round 18 involves only limited geographic dispersion and does not allow for the greater hours per case required for travel

to in-person interviews. Other than interviews that cannot be completed by telephone (prison, disability, etc.), we expect that remaining in-person interviews will come about primarily because cases required in-person locating and gaining cooperation activities to secure an interview. We will not forego an opportunity for an in-person interview with a chronically non-cooperative individual if such an opportunity presents itself.

To better understand the effect of mode change on Round 18, we also plan a small ‘in-person first’ group. These randomly-selected cases from among historically less-cooperative cases would be fielded for in-person interviewing as has been the historical norm. We expect 10 percent of completed interviews to have been designated for ‘in-person first’ fielding. Including these, we anticipate that approximately 25 percent of all Round 18 interviews will be conducted in-person.

With the move to predominantly telephone interviewing on the NLSY97, we dramatically reduce our use of audio-self-administration for collection of sensitive items. Only questions related to experiences with the criminal justice system will be available for computer-assisted audio self-administration for a subset of respondents who are interviewed in-person. CAI interviews can reduce respondent burden and produce data that can be prepared for release and analysis faster and more accurately than is the case with pencil-and-paper interviews. Mode experiments on another NLS cohort showed that the same interview took 10 percent less time to administer using a computer. In addition, cases collected by phone ran approximately 5 minutes longer than those collected in-person in Round 17 of the NLSY97.

As planned at the outset of Round 17, the phone rate was 25 percent, up from 16 percent in Round 16. The contractor attempted to direct this increase and randomly selected cooperative cases that were approached for phone interviews. From the pool of cases that completed at least one interview Round 14 to Round 16 and completed less than half of those interviews by phone, 1/7^th were selected for the preferred phone activity with a target that 85 percent of completed interviews with this group would be by telephone. There were 6934 cases in the phone eligible pool, with 972 initially selected into the preferred phone group. Due to lower than expected take-up of phone interviews among the preferred phone group, an additional 364 cases meeting the original criteria were designated for preferred phone. Of the 1336 “preferred phone” cases, 766(57.3%) completed by phone and 405 (30.5%) completed in person. This “preferred phone” attempt helped us identify potential challenges in moving cases to telephone in Round 18 One issue was that delay in incentive payment on phone rather than cash at the door might be a discouraging factor. Offering electronic payment during Round 18 will be an attempt to mitigate this problem. Some respondents reported that they preferred an in-person interview and having the same interviewer each round. The contractor is considering ways to maximize consistency in interviewer, even with change in mode for most cases in Round 18. Some respondents reported concerns about losing the self-administered mode for sensitive questions. Some of the most sensitive of these have been dropped in Round 18. In addition, the wording has been changed so that those overhearing responses will not be able to infer the content of the questions from the respondent’s answers. Since text messaging is the preferred mode of communication for the age group of the NLSY97 cohort, field interviewers are issued cell phones by the contractor to increase their responsiveness to text messages and emails. This reduces the response time to the respondents and accommodates the respondent’s preferred mode of communication. These cell phones are returned to the contractor and wiped of all respondent information at the end of the field period. Additionally, in limited situations where the respondent does not have their own phone or the ability to borrow someone’s phone for the length of the interview, the project will send a pre-paid cell phone to the respondent with enough minutes to allow the respondent the opportunity to participate.

Finally, the use of computer-assisted recorded interviewing (CARI) reduces respondent burden by using recordings of the main interview for data quality assurance as a replacement for post-interview validation calls. During Round 18, we also will allow respondents to use web-based technology to schedule their appointments, providing high levels of convenience (and therefore lower burden) for respondents.

4. Efforts to Identify Duplication

We do not know of a national longitudinal survey that samples this age bracket and explores an equivalent breadth of substantive topics including labor market status and characteristics of jobs, education, training, aptitudes, health, fertility, marital history, income and assets, participation in government programs, attitudes, sexual activity, criminal and delinquent behavior, household environment, and military experiences. Data collection for the National Longitudinal Study of Adolescent Health (Add Health) is less frequent and addresses physical and social health-related behaviors rather than focusing on labor market experiences. The studies sponsored by the National Center for Education Statistics do not include the birth cohorts 1980 through 1984. The Children of the NLSY79, also part of the NLS program, spans the NLSY97 age range and touches on many of the same subjects, but does not yield nationally representative estimates for these birth cohorts. Further, the NLSY97 is a valuable part of the NLS program as a whole, and other surveys would not permit the kinds of cross-cohort analyses that are possible using the various cohorts of the NLS program.

The repeated collection of NLSY97 information permits consideration of employment, education, and family issues in ways not possible with any other available data set. The combination of (1) longitudinal data covering the time from adolescence; (2) a focus on youths and young adults; (3) national representation; (4) large minority samples; and (5) detailed availability of education, employment and training, demographic, health, child outcome, and social-psychological variables make this data set and its utility for social science policy research on youth issues unique.

5. Impact on Small Businesses

The NLSY97 is a survey of individuals in household and family units and therefore does not involve small organizations.

6. Consequences of Less Frequent Data Collection

Prior to Round 16, the NLSY97 was conducted annually, and that frequency has been essential for accurately capturing the educational, training, labor market, and household transitions that young people typically experience. Starting in Round 16 data collection changed to a biennial schedule. As NLSY97 respondents age, they experience fewer educational and labor market transitions which make biennial data collection more feasible.

In Round 16 (the first biennial interview), we continued to collect event histories since the date of last interview for employment, marriage, fertility, and schooling in order to maintain the detailed event history data we have collected since Round 1. However, in other domains the reference period for the interview questions was the last 12 months, which would not increase the length of the recall period for most respondents. As in prior rounds, we continued to use memory aides and bounding techniques in our interviewing to elicit accurate recall of events and dates. For Round 18, we do not propose any further changes to the instrument to accommodate biennial interviewing.

