Search BTN:

• HOME
• ARCHIVE
  • Archive By Date
  • Archive By Program
  • Archive By Publication
  • Archive By Theme
• ABOUT
• SUBSCRIBE

How did employment change during the COVID-19 pandemic? Evidence from a new BLS survey supplement

By Alison Aughinbaugh and Donna S. Rothstein

In the spring of 2020, the COVID-19 pandemic triggered job loss in the labor market on a scale not seen since the Great Depression. A year later, the economic situation had improved. Approximately 60 percent of jobs lost had returned, but employment was still down compared to pre pandemic levels.¹In an effort to understand how the COVID-19 pandemic affected labor market experience, the U.S. Bureau of Labor Statistics (BLS) National Longitudinal Survey of Youth 1997 (NLSY97) fielded a short supplemental survey to gather information from its sample members on work and working conditions, among other topics. Data from this new survey sheds light on the work experiences of Americans during the COVID-19 pandemic.

This Beyond the Numbers article examines some of the findings from this new supplement related to work during the pandemic, measured over the past week and past year, including nonwork due to the pandemic, changes in employment, telework, and frequency of contact with others during in-person work.

Description of the data

The NLSY97 is a nationally representative sample of 8,984 men and women born from 1980 to 1984 and living in the United States at the time of the initial survey in 1997. Respondents were interviewed annually from 1997 to 2011 and biennially since then. The NLSY97 collects information on a broad range of topics, with employment history the backbone of the survey. In addition to these regular interviews, a supplementary data collection of the NLSY97 sample members was undertaken from February through May 2021 to clarify the effects of the COVID-19 pandemic on this cohort. The supplement collected point-in-time information on employment, working conditions, children’s schooling, a self-report of general health, and a depression assessment. In addition, the participants were asked about labor market changes over the previous 12 months (losing a job, starting a new job, and changes in their pay) that they had experienced because of the pandemic. The NLSY97 sample members are between 36 and 41 years old at the time of the 2021 supplement.

Employment in the last week

Table 1 shows employment status during the week prior to the supplemental interview, which took place February 2021 through May 2021 for the NLSY97 cohort, by selected characteristics. As a whole, about 77 percent of the cohort was working for pay or profit during the last week, with men more likely to work than women (81 percent compared with 73 percent). In addition, non-Black, non-Hispanic men were more likely to work than non-Black, non-Hispanic women (83 percent compared with 75 percent) and Hispanic men were more likely to work than Hispanic women (82 percent compared with 66 percent); in contrast, Black, non-Hispanic men and women were about equally likely to work in the prior week (a little over 70 percent).

Table 1. Percentage of people born in the years 1980 to 1984 who worked for pay last week (February through May 2021)

Characteristics	All	Men	Women
Total	77.1	81.1	72.9
Race and ethnicity
Non-black, non-Hispanic	79.1	83.4	74.7
Black, non-Hispanic	70.4	70.5	70.4
Hispanic or Latino	74.4	82.1	65.5
Education
Less than a high school diploma	50.1	56.7	43.2
GED	60.1	64.9	53.1
High school graduates, no college	70.5	75.8	63.7
Some college or associate degree	75.9	80	71.4
Bachelor's degree and higher	88.8	93.9	84.5
AFQT percentile test score
Less than 25 percent	62.6	68.5	56.2
25 percent to less than 50 percent	76.9	81	72.9
50 percent to less than 75 percent	81.7	85.2	78.5
75 percent or higher	87	90.3	83.2
Spouse or partner lives in household
Yes	80.3	87	73.3
No	70.2	68.2	72.4
Child < 18 lives in household
Yes	78.2	86.5	71.3
No	75.1	73.4	78
Self-rated health
Excellent/very good	82.6	86.9	78
Good	77.3	80.7	73.7
Fair/poor	60.2	62.7	58
Note: Data are weighted. Source: U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1997.

The percentage of men and women working in the last week rises with education level. About 50 percent of those with less than a high school diploma reported working in the previous week compared with almost 89 percent of those with a bachelor’s degree or more. When examined separately, the percentage of men and women that were working increases with education. However, within each education level, men were also more likely to work than women. Men and women with higher percentile scores on the Armed Forces Qualification Test (AFQT) were more likely to work in the last week than men and women with lower percentile scores.² Men were more likely to work in the last week if a child lived in the household whereas women were less likely to work in the last week if a child lived in the household. Finally, those with poorer self-reported health were much less likely to work in the previous week than those who rate their health as good or excellent/very good.

