Technical Note
The estimates in this release were obtained using data from the first 29 rounds of the
National Longitudinal Survey of Youth 1979 (NLSY79). This survey is conducted by the
National Opinion Research Center at the University of Chicago and the Center for Human
Resource Research at The Ohio State University under the direction and sponsorship of
the U.S. Department of Labor’s Bureau of Labor Statistics.
Sample
The NLSY79 is a nationally representative sample of 12,686 young men and women who were
14 to 22 years of age when first surveyed in 1979. This survey sample was initially
composed of three subsamples:
--A cross-sectional sample of 6,111 youths that was designed to represent the
noninstitutionalized, civilian population of young people living in the U.S. in
1979 and born between Jan. 1, 1957, and Dec. 31, 1964.
--A supplemental sample of 5,295 youths designed to oversample noninstitutionalized,
civilian Black, Hispanic or Latino, and economically disadvantaged nonblack,
non-Hispanic or Latino youths living in the U.S. in 1979 and born between Jan. 1,
1957, and Dec. 31, 1964.
--A military sample of 1,280 youths born between Jan. 1, 1957, and Dec. 31, 1961, and
enlisted in the Army, Air Force, Navy, or Marine Corps as of September 30, 1978.
In 1985, the military sample was discontinued, and, in 1991, the economically disadvantaged
nonblack, non-Hispanic youths were dropped from the supplemental sample. As a result,
the NLSY79 sample now includes 9,964 individuals from the cross-sectional sample and
the Black and Hispanic or Latino supplemental samples. (This sample size is not adjusted
for sample members who have died.)
Individuals were surveyed annually from 1979 to 1994 and biennially since 1994. In 2020-21,
6,535 individuals responded to the survey, for a retention rate of 66 percent (representing
a 74 percent response rate among those sample members who are still living). Only these
individuals are included in the estimates in this release. All results are weighted using
the 2020-21 survey weights that correct for the oversampling, interview nonresponse, and
permanent attrition from the survey. When weighted, the estimates represent all persons
born in the years 1957 to 1964 and living in the U.S. when the survey began in 1979. Not
represented by the survey are U.S. immigrants who were born from 1957 to 1964 and moved
to the U.S. after 1979.
Work history data
The total number of jobs that people hold during their work life is an easy concept to
understand but a difficult one to measure. Reliable estimates require a survey that
interviews the same people over the course of their entire work life and also keeps track
of all the jobs they ever held. The NLSY79 tracks the number of jobs that people have held,
but most of the respondents in this survey are still working and have more years of work
life ahead of them. While the survey has already tracked 41 years of work history, as
the cohort continues to age, more complete information will become available.
A unique feature of the NLSY79 is that it collects the beginning and ending dates of all
jobs held by a respondent so that a longitudinal history can be constructed of each
respondent’s work experiences. The NLSY79 work history data provide a week-by-week work
record of each respondent from Jan. 1, 1978, through the most recent survey date. These
data contain information on the respondent’s labor force status each week, the usual
hours worked per week at all jobs, and earnings for all jobs. If a respondent worked at
more than one job in any week, hours and earnings are obtained for additional jobs. When
a respondent who missed one or more consecutive survey rounds is interviewed again, he
or she is asked to provide information about all time since the last interview.
Interaction between time and age in a longitudinal survey
Because the NLSY79 is a longitudinal survey, meaning the same people are surveyed over time,
the ages of the respondents change with each survey round. It is important to keep in mind
this inherent link between the calendar years and the ages of the respondents. For example,
table 5 reports earnings growth from age 45 to age 56. The youngest respondents in the
sample (birth year 1964) were these ages during 2009-20, whereas the oldest respondents
(birth year 1957) were these ages during 2002-13.
Although participants in the NLSY79 were ages 55 to 64 during the 2020-21 interviews, this
release covers only the period while the respondents were ages 18 to 56. The reason for not
including older ages is that the sample sizes were still too small to provide statistically
reliable estimates for age groups older than 56. As the NLSY79 continues to be administered
and the respondents age, subsequent rounds of the survey will enable analyses to be conducted
for older age groups.
As with age, the educational attainment of individuals may change from year to year. In
the tables and analysis presented in this report, educational attainment is defined as of
the 2020-21 survey. This definition is used even when data on age and educational attainment
are presented together. For example, table 1 reports the number of jobs held during different
age categories. Suppose that a respondent had completed a bachelor’s degree at age 56. That
respondent would be included in the “Bachelor’s degree and higher” educational category in
all age categories shown on the table, even though he or she did not have a bachelor’s degree
at any point from age 18 to age 55.
Definitions
Job. A job is defined as an uninterrupted period of work with a particular employer. Jobs
are therefore employer-based, not position-based. If a respondent indicates that he or she
left a job but in a subsequent survey returned to the same job, it is counted as a new job.
For example, if an individual worked in a retail establishment during the summer, quit at
the end of summer, and then resumed working for the same employer the following spring, this
sequence would count as two jobs, rather than one. For self-employed workers, each “new”
job is defined by the individuals themselves.
Unemployment. If respondents indicate a gap between employers, they are asked how many of
those weeks they spent searching for employment or on layoff. For that number of weeks, they
are considered unemployed. For the remaining weeks, they are coded as not in the labor force.
No probing for intensity of job search is done.
Usual earnings. Respondents can report earnings over any time frame (hour, day, week, month,
year). For those who do not report an hourly wage, one is constructed using usual hours
worked over that time frame. Wages greater than $100 per hour and less than $1 per hour
(in 1979 dollars) were not included in the analysis of earnings growth because the reported
earnings levels were almost certainly in error. For the same reason, individuals who had
inflation-adjusted earnings growth greater than 100 percent were not included in the analysis.
Marital Status. Marital status is determined for the interview date in which the respondent
first turned a specific age (56, for example). If not interviewed at that age, then marital
status is determined for the prior age.
Health limits kind or amount of work respondent can do: Health limit is determined for the
interview date in which the respondent first turned a specific age (56, for example). If
not interviewed at that age, then health limit is determined for the prior age.
Race and ethnicity groups. In this release, the findings are reported for non-Hispanic White
persons, non-Hispanic Black persons, and Hispanic or Latino persons. These three groups are
mutually exclusive but not exhaustive. Other race groups, which are included in the overall
totals, are not shown separately because their representation in the survey sample is not
sufficiently large to provide statistically reliable estimates. In other BLS publications,
estimates usually are published for White persons, Black persons, and Hispanic or Latino
persons, but these groups are not mutually exclusive. The term Hispanic or Latino is
considered to be an ethnicity group, and Hispanic or Latino persons can be of any race.
Most other BLS publications include Hispanic or Latino persons in the White and Black race
groups in addition to the Hispanic or Latino ethnicity group.
Information in this release will be made available to sensory impaired individuals upon
request. If you are deaf, hard of hearing, or have a speech disability, please dial 7-1-1
to access telecommunications relay services.
United States Department of Labor
