Department of Labor Logo United States Department of Labor
Dot gov

The .gov means it's official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you're on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Article
March 2023

Employment, telework, and child remote schooling from February to May 2021: evidence from the National Longitudinal Survey of Youth 1997

In this article, we use data on work and telework from a COVID-19 supplement to the National Longitudinal Survey of Youth 1997 collected from February to May 2021. We examine whether the likelihoods of work and telework are associated with background characteristics related to demographics, education, geography, and prepandemic jobs. We also focus on differences between women and men and differences between individuals with children enrolled in school and those without children enrolled in school. Additionally, we examine factors associated with the likelihood of children in the individual’s household attending remote or in-person schooling.

In March 2020, the coronavirus disease 2019 (COVID-19) pandemic triggered a crisis in the U.S. labor market on a scale not seen since the Great Depression (1929 to 1940). In March and April 2020 (combined), nonfarm payroll employment declined by over 21 million and the unemployment rate rose to 14.7 percent.1 A year later, the economic situation had improved. In April 2021, employment was still down about 8 million jobs compared with its prepandemic level. The unemployment rate was 6.1 percent in April 2021.2

The pandemic also changed the way that work is performed, as some employees switched from in-person work to telework (or working from home). Matthew Dey, Harley Frazis, David S. Piccone Jr, and Mark A. Loewenstein use data from the Current Population Survey to show that from May to June 2020 about a third of workers reported that they teleworked because of the pandemic.3 The percentage of individuals who teleworked fell to 25 percent from July to September 2020 and to 22 percent from October to December 2020. In April 2021, about a year after the start of the pandemic, nearly 18 percent of workers reported teleworking because of the pandemic.4

In March 2020, school closures were implemented to attempt to prevent the spread of COVID-19. By March 25, 2020, all U.S. public school buildings were closed.5 Only two states (Wyoming and Montana) reopened their schools before the end of the 2019–20 academic year.6 Disruptions continued well into the 2020–21 school year. The National Assessment of Educational Progress (NAEP) Monthly School Survey showed that nearly a year passed before most U.S. schools were open in person and full time for all students. In February 2021, 49 percent of public schools with fourth or eighth grade were open full time and in person for all students. This percentage had grown to 63 percent by May 2021.7

The National Longitudinal Survey of Youth 1997 (NLSY97) fielded a short NLSY97 COVID-19 Supplement from February to May 2021 to gather information on work, working conditions, health, and children’s schooling.8 In this article, we use data from the NLSY97 COVID-19 Supplement to examine the likelihood of work in the week before the survey and the likelihood the work was performed by telework. The outcome measures of work and telework are analyzed in relation to employment and job characteristics measured during the 6 months preceding the pandemic. We examine differences in these relationships between women and men overall and between women and men with children enrolled in school. We also look at teleworking behavior and patterns of remote and in-person schooling incidence by sex and race during spring 2021.

Data from the NLSY97

The NLSY97 is a nationally representative sample of 8,984 men and women born from 1980 to 1984 who were living in the United States during their first interview in 1997. NLSY97 respondents were interviewed annually from 1997 to 2011 and biennially since 2011. The latest public-release full interview, round 19 of data collection for the NLSY97, took place from September 2019 to July 2020. About 90 percent of data collected during round 19 occurred before March 2020.9 In spring 2021, the NLSY97 fielded a supplement on the effects of the COVID-19 pandemic called the NLSY97 COVID-19 Supplement.10 The supplement interviews were conducted from February to May 2021. The supplement data include information on employment, telework, health, and children’s schooling. NLSY97 respondents were between ages 36 and 41 at the time of the supplement.11  

In this article, we limit our sample to men and women who participated in both the NLSY97 COVID-19 Supplement and the NLSY97 round 19 data collection. Furthermore, we delete a small number of observations for which key variables were missing. The sample size for this analysis is 5,238 when these observations are excluded.12 In addition to using our full sample, we limit some of our analysis to respondents with children under 18 in their households who were enrolled in K–12 public, private, or other schools.13 We use data from this subsample of 2,819 respondents to examine the relationship between various background characteristics and the mode of children’s schooling (remote schooling or in-person schooling).

Variables and descriptive statistics

Table 1 shows the outcome measures that are dependent variables in the regression models: employment, telework, and whether children’s schooling was remote or in person. All measures pertain to the time of the NLSY97 COVID-19 Supplement interview. About 77 percent of individuals in the full sample worked in the week before their supplement interview. Women were less likely to be working than men (73 percent of women compared with 81 percent of men). Women and men in the subsample with children in school were more likely to be working compared with those in the full sample (75 percent of women and 87 percent of men in the school sample).

Table 1. Employment, telework, and children’s schooling mode in the week before the NLSY97 COVID-19 Supplement interview, February to May 2021 [1]
VariablesFull SampleSample with children in school [2]Full sampleSample with children in school [2]
AllAllWomenMenDifferenceWomenMenDifference

Employment last week

Worked

77.080.072.681.2-8.674.587.0-12.5

If worked

Any telework

46.446.151.042.58.549.043.06.0

Full telework

25.223.329.821.38.527.718.59.2

Children’s schooling mode last week

Only in-person schooling

31.328.235.3-7.1

Only remote schooling

30.032.327.05.2

Hybrid: both in-person and remote

36.237.035.31.7

Neither in-person nor remote

2.52.52.40.1

[1] NLSY97 is the National Longitudinal Survey of Youth 1997, and COVID-19 is coronavirus disease 2019.

[2] Children are in K–12 and under age 18 and at least one child is not homeschooled.

Note: The means in the table are weighted means.

Source: U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1997.

Table 1 shows that among individuals who worked in the week before the COVID-19 Supplement interview, 46 percent teleworked some of the time and 25 percent teleworked all the time in that week. Almost 54 percent of individuals who worked in the prior week did not telework. Women in the full sample were more likely to telework than men (51 percent versus 43 percent for some telework and 30 percent versus 21 percent for full telework). Table 1 shows that levels of telework are similar in the schooling sample.

By late winter and spring 2021, a growing number of schools were at least partially in person, but distance learning remained in effect at many schools.14 The bottom part of table 1 shows the percentage of individuals with children that attended only in-person, only remote, and hybrid (both in-person and remote) K–12 schooling in the last week. Among individuals with children enrolled in K–12, 31 percent had children who attended only in-person schooling, 30 percent had children who attended only remote schooling, and 36 percent had children who attended hybrid schooling. Note that if an individual had one child who attended in person and another child that attended remotely, then they would be placed into the hybrid category in table 1. Women were less likely than men to have children who attended only in-person schooling (28 percent compared with 35 percent). Women were more likely than men to have children who attended only remote schooling (32 percent compared with 27 percent).