Respondents who were not working in the last week (a small percentage of whom were temporarily absent) were asked whether it was because of the COVID-19 pandemic. Of those not working, almost 41 percent attributed their non-work to the COVID-19 pandemic. Chart 1 depicts the distribution by race and ethnicity, as well as education. Black, non-Hispanic men and women who did not work in the previous week were much more likely to report that their main reason for not working was because of the COVID-19 pandemic (56 percent) compared with non-Black, non-Hispanic (35 percent), or Hispanic (42 percent) men and women. Those with a GED or less than a high school diploma were much more likely to report that their main reason for not working was because of the COVID-19 pandemic than those with higher levels of education.

Hours of work and work from home among those working in the last week

Among those working in the last week, BLS collected information on hours of work and the number of those hours worked from home. Table 2 displays hours of work during the last week (February 2021 through May 2021) by selected characteristics. It also shows how many of those hours worked were at home (none, some, or all). Of those working, 84 percent worked at least 35 hours per week, with 53 percent having no work from home, 21 having at least partial work from home, and 25 percent working all hours from home. Men who worked in the prior week were more likely to work full time than women (89 percent compared with 78 percent). About 22 percent of men worked all hours from home compared with 30 percent of women.

Table 2. Percentage distribution of hours and hours worked from home for people born in the years 1980 to 1984 who worked last week (February through May 2021)

Characteristics	All					Men					Women
	Hours worked at all jobs		Work from home			Hours worked at all jobs		Work from home			Hours worked at all jobs		Work from home
	Less than 35	35 or more	None	Some	All	Less than 35	35 or more	None	Some	All	Less than 35	35 or more	None	Some	All
Total	15.9	84.1	53.3	21.3	25.4	10.7	89.3	56.8	21.4	21.8	22.1	77.9	49.2	21.1	29.7
Race and ethnicity
Non-Black, non-Hispanic	15.5	84.5	51.6	22.2	26.2	9.5	90.5	54.7	22.1	23.2	22.4	77.6	48.1	22.3	29.6
Black, non-Hispanic	17	83	62	15.9	22.1	13.9	86.1	68.1	17.3	14.6	20.3	79.7	55.5	14.5	30
Hispanic or Latino	17	83	54.4	21.5	24.2	13.5	86.5	58.6	20.7	20.7	22	78	48.2	22.7	29.2
Education
Less than a high school diploma	32.2	67.8	82.3	9.7	8	25	75	84.3	8.9	6.7	42.4	57.6	79.6	10.7	9.6
GED	24.9	75.1	79	12.7	8.3	18.3	81.7	79.3	13.2	7.5	36.8	63.2	78.6	11.8	9.6
High school graduates, no college	16.1	83.9	76.2	13.2	10.6	9.4	90.6	79.6	13.9	6.5	26.6	73.4	71	12.2	16.9
Some college or associate degree	17.1	82.9	61.9	18.4	19.6	11.8	88.2	64.2	18.2	17.6	23.7	76.3	59.1	18.7	22.2
Bachelor's degree and higher	11.9	88.1	32.1	28.5	39.4	7.3	92.7	32.9	30.2	36.9	16.2	83.8	31.3	26.8	41.8
AFQT percentile test score
Less than 25 percent	19.9	80.1	75	12.4	12.6	13.9	86.1	77.7	14.6	7.8	28.1	71.9	71.5	9.4	19.1
25 percent to less than 50 percent	15.3	84.7	63.4	17	19.6	10.5	89.5	67.6	18.1	14.3	20.7	79.3	58.8	15.8	25.4
50 percent to less than 75 percent	14.5	85.5	50.7	22.3	27	8.3	91.7	56.6	19.9	23.5	21	79	44.6	24.7	30.6
75 percent or higher	15.8	84.2	33.6	28.1	38.3	11.9	88.1	32.9	29.9	37.2	20.5	79.5	34.5	25.9	39.6
Spouse or partner lives in household
Yes	14.5	85.5	49.1	23.7	27.3	8.7	91.3	52.5	24.2	23.3	21.9	78.1	44.8	22.9	32.3
No	19.6	80.4	64.1	15.4	20.5	16.5	83.5	69.4	13.4	17.2	22.8	77.2	58.9	17.4	23.7
Child < 18 lives in household
Yes	15.5	84.5	52.1	22.9	25	8.3	91.7	54.7	24.3	21	22.8	77.2	49.4	21.5	29.1
No	17	83	55.9	17.7	26.5	15.1	84.9	60.4	16.1	23.4	20.2	79.8	48.5	20.1	31.4
Self-rated health
Excellent/very good	14.8	85.2	50.7	22.8	26.5	9.6	90.4	54.7	22.9	22.5	21.1	78.9	45.8	22.7	31.5
Good	16.4	83.6	56	20.4	23.6	10.2	89.8	59.8	20.5	19.7	23.7	76.3	51.5	20.4	28.1
Fair/poor	19.5	80.5	56	17.4	26.6	17.5	82.5	56.8	17.6	25.6	21.6	78.4	55.2	17.2	27.5
Note: Data are weighted. Source: U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1997.