We did experience a decline in retention rate in Rounds 16 and 17, much like the decline experienced by the NLSY79 in its transition to biennial interviewing. The likely causes of the decline in Round 16 were manifold, including a pause in data collection related to a partial shutdown of the federal government, as well as an extremely severe winter that hampered travel, staff and respondent availability and communications. We are still investigating any bias resulting from this decline in retention rate. Tables 1a and 1b below provide an initial examination, comparing the proportion of NLSY97 respondents not completing the Round 16 and Round 13 interviews across the main sampling cells (gender*birth year*race/ethnicity). We see the same patterns in both rounds: men are more likely to be non-respondents than women, non-black/non-Hispanic respondents are more likely to be non-respondents than black and/or Hispanic respondents, and the two youngest birth cohorts (1983 and 1984) are somewhat less likely to non-respond. The differences in proportions are larger for every comparison in Round 16 than in Round 13.

Table 1a. Proportion of NLSY97 Respondents Not Completing Round 16 Interview, by Gender, Birth Year, and Race/Ethnicity
Row Labels	Black/AfrAm, Non- Hispanic	Latino/ Hispanic	Mixed Race	Non-Black, Non-Hispanic	Total	# of cases
Male	0.196	0.225	0.150	0.248	0.229	4599
1980	0.236	0.250	0.182	0.280	0.261	854
1981	0.175	0.171	0.333	0.237	0.207	947
1982	0.233	0.267	0.100	0.272	0.258	953
1983	0.152	0.191	0.000	0.233	0.203	934
1984	0.181	0.251	0.111	0.223	0.218	911
Female	0.114	0.162	0.233	0.220	0.180	4385
1980	0.148	0.189	0.200	0.222	0.192	837
1981	0.146	0.135	0.400	0.234	0.194	927
1982	0.078	0.141	0.333	0.209	0.162	888
1983	0.111	0.175	0.200	0.212	0.178	873
1984	0.080	0.179	0.000	0.222	0.173	860
Grand Total	0.155	0.195	0.193	0.235	0.205	8984
# of cases	2335	1901	83	4665	8984

Table 1b. Proportion of NLSY97 Respondents Not Completing Round 13 Interview, by Gender, Birth Year, and Race/Ethnicity

Row Labels	Black/AfrAm, Non- Hispanic	Latino/ Hispanic	Mixed Race	Non-Black, Non-Hispanic	Total
Male	0.168	0.162	0.100	0.189	0.177
1980	0.193	0.176	0.091	0.203	0.193
1981	0.167	0.195	0.167	0.222	0.202
1982	0.165	0.162	0.100	0.198	0.180
1983	0.161	0.134	0.000	0.176	0.163
1984	0.153	0.141	0.111	0.148	0.147
Female	0.087	0.120	0.116	0.175	0.139
1980	0.106	0.126	0.200	0.183	0.148
1981	0.106	0.125	0.200	0.192	0.157
1982	0.069	0.073	0.111	0.158	0.116
1983	0.098	0.119	0.100	0.166	0.137
1984	0.053	0.163	0.000	0.172	0.137
Grand Total	0.128	0.142	0.108	0.182	0.159

Note: Tables 1a and 1b comparisons do not take into account mortality.

Table 1c shows the retention rates experienced on the NLSY79 and the NLSY97 just before and after the transition from annual to biennial interviewing. Although there were important extenuating circumstances affecting the Round 16 NLSY97 fielding, the drop in response rates in the first and second biennial year was 2.3 and 2.1 percent for the NLSY79 and 3.1 and 0.2 percent for the NLSY97.

Table 1c. NLSY79 and NLSY97 Retention Rates Surrounding the Transition to Biennial Interviewing

	NLSY79 Retention Rate* (Year)	NLSY97 Retention Rate* (Year) (Round)
2 rounds prior to transition to biennial	92.1 (1993) (R15)	83.7 (2010) (R14)
Round prior to transition to biennial	91.1 (1994) (R16)	83.9 (2011) (R15)
First round after 2-year gap	88.8 (1996) (R17)	80.8 (2013) (R16)
2nd round after transition to biennial	86.7 (1998) (R18)	80.6 (2015) (R17)
3rd round after transition to biennial	84.3 (2000) (R19)	78.4 (2017) (R18)

* Retention rates exclude deceased and out of sample cases.

7. Special Circumstances

None of the listed special circumstances apply.

8. Federal Register Notice and Consultations

No comments were received as a result of the Federal Register notice published in 81 FR 55485 on August 19, 2016.

There have been numerous consultations regarding the NLSY97. In 1988, the National Science Foundation sponsored a conference to consider the future of the NLS. This conference consisted of representatives from a variety of academic, government and nonprofit research and policy organizations. The participants endorsed the notion of conducting a new youth survey. The NLSY97 incorporates many of the major recommendations that came out of that conference.

The NLS program also has a technical review committee that provides advice on interview content and long-term objectives. This group typically meets twice each year. Table 2 below shows the current members of the committee.

Table 2. National Longitudinal Surveys Technical Review Committee (2016)

Shawn Bushway Rockefeller College of Public Affairs and Policy Department of Public Administration and Policy University at Albany,	John Cawley Department of Policy Analysis & Management and Department of Economics Cornell University	Lingxin Hao Department of Sociology Johns Hopkins University
Judith Hellerstein Department of Economics University of Maryland	David Johnson Survey Research Center University of Michigan	Lance Lochner Department of Economics University of Western Social Science Centre
Alex Mas Industrial Relations Section Princeton University	Kathleen McGarry Department of Economics University of California, Los Angeles	Kristen Olson Department of Sociology University of Nebraska-Lincoln
Kelly Raley Department of Sociology University of Texas	Jesse Rothstein Richard & Rhoda Goldman School of Public Policy University of California, Berkeley	Rebecca Ryan Department of Psychology Georgetown University

9. Payment to Respondents

The NLSY97 is a long-term study in which the same respondents have been interviewed on an annual basis. Beginning with Round 16, the interval period between interviews moved to two years. Because minimizing sample attrition is critical to sustaining this type of longitudinal study, respondents in all prior rounds have been offered financial and in-kind incentives as a means of securing their long-term cooperation and slowing the decline in response rates.