The reports of schooling mode in the NLSY97 COVID-19 Supplement closely match those in the NAEP Monthly School Survey over the same months.15 In addition, a comparison between the child in-person and remote schooling reports from the NLSY97 COVID-19 Supplement with data from the Census Bureau's Household Pulse Survey (HPS) shows consistent patterns by sex (chart 1) and by race and ethnicity (chart 2).16 The NLSY97 COVID-19 Supplement and the HPS show that women in the 1980–84 cohort were less likely than men to report that their children attended only in-person schooling. Women were more likely than men to report that their children attended only remote schooling in the prior week. Both surveys also show that non-Black, non-Hispanic individuals were more likely to report that children in their household attended only in-person schooling than Black or Hispanic individuals. Also, non-Black, non-Hispanic individuals were less likely to report that children in their household attended only remote schooling.

Table 2 displays weighted descriptive statistics for background variables used in our employment and schooling regressions. This table shows the descriptive statistics for our full sample and subsample of individuals with children in the household under age 18 who were enrolled in K–12 in the week before the NLSY97 COVID-19 Supplement interview. The following information is presented in table 2:

  • Employment and job characteristics as of the round 19 interview, including whether the job is conducive to telework (described below)
  • Demographics (sex, race, and ethnicity)
  • Percentile score on the Armed Forces Qualification Test (AFQT)17
  • Highest educational degree
  • Household composition at the time of the NLSY97 COVID-19 Supplement interview
  • Urbanicity (whether the individual lived in an urban area)
  • County-level change in activity at workplaces (described below)
  • Whether health conditions limit work at the round 19 interview

We use reports of respondents about the kinds of tasks that make up their job to classify the jobs reported at the prior round 19 interview as conducive to telework. We use these reports to define a job as conducive to telework if (1) respondents report that less than half their day is spent doing physical tasks and (2) respondents report that their job requires a moderate or low amount of face-to-face contact with people other than coworkers or supervisors.18 About the same percentage of men and women who were working at the round 19 interview in the full sample were in jobs conducive to telework. In the school sample, women were more likely than men to be in jobs conducive to telework. This result shows occupational differences between the full sample and the school sample.19

County-level change in activity at workplaces is defined as the percent change in visits to places of work between a baseline period before the COVID-19 pandemic (January 3, 2020, to February 6, 2020) and the period of the supplement interview (February to May 2021) as measured by Google cell-phone location data that provide information on visits to workplaces in the respondent’s round 19 county of residence.20

Table 2. Descriptive statistics of background variables from NLSY97 round 19 and NLSY97 COVID-19 Supplement interview [1]
VariablesFull sampleSample with
children in school [1]
AllWomenMenAllWomenMen

Employment and job characteristics at round 19 interview

Not working

15.419.611.413.518.47.2

If working at job at round 19 interview

Military

1.40.71.91.40.72.2

Job is conducive to telework

29.230.927.628.330.525.9

Demographics

Female

48.7100.00.056.0100.00.0

Race/ethnicity

Non-black, non-Hispanic

70.671.070.170.770.670.9

Black, non-Hispanic

15.315.515.114.415.213.4

Hispanic

12.812.413.213.413.013.8

Other

1.31.11.51.51.21.9

Quartile of AFQT score [2]

1st

21.020.321.720.920.821.1

2nd

23.424.322.524.826.522.5

3rd

25.826.824.827.027.226.7

Highest

29.928.631.127.325.629.7

Highest degree completed

Less than high school

4.84.94.74.95.54.1

GED [3]

7.96.79.17.86.29.8

High school diploma

19.717.521.721.119.223.5

Some college

29.028.129.929.329.928.6

Bachelor’s degree or higher

38.642.934.536.939.134.1

Household composition

Spouse/partner in household

69.168.669.579.872.389.4

Children under age 18 in household

No children

31.823.739.60.00.00.0

Only children under age 6

13.313.013.67.66.39.3

Only children ages 6 to 17

35.743.128.860.263.955.6

Children both under age 6 and ages 6 to 17

19.120.218.132.129.835.1

Geography at round 19 interview

Urban

79.778.980.576.777.475.8

County-level change in activity at workplaces

-26.9-27.0-26.8-26.0-26.1-25.8
(8.6)(8.6)(8.6)(8.3)(8.4)(8.2)

Health at round 19 interview

Health condition limits work

9.211.47.17.49.54.7

Sample size

5,2382,8322,4062,8191,7411,078

[1] Children are in K–12 and under age 18 and at least one child is not homeschooled. NLSY97 is the National Longitudinal Survey of Youth 1997. COVID-19 is coronavirus disease 2019.

[2] AFQT is the Armed Forces Qualification Test.

[3] The GED is the General Educational Development test.

Note: The means in the table are weighted means. Standard deviations are in parentheses.

Source: U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1997.

Table 2 shows that 11 percent of men and 20 percent of women were not working at the time of their round 19 interview. The sex difference in the percentage not working at the round 19 interview is slightly larger between men and women with school children than between men and women in the full sample. Among those with children in school, 18 percent of women and 7 percent of men were not working at the time of the round 19 interview.

In addition, the full sample shows some differences by sex with respect to the presence and age of children in the household at the date of the supplement interview. Table 2 shows that men in the full sample were more likely than women to have no children under 18 in their household (40 percent compared with 24 percent). Women in the full sample were more likely than men to have children ages 6 to 17 in the household (43 percent compared with 29 percent).21 Similar percentages of women and men in the full sample have children in the household who are only under age 6 or both under age 6 and ages 6 to 17.

The last three columns of table 2 present the statistics for the subsample of individuals who had children under age 18 in K–12 when the supplement was fielded. These columns compared with the first three columns show some differences between the school sample and the full sample. The school sample is composed of a higher percentage of women (56 percent compared with 49 percent), a higher percentage of individuals with a spouse or partner present in the household at the time of the supplement (80 percent compared with 69 percent), and a higher percentage of individuals with children ages 6 to 17. Overall, education levels and AFQT scores are similar across the two samples.

Empirical specification

We estimate linear probability models for the likelihood of

  • working in the week before the NLSY97 COVID-19 Supplement interview,
  • teleworking conditional on working, and
  • having children in the household who attended only in-person, only remote, or hybrid schooling in the week before the supplement interview.

The estimates control for demographics, AFQT score, education, health, and household composition. Importantly, the work-related estimates also control for whether an individual was working at the round 19 interview and, for those working, whether their job was conducive to telework at that time.

In addition, we include state fixed effects in the models to control for differences across states in the economic environment. These state fixed effects may control for pandemic-related restrictions on economic activity. We include a variable to indicate whether the individual lived in an urban area to capture differences between urban and rural areas. We control for the percent change in county-level activity at workplaces in an attempt to capture within state differences in local behavior during the pandemic. All location variables are based on the individuals’ residence at the round 19 interview date.

Results

Our analysis of work, telework, and children’s schooling during February to May 2021 is divided into two subsections on the basis of the sample used for the estimation: the full sample of the supplement interviews and the sample with children under 18 enrolled in K–12 in the week before the supplement interview.

Full sample

In this subsection, we examine the findings from our analysis of the full sample of the NLSY97 COVID-19 Supplement interviews. First, we consider the probability of employment. Second, we consider the probability of telework among the employed.