Within race and ethnic groups, men were more likely to work full time than women and were more likely than women to have no work from home. Working individuals with a higher level of education were more likely to work 35 or more hours per week and also less likely to have no work from home than those with lower levels of education. For example, compared with individuals with less than a high school diploma, individuals with a bachelor’s degree and higher were about 20 percentage points more likely to have worked at least 35 hours per week and 50 percentage points less likely to have no telework. Similar results by education hold for men and women separately.

Table 2 shows large differentials in the likelihood of working from home by AFQT percentile scores overall and separately by sex. As an example, about 75 percent of working individuals with AFQT test scores less than the 25th percentile had no work from home in the last week compared with almost 34 percent of those with test scores in the 75th percentile or above. About 37 percent of working men with AFQT test scores in the top quartile did full telework compared with only 8 percent of men with AFQT test scores in the lowest quartile. Women with the highest quartile of AFQT scores were over twice as likely to fully work from home than women in the bottom quartile of AFQT scores (40 percent compared with 19 percent).

Those who worked at least some hours outside of their home were asked how frequently the work outside of their home required them to be in close contact (i.e., within 6 feet) with coworkers, customers, or other people not living in their household. Chart 2 depicts the percentage of those working outside of the home who responded they were required to be in close contact most or all of the time (as opposed to some/rarely/not at all), by gender and education. Women who worked outside the home were over 10 percentage points more likely than men to be required by their work to be in close contact with others. Men and women with a bachelor’s degree and higher who worked outside of their home were much less likely to be in close contact with others than those with lower levels of education.

In-person and remote schooling and parental employment in the last week

The COVID-19 pandemic prompted many schools to move from in-person classes to remote learning. By late winter and spring of 2021 more schools were at least partially in-person, but distance learning remained in effect in many schools.³ In table 3, the sample is limited to individuals who have children under age 18 in their household enrolled in K-12 public, private, or other schools.⁴ The table shows the percentage of individuals with children who attended in person and remote-classes in the last week.

Table 3. Percentage of people born in the years 1980 to 1984 with children under 18 in the household who attended in-person and remote schooling last week (February through May 2021)

Characteristics	All		Men		Women
	In-person classes	Remote-learning classes	In-person classes	Remote-learning classes	In-person classes	Remote-learning classes
Total	67.9	65.8	71	62	65.6	68.7
Race and ethnicity
Non-black, non-Hispanic	75.2	61.5	77.7	57.3	73.2	64.8
Black, non-Hispanic	47.4	76.7	51.5	73.5	44.5	79
Hispanic or Latino	52.4	74.6	56.3	71.3	49.2	77.3
Education
Less than a high school diploma	58.1	74.6	69.4	70.4	51.5	77
GED	63.3	66.6	68.7	60.7	56.8	73.7
High school graduates, no college	66.7	66.2	73.5	58.3	60.3	73.6
Some college or associate degree	66.5	69.9	68.9	66.3	64.8	72.5
Bachelor's degree and higher	72.3	60.8	72.7	60.3	72	61.2
AFQT percentile test score
Less than 25 percent	56	68.6	60.5	61.2	52.7	74
25 percent to less than 50 percent	69.4	67.1	73.9	60.5	66.4	71.5
50 percent to less than 75 percent	73.6	62.6	77.4	59.9	70.9	64.6
75 percent or higher	71.2	64.4	71.8	62.6	70.8	66
Work for pay last week
Yes	69.5	64.4	71.9	61	67.4	67.5
No	61.4	71.4	64.8	68.8	60	72.4
If worked last week--work from home/telework
None	72.8	61.2	75.5	58.9	70.1	63.5
Some	73.2	60.7	71.6	57.1	74.9	64.7
All	58	76.1	60.5	72.3	56.4	78.4
Spouse or partner lives in household
Yes	70.3	64	71.9	61.2	68.8	66.7
No	58.7	72.4	63.5	68.8	57.1	73.5
Spouse or partner work for pay last week
Yes	71.8	62.2	73.7	58	70.3	65.6
No	64.1	71.1	66.5	69.3	59.2	74.6
Note: Data are weighted. Source: U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1997.