Evidence shows that incentives can have positive effects on both respondent and interviewer behavior, and therefore are indispensable in reaching the project response rates. Incentives result in conversions among those least likely to participate and quicker cooperation among the more likely. Small increases or one-time bonuses can have a halo effect that results in future participation. Many interviewers find that having something to offer respondents, such as a monetary incentive, additional in-kind offering, or new conversion materials, allows them to open the dialogue with formerly reluctant respondents. Interviewers want a variety of options to respond to the particular needs, issues, and objections of the respondent. Fully loaded, interviewer costs come to around $36 / hour, and interviewers spend an average of almost 7 hours to obtain each interview. If incentives reduce this time by only 5%, nearly $125,000 in interviewer costs are saved.

In Round 18, we propose to move from predominantly in-person interviewing to predominantly phone interviewing. To support that transition, we will employ the Early Bird incentive mechanism that BLS and its contractors have developed and refined over the last several rounds of the NLSY79. Round 18 will also utilize several additional incentive strategies similar to those used in Rounds 13-17, including base, small in-kind, and missed round incentives. In addition, we propose two incentives that have proven effective with the NLSY79 cohort and other large scale studies.

For the purpose of this discussion, a brief summary of the respondent pools and incentive types is provided below.

Respondent Pools

For the Round 18 incentive plan, respondents are grouped into one of four major categories:

1. Respondents in the pretest sample,

2. Respondents in the main sample who completed Round 17,

3. Respondents in the main sample who missed Round 17 but completed at least once since Round 13,

4. Respondents in the main sample who have not completed any interviews during or since Round 13 and are not known to be deceased.

Table 3. Round 18 Respondent Pools

Incentive Types

The Round 18 incentive strategy includes a base incentive, an early bird incentive to respondents who initiate their interview without outreach by interviewing staff, small discretionary supplements, and payments for missed interviews. In addition, a final push incentive for individuals for whom the standard protocol has been unsuccessful in securing cooperation is proposed for Round 18.

Table 4. Round 18 Incentive Structure

Incentive Type	Pretest	R17 Completers: Completing in Early Bird telephone phase	R17 Completers: Completing after Early Bird telephone phase	Missed R17 (and up to 3 previous)	Missed all of the prior 5 rounds
Base	$50	$40	$40	$40	$40
Missed Round(s)	$20 per round	-	-	$20 per round (up to 3 rounds missed)	$20 per round (up to 3 rounds missed)
Early Bird	-	experimental variation: $10/$20/$35		experimental variation: $10/$20/$35	experimental variation: $10/$20/$35
Electronic payment bonus	-	$5	$5	$5	$5
Final push	-		$20	$20	$20
Long-term NIR* supplement	-	-	-		$100 (R18) $60 (R19)
Min	$50	$50	$40	$60	$200
Max	$110	$80	$65	$140	$240
Typical	$50	$60	$40	$60	$200
Token In-Kind	-	< $10 per respondent for up to 200 respondents may be spent on in-kind token gestures. The distribution of these respondents across the almost 9,000 respondents in these three columns is not known ex ante.
‘Talk to us’ prize	-	< $5 per respondent for up to 500 respondents may be spent on in-kind token gestures to get people who are consistent non-contacts to respond to our contacts (e.g., answer the phone, reply to a text, open a door) during a 2-3 week trial period when the response rate is at about 50 percent.

*This incentive will be offered to approximately 200 respondents. We expect fewer than 80 to take it up. For more information, see Table 6.

Base Incentive

A base incentive will be given to all respondents who complete the Round 18 interview. For the pretest sample, the base incentive will remain $50. For main study respondents, we retain the R17 base incentive of $40.

Missed round incentive

Payments for missed rounds are designed to encourage attriters to return to the study. For individuals who missed Round 17 (and possibly earlier consecutive rounds), we would offer an additional incentive for each consecutive missed round just prior to Round 18 if those respondents complete Round 18. We propose to pay $20 per round for one or more consecutive rounds (not to exceed 3). For example, a respondent having last completed the Round 15 interview and then again Round 18 would receive $40 ($20+$20) in missed round incentive, in addition to the $40 base fee for the main study. Rounds missed prior to Round 15 would not contribute to the fee amount for this respondent. The $20 per round is an increase over the $15 per missed round paid through R17, but remains lower than the $30 per round respondents would have received for participating in each missed round. Respondents would receive at most an additional $60 for prior missed rounds. Our objective is always to keep the missed-round payment lower than the payment for a timely interview; the three rounds prior to Round 18 all offered incentives of at least $30 per interview.

Early Bird bonus

In moving to predominantly telephone data collection, we propose to adopt the Early Bird incentive structure that has been well-honed on the NLSY79. Because we do not know the responsiveness of the NLSY97 sample to this type of incentive, and because our introduction of the Early Bird protocol is much less gradual than it was on the NLSY79, we propose an experiment to determine the relative costs and benefits of different premium amounts.

The premise of the Early Bird premium is to offer respondents the opportunity to call us rather than require us to expend resources to seek them out and gain their cooperation. We split the monetary benefits of the reduced effort by allocating a ‘premium’ to the respondent incentive above the base fee of $40. We propose to stratify the sample by prior round cooperativeness at the start of the round, then randomly assign cases by strata to three premium levels.