Probability of employment

Table 3 shows estimates of the probability of working in the week before the NLSY97 COVID-19 Supplement interview. The strongest result is that recent past employment predicts employment in spring 2021. On average, those who were not working at the round 19 interview are 49 percentage points (or 63 percent) less likely to be working at the time of the NLSY97 COVID-19 Supplement interview.22 The coefficient estimate is larger for women than for men. Women and men, respectively, who were not working at the round 19 interview are 52 percentage points and 38 percentage points less likely to be working at the time of the supplement. Moreover, whether the job at round 19 was conducive to telework is related to employment at the time of the supplement interview. Overall, people working in jobs conducive to telework at round 19 are 4 percentage points more likely to be working the week before the supplement interview. Separate estimates for women and men show that women who had jobs conducive to telework at the round 19 interview were 7 percentage points more likely to be working. But no effect is observed for men. Overall, these estimates indicate that women’s employment in spring 2021 was correlated with employment immediately before the COVID-19 pandemic to a greater extent than was men’s employment.

Table 3. Probability of working in the week before the NLSY97 COVID-19 Supplement interview, full sample, February to May 2021 [1]
VariablesAllWomenMen

Demographics

Female

-0.057
(0.011)

Race/ethnicity

Black, non-Hispanic

-0.027-0.011-0.052
(0.016)(0.022)(0.024)

Hispanic

0.000-0.0210.025
(0.016)(0.023)(0.021)

Other

-0.0620.005-0.089
(0.061)(0.105)(0.072)

Quartile of AFQT score [2]

2nd

0.0470.0490.056
(0.019)(0.027)(0.027)

3rd

0.0410.0460.045
(0.020)(0.029)(0.027)

Highest

0.0530.0730.041
(0.020)(0.030)(0.027)

Highest degree completed

GED [3]

0.0250.0170.041
(0.038)(0.053)(0.051)

High school diploma

0.0980.0770.121
(0.033)(0.046)(0.046)

Some college

0.1260.0850.170
(0.033)(0.047)(0.045)

Bachelor’s degree or higher

0.2130.1830.249
(0.033)(0.046)(0.046)

Household composition

Spouse/partner in household

0.018-0.0220.058
(0.014)(0.018)(0.021)

Children under age 18 in household [4]

No children

-0.0400.002-0.066
(0.014)(0.02)(0.021)

Only children under age 6

-0.036-0.039-0.027
(0.016)(0.024)(0.021)

Children both under age 6 and ages 6 to 17

-0.044-0.045-0.037
(0.015)(0.021)(0.021)

Geography at round 19 interview

Urban

0.0000.019-0.031
(0.015)(0.022)(0.021)

County-level change in activity at workplaces

0.0010.0030.000
(0.001)(0.001)(0.001)

State fixed effect

YesYesYes

Employment and job characteristics at round 19 interview

Not working

-0.487-0.521-0.377
(0.021)(0.026)(0.035)

Military

0.0690.1770.037
(0.027)(0.027)(0.037)

Job is conducive to telework

0.0420.0720.010
(0.011)(0.016)(0.016)

Health at round 19 interview

Health condition limits work

-0.114-0.087-0.177
(0.026)(0.033)(0.041)

Sample size

5,2382,8322,406

[1] NLSY97 is the National Longitudinal Survey of Youth 1997, and COVID-19 is coronavirus disease 2019.

[2] AFQT is the Armed Forces Qualification Test.

[3] The GED is the General Educational Development test.

[4] Only children ages 6 to 17 is the omitted category.

Note: The regressions in the table are weighted regressions. Specifications include a control for mode of interview and dummy variables indicating measures are missing for AFQT score, highest degree completed, urbanicity, and job is conducive to telework. Coefficient estimates are reported, and robust standard errors are in parentheses.

Source: U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1997.

In addition, several demographic characteristics are related to employment in spring 2021. Our estimates show that when we control for the other covariates, women are less likely than men to work by 6 percentage points. AFQT scores and education also affected the probability of work. Both men and women with higher AFQT scores are more likely to work by at least 4 to 5 percentage points relative to individuals with AFQT scores in the lowest quartile. Those with higher completed levels of education were also more likely to be working. We find that as the amount of education completed increased, the probability of working for men and women also increased. In the full sample, we find that household composition is related to the likelihood of working in the week before the COVID-19 Supplement interview. Individuals with no children in their household or with any children under the age of 6 are less likely to work compared with individuals whose children are all between ages 6 and 17 (the excluded category).23 This pattern varies between women and men. We find that women with no children are no more likely or no less likely to work than those whose youngest child is age 6 to 17. However, women with a child under age 6 (either those whose children are all under age 6 or who have older children as well) are 4 percentage points less likely to work compared with women whose children are all in the age range of 6 to 17. Disruptions in childcare may have contributed to the lower likelihood of work among women with younger children.24 In contrast, men with no children are less likely to work (the differences based on ages of children are not statistically significant). In general, we find that the presence of a spouse or partner in the household is unrelated to the probability of working. But this estimate masks different patterns for men versus women. Men living with a spouse or partner are 6 percentage points more likely to work.

With respect to geography measured at the round 19 interview, the percent change in county-level activity at workplaces influences the likelihood of working for women only. An increase in county-level activity by 9 percentage points (a little over a standard deviation) increases the probability of working by almost 3 percentage points for women.

Probability of telework

We use the same set of explanatory factors to estimate the probability that workers are teleworking in the week before the NLSY97 COVID-19 Supplement interview. Table 4 shows the probability of any telework in the week before the supplement interview, and Table 5 shows the probability of full telework. The sample is restricted to individuals who worked in the week before the supplement interview. Those who had jobs that were classified as conducive to telework at the time of the round 19 interview were more likely to telework in spring 2021. Women whose job prior to the start of the pandemic was considered conducive to telework were at least 30 percentage points more likely to telework (for either measure of telework, any or full), and men with jobs conducive to telework at that time were 25 percentage points more likely to telework (for either measure of telework, any or full). Notably, having not worked at the time of the round 19 interview is associated with an increase of 7 percentage points in the probability of working fully by telework. But the estimated effect in the any-telework equation is smaller and not statistically different from zero.

Table 4. Probability of any telework in the week before the NLSY97 COVID-19 Supplement interview among those who are working, full sample, February to May 2021 [1]
VariablesAllWomenMen

Demographics

Female

0.050
(0.016)

Race/ethnicity

Black, non-Hispanic

-0.007-0.004-0.016
(0.021)(0.028)(0.031)

Hispanic

0.0090.023-0.007
(0.023)(0.032)(0.032)

Other

0.0660.1100.030
(0.074)(0.095)(0.016)

Quartile of AFQT score [2]

2nd

0.0490.0390.054
(0.025)(0.037)(0.035)

3rd

0.0880.0990.084
(0.026)(0.038)(0.036)

Highest

0.1440.0870.195
(0.028)(0.041)(0.0369)

Highest degree completed

GED [3]

0.0610.0450.070
(0.043)(0.067)(0.057)

High school diploma

0.0600.0720.044
(0.037)(0.058)0.050

Some college

0.1230.1260.115
(0.037)(0.057)(0.050)

Bachelor’s degree or higher

0.3280.3350.318
(0.038)(0.058)(0.052)