Overall, 68 percent of the limited sample described above had a child who attended in-person classes in the last week and 66 percent had a child who attended classes remotely in the last week. Non-Black, non-Hispanic men and women were more likely to have a child attend in-person classes (75 percent) than Black (47 percent) or Hispanic (52 percent) men and women. In contrast, non-Black, non-Hispanic men and women were less likely to have a child attend remote classes (62 percent) than Black (77 percent) or Hispanic (75 percent) men and women. Men and women with less than a high school diploma were much less likely to have a child taking in-person classes (58 percent) than men and women with a bachelor’s degree and higher (72 percent).

Whether a child attended in-person school or did remote schooling may be related to parents’ ability to work from home and juggle a number of responsibilities.⁵ Men who worked last week were more likely to have children who attended in-person schooling (72 percent) than men who did not work last week (65 percent). Similarly, women who worked last week were more likely to have children attend in-person schooling (67 percent) than women who did not work last week (60 percent). Men who worked all hours from home were more likely to have children learning remotely (72 percent) than men who worked some or no hours from home (57 to 59 percent). Women who worked all hours from home were also more likely to have children learning remotely (78 percent) than women who worked some or no hours from home (65 to 64 percent).

Respondents with a child learning remotely in the last week were asked how much they agreed with the statement “Last week, remote or distance learning for children under 18 years old living in my household made it difficult for me to work or do other household tasks.” In chart 3, data are limited to households in which a child under 18 engaged in remote or distance learning. It shows the percentage who agree or strongly agree with the statement above regarding remote learning, by respondent work status and amount of telework in the last week. Additionally, the chart shows that men and women who work on site are the least likely to agree that remote schooling makes it difficult to work or do other household tasks than men and women who did not work or did at least some work from home. For example, 48 percent of women who worked on-site exclusively agreed with the above statement compared with 65 percent of women who worked from home exclusively. Similarly, 42 percent of men who worked on-site exclusively agreed with the statement about remote learning making it difficult to work compared with 58 percent of men who worked from home exclusively.

Employment changes over the last 12 months

Survey respondents were asked about any changes to employment in the last 12 months. At the time that the survey was given, “the last 12 months” roughly coincided with the length of pandemic up to that point. Overall, about 90 percent of men and 83 percent of women worked in the last 12 months. Chart 4 depicts the percentage of men and women who worked for pay or profit over the last 12 months by race, ethnicity, and education. Non-black, non-Hispanic men were more likely to have worked in the last 12 months than non-Black, non-Hispanic women (92 percent compared with 84 percent) and Hispanic men were more likely to have worked than Hispanic women (90 percent compared with 80 percent). Similar percentages of Black men and women had worked in the last 12 months (81 to 83 percent). As with the shorter frame of reference (one week), the longer timeframe also shows that the probability of employment increases with education for both men and women. Of men with less than a high school diploma, 70 percent had worked in the last 12 months, compared with 98 percent of those with a bachelor’s degree and higher. Similarly, 64 percent of women with less than a high school diploma had worked in the last 12 months compared with 91 percent of women with a bachelor’s degree and higher.

To try to measure volatility in respondents’ labor market situation in the last 12 months, the supplement asked those who worked for pay or profit in the last 12 months about changes to work or earnings because of the COVID-19 pandemic. The survey noted that “changes may have occurred because of government restrictions on people’s activities, because of your or others’ COVID-related illnesses, school or day care closings, or because of the overall changes in the economy because of the coronavirus pandemic.”