Type of case	Percent receiving $10 EB premium	Percent receiving $20 EB premium	Percent receiving $35 EB premium
Most cooperative through R17	33.3	33.3	33.3
Middle cooperative through R17	20	40	40
Least cooperative through R17	20	40	40

We propose to develop our measure of cooperativeness based on the experience of the Young Adult Early Bird implementation in Round 27, when that similarly aged-cohort was introduced to the Early Bird protocol. The Early Bird offer is communicated with an expiration date; respondents completing the Round 18 interview after their Early Bird expiration date will not receive an Early Bird premium.

Small in-kind incentive 

Sometimes, a small gift such as a cup of coffee or a toy for a child can facilitate an in-person interview.  We request the ability to give up to 200 respondents small in-kind gifts such as these.  We would spend no more than $10 on any respondent and average only $5.

Final Push Incentive

Starting after the first 12 weeks of the round, cases that have had at least 7 contact attempts or at least one refusal will be eligible for a final push incentive of $20.  In order to facilitate a timely close to the fielding, starting 6 months after the start of fielding, all cases will be eligible for this incentive.  In Round 17, 309 cases received this bonus, almost all of whom received it after the first six months of the field period. Once eligible, cases will receive the final push incentive upon interview completion unless respondents specifically request a lower amount.

Long-term NIR Supplement

We have an increasing number of respondents who have been out more than 5 rounds and who are completing at rates of less than 8 percent. We propose randomly selecting 1/4 of the non-deceased, non-blocked respondents who are out 5 or more rounds for a compound offer: an additional $100 if they complete the Round 18 interview, and an additional $60 if they complete the Round 19 interview. In conjunction with the other pieces of the current incentive schemes, their Round 18 incentive includes $60 for missed round NIR, $40 base fee, and possibly $20 final push and $5 electronic payment; along with the long-term NIR invitation the total possible incentive is $225. Their Round 19 incentive would otherwise be the base fee of $40, possibly plus Early Bird or final push bonuses. Selecting a fraction allows us to both evaluate the success of the offer and target extraordinary outreach efforts. If successful, we envision implementing a rotation across this difficult but valuable group of respondents. Even long-term NIRs not selected for this invitation will receive an increased incentive offer in Round 18 due to the increase of the missed round bonus from $15 to $20 for each of the first three missed rounds (thus from $45 extra to $60 extra). The NLSY97 has successfully implemented many incentive protocols in which amounts decrease relative to prior years without damage to respondent cooperation.

‘Talk to us’ Incentive

We request permission for a small trial of a textable gift (e.g., a Dunkin’ Donuts gift card or iTunes credit) to be offered mid-way through the field period to incentivize respondents whom we believe to have located, but who are not responding to any of our outreach. For a 2-3 week period after 50% of targeted completes have been achieved, we would offer a small group of no more than 500 respondents a small gift (< $5) if they will answer the phone, open the door, or write back to a text or e-mail. This test would be primarily to understand the logistical issues of such an incentive and our qualitative sense of whether or not it is worth a larger scale test in the future. The test is in response to our growing group of respondents whom we are not able to reach in any way although we believe that we have located them.

Incentive Costs

A table listing the total cost of incentives by respondent pool is provided below.

Table 6: Round 18 Incentive Costs by Respondent Pool

Incentive Type	Pretest	R17 Completers	Missed R17 (and up to 3 previous)		Missed all of the prior 5 rounds		Total
Sample size*	196	7054	979		800		9029
Expected R18 completes	150	6554	358		68		7130
Base	$7,500	$262,160	$14,320		$2,720		$286,700
Missed Round(s)	$810		$10,740		$4,080		$15,630
Early Bird		$65,098	$1,225		0		$66,323
Electronic payment bonus		$3,275	$180		$35		$3,490
Final push		$13,100	$720		$140		$13,960
Long-term NIR supplement					$6,800		$6,800
Token In-Kind		$1,500					$1,500
‘Talk to us’ prize	-	$2,500					$2,500
Total	$8,310	$343,633		$27,185		$13,775	$396,903

* Note: Sample sizes exclude deceased cases.

Electronic payments

In the NLSY97 interviews that are conducted in person, the respondent receives a cash incentive payment at the close of the interview. For the interviews done by phone, the payment process has been less streamlined. The case management system identifies a telephone interview as having occurred. Then either cash or a money order must be secured and mailed to the respondent, who receives it at some delay. In the case of loss or neglect by the respondent, the project has no ability to track or restore the payment. On some other OMB-approved data collections such as the National Survey of Early Care and Education for the Administration for Children and Families in the U.S. Department of Health and Human Services, incentives have been paid through electronic means such as e-mailed gift cards. The feasible technologies are rapidly increasing, with Google Wallet and iPhone 6 features as well as PayPal and retailers and financial institutions like Walmart, Amazon and Visa all offering tools for direct electronic payments without cash, check or credit cards. This type of payment system offers the opportunity to achieve project savings in the payment and administration of fees, improve the speed of payment, get more accurate contact information for respondents (such as mobile phone number or e-mail address to which payment can be made), and simplify the process if respondents need help tracking down their payments.

In Round 17 we introduced the use of PayPal as a payment option for respondents who completed the interview by phone. PayPal allowed the project to reduce administration fees, improve the speed of paying the respondent, confirm contact information (such as mobile phone number or e-mail address to which payment can be made), and allow the project to be proactive about resolving issues related to the transaction. The R17 experiment with PayPal was an overall success with a 17% overall selection of PayPal by telephone respondents as their choice of payment method when completed by phone. First, the project was able to improve service to respondents by expediting the turnover time for being paid for their participation, with mailing payment turnaround time being roughly 15 days whereas PayPal turnaround time is 48 hours or less. Second, the project was able to cut costs of labor, processing and delivery fees for the approximately 300 respondents who chose to be paid by via electronic payment method. Finally, one of the limitation of traditional methods of mailing is the inability to be proactive about resolving delivery or payment issues. With PayPal, the project and respondents experienced significantly fewer issues, and the project was able to proactively resolve any issues that did arise, actively minimizing respondent dissatisfaction.