Household composition

Spouse/partner in household

0.0800.0710.091
(0.018)(0.024)(0.028)

Children under age 18 in household [4]

No children

-0.033-0.025-0.044
(0.019)(0.028)(0.029)

Only children under age 6

0.0190.0250.008
(0.025)(0.036)(0.035)

Children both under age 6 and ages 6 to 17

0.0150.033-0.008
(0.021)(0.030)(0.031)

Geography at round 19 interview

Urban

0.0290.0360.018
(0.021)(0.031)(0.030)

County-level change in activity at workplaces

-0.008-0.009-0.006
(0.001)(0.002)(0.002)

State fixed effect

YesYesYes

Employment and job characteristics at round 19 interview

Not working

0.0430.0080.062
(0.033)(0.045)(0.051)

Military

0.0100.157-0.065
(0.068)(0.138)(0.078)

Job is conducive to telework

0.2870.3140.250
(0.018)(0.024)(0.026)

Health at round 19 interview

Health condition limits work

-0.003-0.0480.072
(0.036)(0.044)(0.062)

Sample size

3,9202,0091,911

[1] NLSY97 is the National Longitudinal Survey of Youth 1997, and COVID-19 is coronavirus disease 2019.

[2] AFQT is the Armed Forces Qualification Test.

[3] The GED is the General Educational Development test.

[4] Only children ages 6 to 17 is the omitted category.

Note: The regressions in the table are weighted regressions. Specifications include a control for mode of interview and dummy variables indicating measures are missing for AFQT score, highest degree completed, urbanicity, and job is conducive to telework. Coefficient estimates are reported, and robust standard errors are in parentheses.

Source: U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1997.

Women were more likely than men to do any amount of telework by 5 percentage points and to telework fully by 6 percentage points. The likelihood of teleworking is higher for individuals with higher AFQT scores and more education. Workers with AFQT scores in the highest quartile are 14 percentage points more likely to telework some of the time than those in the lowest quartile. Those with a bachelor’s degree or higher are 33 percentage points more likely to telework some of the time than those with less than a high school diploma. In actuality, the relationship between education and teleworking is even more pronounced than these estimates show because workers with more education are more likely to be in jobs classified as conducive to telework. This finding is similar to other studies on telework during the COVID-19 pandemic.25 Having a bachelor’s degree or higher increases the probability of full telework more for women than for men (19 percentage points versus 13 percentage points). Telework is also related to some aspects of household composition. Workers with a spouse or partner in the household were 8 percentage points more likely to do at least some telework at the time of the supplement. Surprisingly, we do not find a statistically significant effect of the presence and ages of children in the household on telework. As expected, a decline in county-level activity at workplaces is associated with a substantial increase in the likelihood of telework.

Table 5. Probability of full telework in the week before the NLSY97 COVID-19 Supplement interview among those who are working, full sample, February to May 2021 [1]
VariablesAllWomenMen

Demographics

Female

0.057
(0.014)

Race/ethnicity

Black, non-Hispanic

0.0250.057-0.004
(0.018)(0.027)(0.024)

Hispanic

-0.007-0.010-0.017
(0.021)(0.032)(0.027)

Other

0.0570.166-0.039
(0.071)(0.093)(0.094)

Quartile of AFQT score [2]

2nd

0.0210.0060.029
(0.019)(0.033)(0.024)

3rd

0.0340.0070.058
(0.021)(0.035)(0.027)

Highest

0.0570.0020.105
(0.024)(0.038)(0.031)

Highest degree completed

GED [3]

0.024-0.0020.022
(0.033)(0.057)(0.039)

High school diploma

0.0280.0590.000
(0.027)(0.049)(0.030)

Some college

0.0510.0640.035
(0.027)(0.048)(0.032)

Bachelor’s degree or higher

0.1570.1910.125
(0.029)(0.049)(0.036)

Household composition

Spouse/partner in household

0.0340.0540.019
(0.016)(0.022)(0.023)

Children under age 18 in household [4]

No children

0.007-0.0090.012
(0.017)(0.026)(0.024)

Only children under age 6

0.0400.0420.033
(0.023)(0.036)(0.029)

Children both under age 6 and ages 6 to 17

0.008-0.0120.020
(0.019)(0.028)(0.026)

Geography at round 19 interview

Urban

0.0180.0140.023
(0.017)(0.027)(0.023)

County-level change in activity at workplaces

-0.009-0.008-0.010
(0.001)(0.002)(0.001)

State fixed effect

YesYesYes

Employment and job characteristics at round 19 interview

Not working

0.0670.0640.066
(0.028)(0.038)(0.041)

Military

-0.106-0.042-0.140
(0.033)(0.097)(0.031)

Job is conducive to telework

0.2920.3340.251
(0.018)(0.025)(0.027)

Health at round 19 interview

Health condition limits work

0.0540.0640.065
(0.032)(0.043)(0.050)

Sample size

3,9202,0091,911

[1] NLSY97 is the National Longitudinal Survey of Youth 1997, and COVID-19 is coronavirus disease 2019.

[2] AFQT is the Armed Forces Qualification Test.

[3] The GED is the General Educational Development test.

[4] Only children ages 6 to 17 is the omitted category.

Note: The regressions in the table are weighted regressions. Specifications include a control for mode of interview and dummy variables indicating measures are missing for AFQT score, highest degree completed, urbanicity, and job is conducive to telework. Coefficient estimates are reported, and robust standard errors are in parentheses.

Source: U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1997.

Sample with children in school

In this subsection, we consider the employment and telework outcomes of women and men who live with children enrolled in K–12 and under age 18. We also look at the probability that children in an individual’s household attended various schooling modes (only in-person schooling, only remote schooling, or hybrid schooling) in the week before the NLSY97 COVID-19 Supplement interview.

Probability of employment

The employment outcomes of women and men who have children in school are substantially affected by their round 19 employment status and whether their round 19 job was conducive to telework. (See table 6.)  Not working at the time of the round 19 interview decreases the likelihood of working by 47 percentage points in the week before the NLSY97 COVID-19 Supplement interview. As in the full sample, we also see here that this effect is greater among women than among men (49 versus 36 percentage points). Notably, the estimated effects of previous employment are similar in magnitude to those estimated for the full sample.