Table 4 displays responses about employment changes caused by the pandemic by demographic characteristics and for men and women, for those who worked in the last 12 months. About 26 percent of the NLSY97 sample reported that they stopped work for an employer because of the pandemic and 18 percent of the sample reported they started work for a new employer in the last 12 months. Similar percentages reported that work hours decreased (29 percent), work hours increased (30 percent), earnings decreased (31 percent), and earnings increased (33 percent) over the last 12 months. Black men were much more likely to report that they stopped working for an employer in the last 12 months (38 percent) than non-Black, non-Hispanic men (22 percent) or Hispanic men (27 percent). About 25 percent of non-Black, non-Hispanic women reported stopping work for an employer, 30 percent of Black, non-Hispanic women stopped working for an employer, and 32 percent of Hispanic women stopped working for an employer.

Table 4. Percentage of people born in 1980 to 1984 who worked during the last 12 months (February through May 2021) who reported changes because of the COVID-19 pandemic

Characteristics	Stop working for employer	Start work at new employer	Hours decreased	Hours increased	Earnings decreased	Earnings increased
All
Total	25.8	18.1	29.1	29.6	31.2	33.3
Race and ethnicity
Non-black, non-Hispanic	23.3	16.8	27.3	30.2	30.1	33.8
Black, non-Hispanic	33.9	25.1	33.7	29.5	33.6	33.8
Hispanic or Latino	29.1	16.6	32.3	25.9	34	29.2
Education
Less than a high school diploma	41.8	22.8	48.1	27.4	45.8	30.2
GED	38.6	26.6	43.7	30.4	39.3	31.5
High school graduate., no college	30.5	18.5	34.9	30.9	35.1	31.5
Some college or associate degree	28.4	20.7	31.5	32.7	34.8	35.9
Bachelor's degree and higher	17.8	13.8	19.9	27	24.2	32.9
AFQT percentile test score
Less than 25 percent	32.7	23.5	38.5	30.2	36	28.6
25 percent to less than 50 percent	27.4	18.1	29.5	29.1	30	31.3
50 percent to less than 75 percent	23.6	17.6	24.9	31.7	29.6	32.2
75 percent or higher	19.8	14.2	24	28.9	27.8	37.9
Spouse or partner lives in household
Yes	23	15.8	27.1	28.5	30.4	33.1
No	32.6	23.9	33.8	32.5	33.3	34.1
Child < 18 lives in household
Yes	23.6	16.7	28	28.5	29.8	32.5
No	30.4	21.1	31.3	32.1	34.1	35.3
Self-rated health
Excellent/very good	23.1	16.5	25.3	30.3	27.7	35.1
Good	26.8	19	31.4	28.4	32.8	31.2
Fair/poor	32.7	21.7	36.7	30.2	40.3	32.9
Men
Total	24.8	17.5	26.1	30.7	30.2	36.1
Race and ethnicity
Non-black, non-Hispanic	21.7	15.7	23.5	30.9	28.1	36.7
Black, non-Hispanic	37.8	26.8	34.3	32.5	35.7	37.1
Hispanic or Latino	26.6	17	29.8	27.5	34.1	30.7
Education
Less than a high school diploma	44.1	24	42.4	32	41.8	30.7
GED	37.7	24.6	41.5	30.2	39.2	28.8
High school graduate., no college	27.1	16.4	28.7	33.5	32.4	35.2
Some college or associate degree	25.4	18.4	28.4	33.1	33.7	38.9
Bachelor's degree and higher	18	14.8	17.4	27.4	23.1	36.6
AFQT percentile test score
Less than 25 percent	31.5	20.6	34.9	34.2	35.7	31.8
25 percent to less than 50 percent	23.8	17.5	24.4	29.6	27.8	30.8
50 percent to less than 75 percent	20.8	17.7	22.1	32.8	28.1	37
75 percent or higher	21.5	14	22.2	27.7	26.5	39.6
Spouse or partner lives in household
Yes	21.6	15.2	23.5	29.6	29.2	35.9
No	33.6	23.7	32.9	34.2	33	36.9
Child < 18 lives in household
Yes	21.3	14.8	24.2	30	28.9	35.5
No	30.7	21.9	28.9	32.2	32.2	37.3
Self-rated health
Excellent/very good	22.9	16.6	21.8	31.2	26.3	36.7
Good	25.9	17.8	28.7	29.7	32.1	35.2
Fair/poor	29.6	19.8	36.5	31.5	41.9	36.8
Women
Total	26.8	18.8	32.4	28.3	32.3	30.2
Race and ethnicity
Non-black, non-Hispanic	25.1	18.1	31.6	29.5	32.4	30.5
Black, non-Hispanic	29.9	23.4	33.2	26.6	31.6	30.6
Hispanic or Latino	32.3	16.2	35.6	23.7	34	27.3
Education
Less than a high school diploma	39.3	21.5	54.5	22.2	50.2	29.6
GED	40.1	29.9	47.3	30.8	39.4	36
High school graduate., no college	35.5	21.6	43.9	27.1	39.2	26.1
Some college or associate degree	31.9	23.5	35.2	32.3	36	32.5
Bachelor's degree and higher	17.5	12.8	22.1	26.6	25.2	29.5
AFQT percentile test score
Less than 25 percent	34.3	27	42.9	25.2	36.5	24.7
25 percent to less than 50 percent	31.2	18.8	34.8	28.6	32.3	31.8
50 percent to less than 75 percent	26.3	17.4	27.8	30.6	31	27.4
75 percent or higher	17.8	14.5	26.2	30.2	29.3	35.8
Spouse or partner lives in household
Yes	24.6	16.5	31.4	27.2	31.8	29.7
No	31.7	24	34.8	30.9	33.7	31.3
Child < 18 lives in household
Yes	25.8	18.5	31.5	27.1	30.7	29.6
No	30	19.7	35.3	31.8	37.4	32
Self-rated health
Excellent/very good	23.3	16.4	29.4	29.2	29.4	33.3
Good	27.9	20.3	34.6	27	33.5	26.7
Fair/poor	35.7	23.6	36.9	29.1	38.8	29.2
Note: Data are weighted. Source: U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1997.