As a result of the successes with electronic payment in R17 and with R18 transitioning to mostly phone administration, there is great potential for additional cost savings and continued improvements in respondent’s satisfaction with their payment receipt time. We request permission to continue the use of electronic payments on both the pretest and main sample to telephone respondents in R18. In addition, we propose expanding the options of electronic payments that could be offered to respondents to include other products akin to PayPal, namely Chase QuickPay. With more electronic options, the percentage of respondents using that form of payment could increase, helping in additional cost savings and in increasing respondent satisfaction with the transition to completion via phone. Payments would only be associated with the contractor and not the NLS program, and would be issued from the central office rather than individual interviewers. Telephone respondents would be given the option to choose electronic payment, but all respondents could continue to choose the traditional payment mechanisms if desired.

References

https://aspe.hhs.gov/report/studies-welfare-populations-data-collection-and-research-issues/incentives-panel-studies-0

http://www.oxfordbibliographies.com/view/document/obo-9780199756384/obo-9780199756384-0108.xml#obo-9780199756384-0108-bibItem-0079 search ‘participant incentives’

http://www.goeritz.net/SSCR2.pdf Lotteries in Longitudinal Surveys

http://www.surveypractice.org/index.php/SurveyPractice/article/view/106/html PSID between-wave incentive for contact info. Confirms PSID practice of $10 for update.

http://www.surveypractice.org/index.php/SurveyPractice/article/view/22/html_1 NSHAP finds use of graduated incentives did not affect overall incentive cost in later rounds.

10. Confidentiality of Data

a. BLS Confidentiality Policy

The information that NLSY97 respondents provide is protected by the Privacy Act of 1974 (DOL/BLS – 17 National Longitudinal Survey of Youth 1997 (67 FR 16818)) and the Confidential Information Protection and Statistical Efficiency Act of 2002 (CIPSEA). CIPSEA is shown in attachment 3.

The Confidential Information Protection and Statistical Efficiency Act of 2002 (CIPSEA) safeguards the confidentiality of individually identifiable information acquired under a pledge of confidentiality for exclusively statistical purposes by controlling access to, and uses made of, such information.  CIPSEA includes fines and penalties for any knowing and willful disclosure of individually identifiable information by an officer, employee, or agent of the BLS.

 

Based on this law, the BLS provides respondents with the following confidentiality pledge/informed consent statement:

 

“We want to reassure you that your confidentiality is protected by law. In accordance with the Confidential Information Protection and Statistical Efficiency Act of 2002, the Privacy Act, and other applicable Federal laws, the Bureau of Labor Statistics, its employees and agents, will, to the full extent permitted by law, use the information you provide for statistical purposes only, will hold your responses in confidence, and will not disclose them in identifiable form without your informed consent. All the employees who work on the survey at the Bureau of Labor Statistics and its contractors must sign a document agreeing to protect the confidentiality of your information. In fact, only a few people have access to information about your identity because they need that information to carry out their job duties.

Some of your answers will be made available to researchers at the Bureau of Labor Statistics and other government agencies, universities, and private research organizations through publicly available data files. These publicly available files contain no personal identifiers, such as names, addresses, Social Security numbers, and places of work, and exclude any information about the States, counties, metropolitan areas, and other, more detailed geographic locations in which survey participants live, making it much more difficult to figure out the identities of participants. Some researchers are granted special access to data files that include geographic information, but only after those researchers go through a thorough application process at the Bureau of Labor Statistics. Those authorized researchers must sign a written agreement making them official agents of the Bureau of Labor Statistics and requiring them to protect the confidentiality of survey participants. Those researchers are never provided with the personal identities of participants. The National Archives and Records Administration and the General Services Administration may receive copies of survey data and materials because those agencies are responsible for storing the Nation’s historical documents.

 

BLS policy on the confidential nature of respondent identifiable information (RII) states that “RII acquired or maintained by the BLS for exclusively statistical purposes and under a pledge of confidentiality shall be treated in a manner that ensures the information will be used only for statistical purposes and will be accessible only to authorized individuals with a need-to-know.”

By signing a BLS Agent Agreement, all authorized agents employed by the BLS, the prime contractor and associated subcontractors pledge to comply with the Privacy Act, CIPSEA, other applicable federal laws, and the BLS confidentiality policy. No interviewer or other staff member is allowed to see any case data until the BLS Agent Agreement, BLS Confidentiality Training certification, and Department of Labor Information Systems Security Awareness training certification are on file. Respondents will be provided a copy of the questions and answers shown in attachment 4 about uses of the data, confidentiality, and burden. These questions and answers will appear on the back of the letter that respondents will receive in advance of the Round 18 interviews. Attachment 4 also shows the combination advance letter and locating card for Round 18.

b. Contractor Confidentiality Safeguards

NLS contractors have safeguards to provide for the security of NLS data and the protection of the privacy of individuals in the sampled cohorts. These measures are used for the NLSY97 as well as the other NLS cohorts. Safeguards for the security of data include:

1. Like all federal systems, NLS and its contractors follow the National Institute of Standards and Technology (NIST) guidelines found in special publication 800-53 to ensure that appropriate security requirements and controls are applied to our system. This framework provides guidance, based on existing standards and best practices, for organizations to better understand, manage and reduce cybersecurity risk.

2. Storage of printed survey documents in locked space.

3. Protection of computer files against access by unauthorized individuals and groups. Procedures include using passwords, high-level “handshakes” across the network, data encryption, and fragmentation of data resources. As an example of fragmentation, should someone intercept data files over the network and defeat the encryption of these files, the meaning of the data files cannot be extracted except by referencing certain cross-walk tables that are neither transmitted nor stored on the interviewers’ laptops. Not only are questionnaire response data encrypted, but the entire contents of interviewers’ laptops are now encrypted. Interview data are frequently removed from laptops in the field so that only information that may be needed by the interviewer is retained.