Table 6. Probabilities of working and teleworking in the week before the NLSY97 COVID-19 Supplement interview among those with children in school, February to May 2021 [1]
VariablesProbability worked last weekAmong those who worked last week
Probability any teleworkProbability full telework
AllWomenMenAllWomenMenAllWomenMen

Demographics

Female

-0.0830.0450.068
(0.014)(0.021)(0.019)

Race/ethnicity

Black, non-Hispanic

-0.019-0.023-0.007-0.005-0.005-0.0290.0380.068-0.010
(0.021)(0.028)(0.032)(0.029)(0.036)(0.051)(0.025)(0.036)(0.037)

Hispanic

-0.008-0.0320.025-0.0020.002-0.0380.0060.012-0.019
(0.021)(0.030)(0.029)(0.031)(0.040)(0.049)(0.028)(0.040)(0.040)

Other

-0.0350.048-0.0970.1640.0760.2420.1300.1430.089
(0.076)(0.142)(0.074)(0.081)(0.105)(0.121)(0.093)(0.113)(0.145)

Quartile of AFQT score [2]

2nd

0.0430.0150.0770.0620.0960.0380.0530.0670.038
(0.025)(0.034)(0.038)(0.034)(0.044)(0.053)(0.025)(0.037)(0.035)

3rd

0.0400.0180.0650.0980.1300.0810.0340.0400.028
(0.026)(0.037)(0.038)(0.035)(0.045)(0.054)(0.026)(0.038)(0.037)

Highest

0.0410.0570.0360.1500.1180.1920.0650.0510.079
(0.027)(0.039)(0.040)(0.038)(0.051)(0.060)(0.032)(0.044)(0.046)

Highest degree completed

GED [3]

-0.049-0.042-0.0390.0280.0480.0040.0250.0070.033
(0.051)(0.067)(0.077)(0.061)(0.088)(0.089)(0.044)(0.077)(0.050)

High school diploma

0.0350.0510.0320.0510.0440.0470.0250.0370.022
(0.044)(0.057)(0.070)(0.053)(0.072)(0.082)(0.035)(0.059)(0.037)

Some college

0.0930.0860.1120.0900.1060.0560.0510.0460.061
(0.044)(0.057)(0.070)(0.052)(0.070)(0.082)(0.035)(0.057)(0.041)

Bachelor’s degree or higher

0.1630.1520.1880.2920.3100.2550.1420.1640.119
(0.045)(0.058)(0.070)(0.054)(0.072)(0.086)(0.038)(0.060)(0.046)

Household composition

Spouse/partner in household

-0.013-0.0280.0270.0560.0390.1030.0210.0080.046
(0.020)(0.024)(0.036)(0.027)(0.032)(0.052)(0.022)(0.028)(0.036)

Children under age 18 in household [4]

Only children under age 6

-0.0260.001-0.0590.0310.0510.0210.0390.0680.022
(0.024)(0.036)(0.035)(0.040)(0.052)(0.060)(0.035)(0.055)(0.048)

Children both under age 6 and ages 6 to 17

-0.050-0.048-0.0470.0250.053-0.0110.0170.0010.023
(0.016)(0.022)(0.022)(0.023)(0.031)(0.034)(0.020)(0.029)(0.028)

Geography at round 19 interview

Urban

-0.015-0.005-0.0350.0240.0290.0040.018-0.0030.022
(0.018)(0.026)(0.027)(0.027)(0.037)(0.041)(0.022)(0.033)(0.031)

County-level change in activity at workplaces

0.0020.0040.000-0.007-0.010-0.005-0.007-0.007-0.008
(0.001)(0.002)(0.002)(0.002)(0.002)(0.003)(0.001)(0.002)(0.002)

State fixed effect

YesYesYesYesYesYesYesYesYes

Employment and job characteristics at round 19 interview

Not working

-0.472-0.491-0.361-0.0020.041-0.0780.0220.068-0.014
(0.029)(0.033)(0.062)(0.047)(0.055)(0.087)(0.035)(0.043)(0.058)

Military

0.0510.1620.0330.0530.076-0.031-0.079-0.055-0.100
(0.028)(0.034)(0.037)(0.091)(0.175)(0.110)(0.045)(0.132)(0.042)

Job is conducive to telework

0.0390.0550.0130.3000.3420.2370.2970.3460.243
(0.014)(0.021)(0.021)(0.024)(0.031)(0.040)(0.025)(0.033)(0.038)

Health at round 19 interview

Health condition limits work

-0.095-0.085-0.1500.0670.0140.1940.1260.1530.098
(0.033)(0.039)(0.070)(0.051)(0.058)(0.105)(0.047)(0.061)(0.076)

Sample size

2,8191,7411,0782,1901,2679232,1901,267923

[1] NLSY97 is the National Longitudinal Survey of Youth 1997, and COVID-19 is coronavirus disease 2019.

[2] AFQT is the Armed Forces Qualification Test.

[3] The GED is the General Educational Development test.

[4] Only children ages 6 to 17 is the omitted category.

Note: The regressions in the table are weighted regressions. Specifications include a control for mode of interview and dummy variables indicating measures are missing for AFQT score, highest degree completed, urbanicity, and job is conducive to telework. Coefficient estimates are reported, and robust standard errors are in parentheses.

Source: U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1997.

Having a job at the round 19 interview that is categorized as conducive to telework increases the probability of working during the week before the supplement interview by 4 percentage points. Again, the estimated effect is roughly the same size as that estimated for the full sample. For the school subsample, having a job conducive to telework during the round 19 interview has no discernable effect for men on the probability of working. In contrast, we find that having a job conducive to telework during the round 19 interview increased the probability of working by 6 percentage points for women with children in school.

Women with children in school are 8 percentage points less likely to work than men with children in school, which is a larger difference than the estimate in the full sample. Highest degree completed has a weaker effect on the probability of working during spring 2021 in the school sample than in the full sample. But this difference between the two samples in the estimates of education level on the probability of working only appears among men. The effect on the probability of working is 6 to 9 percentage points smaller at each level of education for men who live with children enrolled in school than for men in the full sample. We also find that adults with children in school had a lower likelihood of working in spring 2021 if these adults reported in the round 19 interviews a health condition that limited their ability to work. A prior health condition reduced the probability of work during spring 2021 more for men than women (15 and 9 percentage points, respectively).

We find no effect of the presence of either a spouse or partner in the household or the presence of children of different ages on the probability of working when the sample is limited to those with children in school. The change in county-level activity at workplaces is related to the likelihood of employment. That is, greater county-level activity increased the probability of working.

Probability of telework

As noted above, we found that having a job conducive to telework at the round 19 interview was associated with an increase in the likelihood of work by 4 percentage points for individuals in the school sample. (See table 6.) Not surprisingly, among workers, persons having a job at the round 19 interview conducive to telework also affects the likelihood of telework. Specifically, having a job conducive to telework is associated with an increase in the likelihood of telework (either any or full) by 30 percentage points in the school sample. (See table 6.) The effect varies by sex; it is about 10 percentage points larger for women than men. These estimates are similar to those from the full sample.

Working women with children in school were 5 percentage points more likely to do any telework and 7 percentage points more likely to have all work be telework than working men with children enrolled in school. Workers with a bachelor’s degree or higher were more likely to do any telework than those with less than a high school diploma. Women with a bachelor’s degree or higher were 31 percentage points more likely to do any telework, and men with a bachelor’s degree or higher were 26 percentage points more likely to do any telework. The probability of full telework also increased for those with a bachelor’s degree or higher compared with those with less than a high school diploma: 16 percentage points for women and 12 percentage points for men.

Men who had a spouse or partner in the household and children in school were more likely to do some telework. The ages of the children did not affect the probability of telework, which is the same as in the full sample.

The percent change in county-level activity at workplaces was negatively associated with the propensity to telework some or all the time. An increase in county-level workplace activity by 9 percentage points (a little over a standard deviation) decreased the probability of any or full telework in the prior week by 6 percentage points.