Men and women with a bachelor’s degree and higher were less likely to report volatility in their labor market activity in the last 12 months, compared with their counterparts with less education as shown in Table 4 and Charts 5 and 6. (Job volatility is measured by stopping a job, experiencing a decrease in hours, or experiencing a decrease in earnings.) For example, 18 percent of men with a bachelor’s degree and higher reported stopping work for an employer, 17 percent reported a decrease in hours, and 23 percent reported a decrease in earnings. In contrast, 44 percent of men with less than a high school diploma reported stopping work for an employer, 42 percent reported a decrease in hours, and 42 percent reported a decrease in earnings. For women, there were also stark differences in labor market volatility in the last 12 months by education, with those with a bachelor’s degree and higher showing more stability than other groups. About 18 percent of women with a bachelor’s degree and higher reported stopping work for an employer, 22 percent reported a decrease in hours, and 25 percent reported a decrease in earnings. In contrast, 39 percent of women with less than a high school diploma reported stopping work for an employer, 55 percent reported a decrease in hours, and 50 percent reported a decrease in earnings.

As shown in Table 4, both men and women who had a spouse, partner, or a child under 18 in the household were less likely to stop working for an employer in the last 12 months. Individuals with poorer health had more labor market volatility than those with excellent or very good health.

Summary

The results from the special supplement to the NLSY97 during February through May 2021 illustrate some of the effects of the COVID-19 pandemic on men’s and women’s labor market experiences during that time. The findings show differences across demographic groups; those with lower levels of education, poorer health, and minority workers were often affected more than others. As longitudinal NLSY97 data are released detailing employment before, during, and after the pandemic, researchers will be able to analyze these relationships more fully by controlling for employment history, occupational characteristics, and geographic location. The complete NLSY97 data will permit researchers to trace the impact of the pandemic on labor market experiences over the shorter and (eventually) longer term—as the effects may reverberate over many aspects of peoples’ lives.

Note:  The views expressed are those of the authors and do not reflect the policies of the Bureau of Labor Statistics or the views of other BLS staff members.

This Beyond the Numbers article was prepared by Alison Aughinbaugh and Donna S. Rothstein, research economists in the Office of Employment and Unemployment Statistics (OEUS), U.S. Bureau of Labor Statistics. E-mail: NLS_INFO@bls.gov; telephone: 202-691-7410; E-mails: aughinbaugh.alison@bls.gov and rothstein.donna@bls.gov.