4. Protection of computer files against access by unauthorized persons and groups. Especially sensitive files are secured via a series of passwords to restricted users. Access to files is strictly on a need-to-know basis. Passwords change every 90 days.

Protection of the privacy of individuals is accomplished through the following steps:

1. Oral permission for the interview is obtained from all respondents, after the interviewer ensures that the respondent has been provided with a copy of the appropriate BLS confidentiality information and understands that participation is voluntary.

2. Information identifying respondents is separated from the questionnaire and placed into a nonpublic database. Respondents are then linked to data through identification numbers.

3. After the final interview round, respondent identifier computer files will be destroyed.

4. The public-use version of the data, available on the Internet, masks data that are of sufficient specificity that individuals could theoretically be identified through some set of unique characteristics.

5. Other data files, which include variables on respondents’ State, county, metropolitan statistical area, zip code, and census tract of residence and certain other characteristics, are available only to researchers who undergo a review process established by BLS and sign an agreement with BLS that establishes specific requirements to protect respondent confidentiality. These agreements require that any results or information obtained as a result of research using the NLS data will be published only in summary or statistical form so that individuals who participated in the study cannot be identified. These confidential data are not available on the Internet.

6. Some questions of a more private nature are contained in self-administered portions of the survey for some respondents so their answers are concealed both from the interviewer and anyone in the household who might overhear the interview. Where self-administration will not occur, we have attempted to reframe questions so that individuals overhearing the respondent’s answers only would not be able to infer content.

7. In Round 18, the project team will continue several training and procedural changes that were begun in Round 11 to increase protection of respondent confidentiality. These include an enhanced focus on confidentiality and data security in training materials, clearer instructions in the Field Interviewer Manual on what field interviewers may or may not do when working cases, continued reminders of respondent confidentiality in field communications throughout the field period and formal separation procedures when interviewers complete their project assignments. Also, online and telephone respondent locating activities have been moved from geographically dispersed field managers to locating staff in central offices.

8. Date of birth will be verified at the beginning of the interview; thus, preventing the interviewer from starting the interview with the wrong person and then recognizing the error during the interview when the preloaded data were questioned. This check has been in place since Round 16.

11. Sensitive Questions

Continuing the practice of the last few rounds of the NLSY97, the Round 18 questionnaire includes a variety of items that permit the respondents to provide more qualitative information about themselves. Informal feedback from the interviewers and respondents indicates that this type of subjective data carries greater resonance with respondents as being informative about who they are, rather than the behavioral data that are the mainstay of the NLSY97 questionnaire. The items selected for self-description are all hypothesized in the research literature to be predictive of or correlated with labor market outcomes. There are several broad sets of questions in the NLSY97 data-collection instruments that may be considered sensitive. We address each of these categories separately below.

a.) Sexual Activity

Because puberty and the initiation of sexual activity occurred for many of the sample members during the first few survey rounds, this information has been carefully collected. Results from a number of different surveys, including early rounds of the NLSY97, indicate that a significant proportion of adolescents between the ages of 13 and 17 report that they are sexually active. It is vital that we continue to trace the progression of sexual activity in relation to the realization of educational and occupational goals and with respect to the promotion of good health practices. The level of sexual activity and contraceptive use are important indicators of how serious young people are about reaching higher levels of educational and occupational attainment, and there should be significant congruence between anticipated life goals, sexual activity, and its associated outcomes.

The survey will continue to collect information on the number of times male respondents have made a woman pregnant, as well as information on the live births from those pregnancies. Few studies have examined the linkages between early childbearing for men and subsequent education and employment outcomes in relation to family commitments. However, there is now some research indicating a modest connection between male labor supply and how many children they have. In an age where social responsibility is a salient public issue, longitudinal collection of data on the number of children men have fathered is essential to guarantee adequate representation of their children. At a minimum, collection of information about the offspring of male respondents is necessary for linking economic outlays of child support or lack of outlays of child support with potential determinants of both men’s and women’s labor supply behavior. Cross-sectional estimates of childbearing data for men can underestimate the number of children ever born to males.

The interviewer will administer these questions directly, although the questions have primarily been asked using the A-CASI technology prior to Round 18. Questions are worded so that respondents may answer without having to speak aloud sensitive terms such as contraception methods, which could be overheard in their immediate environment. No respondent will be pressured to answer the questions, and interviewers will be instructed to accept refusals without attempting to encourage response. Previous experience indicates that respondents usually recognize the importance of these questions, and field interviewers generally have not reported difficulties with these types of questions

b.) Anti-Social Behavior

The educational and labor force trajectory of individuals is strongly affected by their involvement in delinquent and risk-taking behaviors, criminal activity, and alcohol and drug use. There is widespread interest in collecting data on such behaviors. The challenge, of course, is to obtain accurate information on activities that are socially unacceptable or even illegal. Questions on these activities have been asked in the self-administered portions of the NLSY97, but will be asked by interviewers directly for 75 percent of interviews conducted by telephone in Round 18.

Crime and delinquency. The longitudinal collection of self-reported criminal behavior permits examination of the effects of these deviant behaviors on employment activity. This includes the ability to study whether there is a sustained pattern of criminal activities through the life cycle and how these patterns are related to employment difficulties. An additional area of study is the ways in which deviant behaviors may be causally associated with a disposition towards other aberrant behavior such as excessive alcohol and drug use. Use of both self-reports of behavior and of official disciplinary and court actions allows the NLSY97 to separate the effects of criminal activity that lead to an arrest or other legal action versus criminal activity that remains unpunished.