Probability of in-person, remote, and hybrid schooling

Many schools had returned to in-person schooling by late winter and spring 2021. But in a substantial number of schools, especially those in large urban areas, remote schooling remained in effect, at least in a hybrid format.26

In table 7, we examine the probability that children in the household attended only in-person schooling, only remote schooling, or hybrid schooling in the week before the NLSY97 COVID-19 Supplement interview. Disparities in whether children had only in-person or only remote schooling are evident by many of the characteristics that we examine.

Table 7. Probability children in the household attended in-person school only, remote school only, or hybrid school in the week before the NLSY97 COVID-19 Supplement interview, February to May 2021 [1]
VariablesProbability in-person school onlyProbability remote school onlyProbability hybrid school
AllWomenMenAllWomenMenAllWomenMen

Demographics

Female

-0.0570.0400.016
(0.019)(0.018)(0.021)

Race/ethnicity

Black, non-Hispanic

-0.167-0.148-0.1840.2080.2010.227-0.070-0.090-0.066
(0.025)(0.030)(0.044)(0.027)(0.034)(0.045)(0.028)(0.035)(0.047)

Hispanic

-0.102-0.070-0.1240.1170.1230.119-0.035-0.074-0.010
(0.025)(0.029)(0.042)(0.028)(0.053)(0.045)(0.028)(0.036)(0.045)

Other

-0.1250.013-0.2380.0600.1030.0350.075-0.1010.206
(0.071)(0.093)(0.097)(0.090)(0.125)(0.127)(0.093)(0.123)(0.131)

Quartile of AFQT score [2]

2nd

0.0200.0250.004-0.033-0.026-0.0340.0450.0370.053
(0.028)(0.032)(0.051)(0.028)(0.036)(0.044)(0.031)(0.040)(0.049)

3rd

0.0140.036-0.014-0.022-0.013-0.0220.0370.0100.060
0.030(0.036)(0.051)(0.029)(0.038)(0.046)(0.033)(0.044)(0.052)

Highest

-0.015-0.010-0.0160.0150.0190.0320.0300.0090.025
(0.033)(0.041)(0.055)(0.033)(0.043)(0.051)(0.036)(0.048)(0.056)

Highest degree completed

GED [3]

0.0450.0840.010-0.029-0.004-0.041-0.017-0.058-0.008
(0.045)(0.053)(0.079)(0.051)(0.069)(0.079)(0.054)(0.070)(0.088)

High school diploma

0.0880.0920.064-0.041-0.054-0.028-0.038-0.019-0.049
(0.038)(0.045)(0.072)(0.044)(0.059)(0.070)(0.047)(0.060)(0.080)

Some college

0.0700.1060.015-0.057-0.079-0.0410.0000.0080.009
(0.037)(0.044)(0.070)(0.045)(0.059)(0.071)(0.047)(0.060)(0.081)

Bachelor’s degree or higher

0.1570.2100.079-0.096-0.120-0.076-0.046-0.048-0.030
(0.039)(0.046)(0.073)(0.046)(0.060)(0.073)(0.048)(0.062)(0.083)

Household composition

Spouse/partner in household

0.0080.010-0.027-0.036-0.036-0.0280.0290.0270.043
(0.022)(0.026)(0.045)(0.023)(0.027)(0.045)(0.025)(0.030)(0.049)

Children under age 18 in household [4]

Only children under age 6

0.2480.2200.287-0.082-0.079-0.098-0.186-0.159-0.212
(0.038)(0.052)(0.056)(0.033)(0.045)(0.048)(0.032)(0.045)(0.047)

Children both under age 6 and ages 6 to 17

0.008-0.0160.037-0.011-0.0420.016-0.0070.051-0.064
(0.019)(0.025)(0.032)(0.019)(0.025)(0.030)(0.022)(0.029)(0.034)

Geography at round 19 interview

Urban

-0.066-0.057-0.0700.0490.0570.0510.017-0.0040.026
(0.025)(0.032)(0.043)(0.021)(0.028)(0.034)(0.027)(0.035)(0.042)

County-level change in activity at workplaces

0.0060.0060.005-0.008-0.009-0.0070.0020.0020.000
(0.001)(0.002)(0.002)(0.001)(0.002)(0.002)(0.002)(0.002)(0.003)

State fixed effect

YesYesYesYesYesYesYesYesYes

Sample size

2,8191,7411,0782,8191,7411,0782,8191,7411,078

[1] NLSY97 is the National Longitudinal Survey of Youth 1997, and COVID-19 is coronavirus disease 2019.

[2] AFQT is the Armed Forces Qualification Test.

[3] The GED is the General Educational Development test.

[4] Only children ages 6 to 17 is the omitted category.

Note: The regressions in the table are weighted regressions. Specifications include a control for mode of interview and dummy variables indicating measures are missing for AFQT score, highest degree completed, and urbanicity. Coefficient estimates are reported, and robust standard errors are in parentheses.

Source: U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1997.

Controlling for other characteristics, we find that women were 6 percentage points less likely to have children with only in-person schooling and 4 percentage points more likely to have children with only remote schooling. These differences closely correspond to the differences in schooling mode seen in the descriptive statistics reported in table 1. Women and men have a similar likelihood of having children in their household receiving hybrid schooling (both in-person schooling and remote schooling).

Black, non-Hispanic individuals were 17 percentage points less likely than non-Black, non-Hispanic individuals to have children attend school only in person and 21 percentage points more likely to have children attend only remote school. Hispanic individuals were 10 percentage points less likely than non-Black, non-Hispanic individuals to have children attend school only in person and 12 percentage points more likely to have children attend only remote school. Data from the NAEP Monthly School Survey on the reopening of schools and attendance also indicate that Black and Hispanic students were more likely to be enrolled in remote learning than White students during the spring semester of 2021.27 Andrew Camp and Gema Zamarro use data from a COVID-19 tracking survey to find that a combination of factors contribute to the in-person schooling racial gap.28 They find that schooling mode offerings at the school district level, local COVID-19 outbreaks, and election polling responses for the 2020 presidential election have large statistically significant effects on the probability of a student attending in-person or remote schooling during the fall of 2020. 

Individuals with more education were more likely to have children only attending school in person. This result is mostly driven by women. Specifically, women with some college or with a bachelor’s degree (compared with women with less than a high school diploma) were more likely to have children only attend school in person by 11 and 21 percentage points, respectively. Education level did not have an effect on the likelihood of hybrid schooling mode for women or men.

Individuals in households that included only children under age 6 were more likely to have children attending only in-person schooling (25 percentage points), and they were less likely to have children attending only remote schooling (8 percentage points) and hybrid schooling (19 percentage points). (See table 7.) These distinctions in schooling mode may reflect attempts by many school districts to offer full-time in-person schooling to their youngest learners while offering hybrid or remote-only schooling to older students.29

Location is related to schooling mode. Individuals who resided in urban areas at the round 19 interview were 7 percentage points less likely to have children attending school only in person and 5 percentage points more likely to have children attending school only remotely in the week before the COVID-19 Supplement interview. Change in county-level activity at workplaces is positively associated with the likelihood of having children attending school only in person and negatively associated with the likelihood of having children attending school only remotely. An increase in county-level activity by 9 percent (a little over a standard deviation) was associated with a 5-percentage-point increase in the likelihood of having children attending school only in person and a 7-percentage-point reduction in the likelihood of having children attending school only remotely. This finding likely reflects the fact that restrictions on in-person schooling and work activities go hand in hand. These correlations also might reflect that individuals are reluctant both to send their kids to school and to work in person when the incidence of COVID-19 in a county is high.

Employment, telework, and child schooling

The prior section shows that demographic characteristics, child age, and location are related to schooling mode. We now examine whether schooling mode is related to the likelihood of work and telework.

Chart 3 shows the probability of working in the prior week by schooling type and sex using data from the HPS and the NLSY97 COVID-19 Supplement. Both the HPS and the NLSY97 COVID-19 Supplement show that the likelihood of men working in the 1980–84 cohort did not vary by schooling type. In contrast, women whose children attended only in-person schooling were more likely to work than were women whose children attended only remote schooling (78 percent compared with 71 percent in the NLSY97 COVID-19 Supplement and 77 percent compared with 57 percent in the HPS). However, we find that the effect of in-person schooling disappears when we estimate a regression equation that includes schooling mode along with the explanatory variables listed in table 2. This finding indicates that the relationship depicted in chart 3 is likely due to other variables that are correlated with schooling mode.

Chart 4 displays the probability of full telework (conditional on working) by schooling type and sex using data from the NLSY97 COVID-19 Supplement only. Both men and women who had children who attended only in-person schooling were much less likely to have full telework than those who had children who attended only remote schooling. Men with children who attended only in-person schooling were 12 percentage points less likely to telework all the time. Women with children who attended only in-person schooling were 20 percentage points less likely to telework all the time. The positive effect of remote schooling on teleworking remains when we estimate a regression equation that includes schooling mode along with the explanatory variables listed in table 2. This finding suggests that remote schooling itself may make teleworking more likely.

However, the potential endogeneity of schooling modes complicates the analysis of the effect of schooling mode on both the probability of working and the probability of teleworking. That is, the choice of schooling mode may depend on unmeasured factors, such as unobserved geographic, household, or child characteristics, that also affect the likelihood of work and telework. Thus, one cannot interpret the effects of schooling mode as causing the work or telework outcomes. A detailed analysis of the effects of schooling mode on work and telework is beyond the scope of this article. But we are exploring this topic further in our ongoing work.

Discussion and conclusion

The repercussions of the COVID-19 pandemic on employment and children’s schooling continued well into spring 2021 when the National Longitudinal Survey of Youth 1997 (NLSY97) administered the NLSY97 COVID-19 Supplement interview. Descriptive statistics from the supplement show that a substantial percentage of working men and women were teleworking some hours (46 percent) or all hours (25 percent) in the week before the supplement interview. The descriptive statistics show that 30 percent of parents had children under age 18 participating in only remote schooling, 31 percent had children participating in only in-person schooling, and 36 percent had children participating in hybrid schooling (both in-person and remote).

Estimation results from the supplement data suggest that working at a job that was conducive to telework (few physical tasks and limited face-to-face contact) at the earlier round 19 interview increased women’s likelihood of working in the week before the NLSY97 COVID-19 Supplement interview. Having a prior job conducive to telework increased the likelihood of any and full telework for both men and women. Controlling for observed job characteristics, we find that men and women with a bachelor’s degree or higher were much more likely to telework than those with lower levels of education.

Data from both the NLSY97 COVID-19 Supplement and the Household Pulse Survey show that women born from 1980 to 1984 were more likely than men of a similar age to have children attending only remote schooling in spring 2021 and were less likely than men of a similar age to have children attending only in-person schooling. Black and Hispanic individuals born from 1980 to 1984 were more likely than non-Black, non-Hispanic individuals born from 1980 to 1984 to have children attending only remote schooling and less likely to have children attending only in-person schooling. Linear probability estimation with the NLSY97 COVID-19 Supplement data shows that these results remain with controls for background variables and state fixed effects. Residing in an urban area increased the likelihood of only remote schooling and decreased the likelihood of only in-person schooling. Individuals with children only under age 6 were more likely to have children attending only in-person schooling and much less likely to have children attending only remote schooling. Exploratory descriptive statistics imply that those living with children who were only attending remote schooling were more likely to have full telework.

The effects of COVID-19 and accompanying restrictions reduced activity at workplaces at the beginning of the pandemic. County-level activity at workplaces in spring 2021 was still substantially below its prepandemic level in many locations. Data from the NLSY97 COVID-19 Supplement show that reduced county-level activity at workplaces was associated with a reduced probability that women were working in spring 2021. In our regression results, we find that reduced county-level activity at the workplace was associated with an increase in the likelihood of teleworking for both men and women. This result reflects the fact that lessened activity at the workplace during the pandemic was generally associated with increased remote work since many individuals switched from working in the office to working at home. Our results also indicate that a decline in county-level activity at the workplace was associated with an increase in the likelihood of only remote schooling and a concomitant decrease in the likelihood of only in-person schooling.

The next NLSY97 data release will include the 2021–22 interview. In the 2021–22 interview, the survey collected information about employment since the date of the 2019–20 interview. In addition, the 2021–22 data will include information on health, program participation, income, family, and much more. Combining these data with earlier data on respondents’ employment histories and background will allow researchers to examine in greater detail how the pandemic affected people’s lives.

Suggested citation:

Alison Aughinbaugh, Jeffrey A. Groen, Mark A. Loewenstein, Donna S. Rothstein, and Hugette Sun, "Employment, telework, and child remote schooling from February to May 2021: evidence from the National Longitudinal Survey of Youth 1997," Monthly Labor Review, U.S. Bureau of Labor Statistics, March 2023, https://doi.org/10.21916/mlr.2023.5

Notes


1 The Employment Situation: March 2020, USDL-20-0521 (U.S. Bureau of Labor Statistics, April 3, 2020), https://www.bls.gov/news.release/archives/empsit_04032020.pdf; The Employment Situation: April 2020, USDL-20-0815 (U.S. Bureau of Labor Statistics, May 8, 2020), https://www.bls.gov/news.release/archives/empsit_05082020.pdf.

2 The Employment Situation: April 2021, USDL-21-0816 (U.S. Bureau of Labor Statistics, May 7, 2021), https://www.bls.gov/news.release/archives/empsit_05072021.pdf.

3 Matthew Dey, Harley Frazis, David S. Piccone Jr, and Mark A. Loewenstein, “Teleworking and lost work during the pandemic: new evidence from the CPS,” Monthly Labor Review, U.S. Bureau of Labor Statistics, July 2021, https://doi.org/10.21916/mlr.2021.15.