The design of the crime and delinquency module for the NLSY97 has taken great care to avoid weaknesses contained in other surveys’ instruments. As a result, unlike other surveys, the NLSY97 elicits information on a wider scope of activities and experiences related to crime. This includes questions concerning the type and frequency of criminal activity as well as self-reports about convictions, time served, and income received as a result of criminal activity.

Experiences with the correctional system. The Round 18 questionnaire continues to ask several questions on incarceration and parole that were added in Round 12. These include finer detail on parole and probation status as well as violations of that status, questions about experiences and services received while incarcerated, and questions about experiences and behaviors since release from incarceration. The questions on experiences and services received while incarcerated will be asked of currently incarcerated respondents and those who were incarcerated and released since the last interview. Respondents who were released from incarceration since the last interview also will be asked about their experiences and behaviors since they were released. For those respondents who are interviewed in person, these questions will continue to appear in the self-administered section, but will be asked by interviewers directly for the 75 percent of interviews conducted by telephone in Round 18.

Substance use. To quote a report based on data from the 1990 Youth Risk Behavior Surveillance System (U.S. Department of Health and Human Services), “Patterns of tobacco, alcohol and other drug use usually are established during youth, often persist into adulthood, contribute substantially to the leading causes of mortality and morbidity, and are associated with lower educational achievement and school dropout.” It is important that the NLSY97 continue to collect this information because of the potential impact of substance use on education and employment outcomes.

c.) Mental health

The literature linking mental health with various outcomes of interest to the NLSY97, including labor force participation, is fairly well-established. The Round 18 questionnaire includes questions on how many times the respondent has been treated for emotional, mental or psychiatric problems and how many times the respondent missed work or activities because of such problems in the past year.

d.) Religion

The NLSY97 has included questions about religious identification and attendance in most rounds. In Round 18, we are also including questions on religious attendance as have been asked in several prior rounds.

e.) Income, Assets, and Program Participation

The questionnaire asks all respondents about their income from wages, salaries, and other income received in the last calendar year. Other income is collected using a detailed list of income sources such as self-employment income, receipt of child support, interest or dividend payments, or income from rental properties. Respondents also are asked about their participation in government programs. Included are specific questions regarding a number of government assistance programs such as Unemployment Compensation, AFDC/TANF/ADC, and food stamps.

In addition to income, respondents are periodically asked about current asset holdings. Questions include the market value of any residence or business, whether the respondent paid property taxes in the previous year, and the amount owed on motor vehicles. Other questions ask about the respondent’s current checking and savings account balances, the value of various assets such as stocks or certificates of deposit, and the amount of any loans of at least $200 that the respondent received in the last calendar year. To reduce respondent burden, the asset questions are not asked of each respondent in every round. These questions have been asked in the first interview after the respondent turns 18, and the first interviews after the respondent’s 20^th, 25^th, 30^th, and 35^th birthdays. Because asset accumulation is slow at these young ages, this periodic collection is sufficient to capture changes in asset holdings. Round 17 (begun in Fall 2015) was the first time any respondents were asked to report their assets as of age 35. No substantive changes were necessary for the age 30 section to be made appropriate for administration at age 35.

Given the high fraction of household wealth associated with home ownership, the NLSY97 questionnaire collects home ownership status and (net) equity in the home from respondents each year that they are not scheduled for the full assets module. In Round 14 we added a question on present value of the home. We also ask respondents who owned a house or other dwelling previously and no longer live there about what happened to their house or dwelling. To allow for respondent-interviewer rapport to build before these potentially sensitive items are addressed, the housing value questions were moved from the (first) Household Information section to the Assets section, which occurs late in the interview. This question series also permits the respondent to report the loss of a house due to foreclosure.

The Rounds 1-13 NLSY97 questionnaires collected month-by-month participation status information for several government programs. These questions have been substantially reduced so that monthly data were no longer collected beginning with the Round 14 interview except from respondents who did not complete the Round 13 interview. The relatively low levels of participation reported did not seem to justify the respondent burden imposed by these questions. More detailed questions about receipt of income from government programs are now asked in the Income section with other sources of income.

f.) Financial Health

To get a better understanding of the financial well-being of the respondents, in Round 18 we continue questions we introduced in Round 10 about a respondent’s financial condition. We ask a set of questions to measure the financial distress of the respondents in the past 12 months. In particular, we ask whether respondents have been 60 days late in paying their mortgage or rent, and whether they have been pressured to pay bills by stores, creditors, or bill collectors. In addition, we ask respondents to pick the response that best describes their financial condition from the following list:

1. very comfortable and secure

2. able to make ends meet without much difficulty

3. occasionally have some difficulty making ends meet

4. tough to make ends meet but keeping your head above water

5. in over your head

The goal of these questions is to understand better the financial status of these respondents and how this status affects and is affected by their labor market activities.

Respondents are free to refuse to answer any survey question, including the sensitive questions described above. Our experience has been that participants recognize the importance of these questions and rarely refuse to answer.

12. Estimation of Information Collection Burden

The Round 18 field effort will seek to interview each respondent identified when the sample was selected in 1997. We will attempt to contact approximately 8,800 sample members who are not known to be deceased. BLS expects that interviews with approximately 6,950 of those sample members will be completed. The content of the interview will be similar to the interviews in Round 17, although extensive minor edits have been made to adapt the instrument for predominantly telephone administration, such as in references to showcards, introductory statements, reduction of self-administered content, and shortening of code frames. Based upon interview length in past rounds, we estimate the interview will require about 72 minutes.

Interview length will vary across respondents. For example, the core of the interview covers schooling and labor market experience. Naturally, respondents vary in the number of jobs they have held, the number of schools they have attended, and their experiences at work and at school. Our aim is to be comprehensive in the data we collect, and this leads to variation in the time required for the respondent to remember and relate the necessary information to the interviewer. For these reasons, the timing estimate is more accurate on average than for each individual case.