4 The percentage of workers teleworking because of the pandemic was 21.5 percent for workers ages 35 to 44. This group of workers, ages 35 to 44, is similar in age to the workers analyzed in this present article. “Effects of the coronavirus COVIDÔÇÉ19 pandemic (CPS),” Labor Force Statistics from the Current Population Survey (U.S. Bureau of Labor Statistics, last modified on November 2, 2022), https://www.bls.gov/cps/effects-of-the-coronavirus-covid-19-pandemic.htm.

5 “The coronavirus spring: the historic closing of U.S. schools (a timeline),” Education Week, July 1, 2020, https://www.edweek.org/leadership/the-coronavirus-spring-the-historic-closing-of-u-s-schools-a-timeline/2020/07.

6 Ibid.

7 “Monthly School Survey dashboard,” Monthly School Survey (Institute of Education Sciences, U.S. Department of Education), https://ies.ed.gov/schoolsurvey/mss-dashboard/.

8 “Appendix 14: NLSY97 COVID-19 Supplement,” National Longitudinal Surveys (U.S. Bureau of Labor Statistics), https://www.nlsinfo.org/content/cohorts/nlsy97/other-documentation/codebook-supplement/appendix-14-nlsy97-covid-19.

9 The estimation sample contains 5,238 interviews. In this sample, 453 (9 percent) of round 19 interviews of the National Longitudinal Survey of Youth 1997 (NLSY97) were conducted in March 2020 or later. The inclusion of a dummy variable for whether a round 19 interview was conducted in March 2020 or later does not change the estimation results. And the dummy variable is not statistically significant in the regressions.

10 “Appendix 14: NLSY97 COVID-19 Supplement.”

11 Respondents could answer the NLSY97 Supplement survey on a website designed for this purpose (85 percent completed via web) or with an interviewer via telephone. This was the first time that NLSY97 respondents had a web option. For more information about the COVID-19 Supplement, see Alison Aughinbaugh and Donna S. Rothstein, “How did employment change during the COVID-19 pandemic? Evidence from a new BLS survey supplement,” Beyond the Numbers: Employment & Unemployment, vol. 11, no. 1 (U.S. Bureau of Labor Statistics, January 2022), https://www.bls.gov/opub/btn/volume-11/how-did-employment-change-during-the-covid-19-pandemic.htm; and “Appendix 14: NLSY97 COVID-19 Supplement.”

12 The descriptive statistics and regressions in the tables in this article use NLSY97 weights created with the National Longitudinal Surveys (NLS) custom weighting program. The weights are designed so that estimates from the current sample can be taken as representative of the cohort born from 1980 to 1984 and living in the United States in 1997. “NLS Youth 1997,” National Longitudinal Surveys (U.S. Bureau of Labor Statistics), https://www.nlsinfo.org/weights/nlsy97.

13 Individuals with only homeschooled children are not included.

14 Education in a pandemic: the disparate impacts of COVID-19 on America’s students (Office for Civil Rights, U.S. Department of Education, June 2021), https://www2.ed.gov/about/offices/list/ocr/docs/20210608-impacts-of-covid19.pdf.

15 “Monthly School Survey dashboard.”

16 We limit the sample in the Household Pulse Survey (HPS) to those born from 1980 to 1984 to be comparable with the NLSY97. We also limit the HPS sample to data collected from April 14 to June 7, 2021.

17 The Armed Forces Qualification Test (AFQT) covers four sections of the Armed Services Vocational Aptitude Battery and measures math and verbal aptitude. This test was given to NLSY97 respondents in 1997–98.

18 If the individual was not working at a job at the round 19 interview, then the “job is conducive to telework” variable is set to 0 and the “not working at round 19” variable is set to 1.

19 Instead of using information in the survey on the job tasks that workers perform, we could have used an Occupational Information Network measure based on their reported occupation, as in Dey et al., “Teleworking and lost work during the pandemic.” Both measures give similar results in our analysis. For simplicity, we have chosen to use only the task-based measure in this article but are exploring the two measures further in ongoing work. Our results show that the measures are correlated but that each has relevant information as to whether a job is conducive to telework.

20 “See how your community moved differently due to COVID-19,” COVID-19 Community Mobility Reports (Google), no longer updated as of 10/15/2022, https://www.google.com/covid19/mobility/.

21 Consistent with the patterns in the NLSY97 COVID-19 Supplement, data from the March 2019 Current Population Survey show that 96 percent of children lived with a mother figure and 74 percent lived with a father figure. These data show that the percentage of children who live with only a mother increases with age. Lydia R. Anderson, Paul F. Hemez, and Rose M. Kreider, “Living arrangements of children: 2019,” Household Economic Studies (U.S. Census Bureau, February 2022), https://www.census.gov/content/dam/Census/library/publications/2022/demo/p70-174.pdf.

22 Calculation of 63 percent: coefficient on the “not working at round 19” variable in table 3 (–0.487) divided by the percentage of the sample who worked last week (77 percent, as shown in table 1), which is –0.487/0.77 = –0.632.

23 The child age variables are a set of mutually exclusive dummy variables. The estimated effects of child age are relative to the excluded category (dummy variable).

24 Even without disruptions in childcare, mothers with young children are expected to work less; however, note that we include a control for previous employment in our regression equations.

25 Cevat Giray Aksoy, Jose Maria Barrero, Nicholas Bloom, Steven J. Davis, Mathias Dolls, and Pablo Zarate, “Working from home around the world,” Brookings Papers on Economic Activity (Brookings Institution, September 7, 2022), https://www.brookings.edu/bpea-articles/working-from-home-around-the-world/.

26 Education in a pandemic.

27 School responses to COVID-19 (Institute of Education Sciences, U.S. Department of Education), https://ies.ed.gov/schoolsurvey/.

28 Andrew Camp and Gema Zamarro, “Determinants of ethnic differences in school modality choices during the COVID-19 crisis” (EdWorkingPaper: 21-374, Annenberg Institute at Brown University, April 2021), https://doi.org/10.26300/pmyy-nh92.

29 Eliza Shapiro and Kate Taylor, “Why school districts are bringing back younger children first,” New York Times, November 30, 2020, https://www.nytimes.com/2020/11/30/nyregion/elementary-schools-reopening.html.

 

article image
About the Author

Alison Aughinbaugh
aughinbaugh.alison@bls.gov

Alison Aughinbaugh is a research economist in the Office of Employment and Unemployment Statistics, U.S. Bureau of Labor Statistics.

Jeffrey A. Groen
groen.jeffrey@bls.gov

Jeffrey A. Groen is a research economist in the Office of Employment and Unemployment Statistics, U.S. Bureau of Labor Statistics.

Mark A. Loewenstein
loewenstein.mark@bls.gov

Mark A. Loewenstein is a senior research economist in the Office of Employment and Unemployment Statistics, U.S. Bureau of Labor Statistics.

Donna S. Rothstein
rothstein.donna@bls.gov

Donna S. Rothstein is a research economist in the Office of Employment and Unemployment Statistics, U.S. Bureau of Labor Statistics.

Hugette Sun
sun.hugette@bls.gov

Hugette Sun is a research economist in the Office of Employment and Unemployment Statistics, U.S. Bureau of Labor Statistics.

close or Esc Key