The estimated burden in Round 18 includes an allowance for attrition that takes place during the course of longitudinal surveys. To minimize the effects of attrition, we will seek to complete interviews with living respondents from Round 1 regardless of whether the sample member completed an interview in intervening rounds. We anticipate that an increased fraction of completed cases will be respondents who are returning to the survey after missing one or more rounds and therefore will have longer interviews than the typical respondent who has been in the survey every year.

With the move to predominantly telephone interviewing, we anticipate somewhat longer timings as has been the pattern in general on the NLSY97. Average timings did increase in Round 17 as the proportion of interviews conducted by telephone rose from 13 to 26 percent.

Household burden will vary with the number of in-scope sample members present, so households with three sample members may require almost four hours, and so forth. Although more than 1,800 households included multiple respondents at the time of the initial interview, by Round 18 many respondents have established their own households, and very few multiple respondent households remain. We are sensitive to the fact that the interviews in households with several sample members theoretically can pose interviewing problems, but that has not been our experience in previous rounds.

During the Round 18 field period, we will conduct validation interviews with no more than 2 percent of respondents to ascertain that the interview took place as the interviewer reported and to assess the quality of the data collected. These cases will be selected purposefully, based on data and assessments by survey management that indicate a field interviewer’s caseload merits further scrutiny. These validation interviews average about four minutes each and will be conducted only for the main fielding. Reasons for validation interview included missing audio on CARI files and High CARI recording refusal rate.

Based on our experience of recording segments of the main interview in Rounds 11 through 13 and use of these segments for data quality assurance in Rounds 14 through 17, respondents will be asked to provide their consent for the recording of segments of the main interview in Round 18.

“My computer is equipped to record this interview for quality control, research, testing and training purposes. As always your confidentiality is protected by Federal law and the policies of the Bureau of Labor Statistics and (Name of Contractor). May I continue with the recording?”

YES

NO

If the respondent objects to the recording of the interview, the interviewer will confirm to the respondent that the interview will not be recorded and then proceed with the interview.

Recording these interviews will enable BLS to improve data quality while reducing respondent burden. These recordings help verify that the interviews actually took place and that the interviewers did not fabricate the data. Recordings also help to ensure that interviewers are reading the questions exactly as worded and entering the responses properly. In addition, they help to identify parts of the interview that might be difficult or causing misunderstanding for interviewers or respondents. Our experiences with OMB-approved interview recordings since Round 10 indicate that respondents are generally quite willing to consent to be recorded, and the quality of recordings is sufficient for meaningful data-quality assurance.

In addition, we employ statistical review of questionnaire data to investigate interviewer performance when recordings are not available (such as in the case of a recording refusal or microphone issues). Statistical review includes, but is not limited to, looking at cases with unusual lengths of interview (too short or too long), the percent of recording refusal, industry and occupation verbatim response quality, response outliers, frequency of comments made by field interviewers in questionnaire, etc. When statistical anomalies are identified in interviewer performance, these interviewers’ cases are then subjected to additional recording or data review.

These two methods of data quality review are conducted by project staff and reduce respondent burden as verification of the interview and its quality can be done without further outreach to the respondent. This methodology also reduces cost as statistical review can be done on an interviewer-level basis to evaluate the particular field interviewer’s data quality.

Table 7: No. of Respondents and Average Response Time, NLSY97 Round 18

Form	Total Respondents	Frequency	Total Responses	Average Time per Response	Estimated Total Burden
NLSY97 Pretest April/May 2017	150	One-time	150	72 minutes	180 hours
Main NLSY97: September 2017-May 2018	6980	One-time	6980	72 minutes	8376 hours
Validation interview: October 2017 – June 2018	139	One-time	139	4 minutes	9.3 hours
TOTALS*	7,130	—	7269	—	8565.3 hours

* The difference between the total number of respondents and the total number of responses reflects the fact that about 6,980 are expected to complete the main interview. In addition, about 139 respondents will be interviewed twice, once in the main survey and a second time in the 4-minute validation interview.

The total response burden for the survey is 8,565.3 hours. The total annualized cost to respondents, based on burden hours and the Federal minimum wage of $7.25 per hour, is $62,098.43.

13. Cost Burden to Respondents or Record Keepers

Respondents for this survey will not incur any capital and start-up costs; respondents will not incur any operation and maintenance or purchase of service costs.

14. Estimate of Cost to the Federal Government

The total estimated cost for the main NLSY97 in Round 18 is $11 million.  This cost includes survey management, questionnaire design, instrument development, pretest and main data collection including incentive payments, cleaning and preparation of data files for users, and services to users of the data files.

15. Change in Respondent Burden

In this round, the NLSY97 will transition from interviews being conducted primarily in-person to being conducted primarily by telephone. With the move to predominantly telephone interviewing, this request contains slightly longer timings. The burden for Round 18 exceeds the projected burden from Round 17 by 883 hours. The difference comes from an increased estimated interview length due to telephone rather than in-person administration of the interview. A good portion of the 11 minute increase in the interview length comes from interviewer administration of components that were previously self-administered by respondents. The change in the data collection mode also resulted in extensive minor edits to the questionnaire in order to adapt the Round 18 instrument for predominantly telephone administration.

16. Plans and Time Schedule for Information Collection, Tabulation, and Publication

The following is the planned schedule for the data collection for Round 18.

Questionnaire Development September 2015 –May 2016

Respondent Materials Development January 2017 – September 2017

Pretest Data Collection April 2017 – May 2017

Main Data Collection September 2017 – May 2018

Data Processing June 2017 – October 2018

Publication of BLS News Release September 2019
Release of Public-Use Main Data Files November 2019

17. Request Not to Display OMB Expiration Date

The OMB number and expiration date will be provided in the advance letter.

18. Exceptions to “Certificate for Paperwork Reduction Act Submissions”

We do not have any exceptions to the “Certificate for Paperwork Reduction Act Submissions” statement.