Wireless Substitution: Early Release of Estimates from the National Health Interview Survey, January – June 2021pdf icon

NATIONAL CENTER FOR HEA LTH STATISTICS

National Health Interview Survey Early Release Program

Wireless Substitution: Early Release of Estimates from the National Health Interview Survey, January-June 2021

Stephen J. Blumberg, Ph.D., and Julian V. Luke

Division of Health Interview Statistics, National Center for Health Statistics

Overview

Since 2007, the National Health Interview Survey (NHIS) Early Release Program has regularly released preliminary estimates of the percentages of adults and children living in homes with only wireless telephones (also known as cellular telephones, cell phones, or mobile phones). These estimates are the most up- to-date estimates available from the federal government concerning the size and characteristics of this population.

Estimates in this report are based on the first six months of 2021. During this time period, 68.0% of adults and 79.1% of children lived in wireless-only households.

NHIS data can also be used to estimate the percentage of adults who live in wireless-only households andhave their own wireless telephone (wireless-only adults). For January-June 2021, 67.2% of adults were wireless-only adults. Demographic subgroups with the highest

health care access and utilization. The survey also includes information about household telephones and whether anyone in the household has a wireless telephone.

To provide access to the most recent information from NHIS, estimates using the January-June 2021 data are being released prior to final data editing and final weighting. These estimates should be considered preliminary. Estimates produced using the final data files may differ slightly from those presented here.

Background

Many health surveys, political polls, and other types of research are conducted using random-digit-dial (RDD) telephone surveys. Most survey research organizations include wireless telephone numbers when conducting RDD surveys.

If they did not, the exclusion of households with only wireless telephones (along with the small proportion of households that have no telephone service) could bias results. This bias- known as coverage bias-could exist if there are differences between people with and without landline telephones for the substantive variables of interest.

Since 2003, NHIS has asked respondents about landlines and wireless telephones in their homes. Compared with adults living in landline households, adults living in wireless-only households are more likely to be people who smoke cigarettes and people who had at least one heavy drinking day in the past year. Adults living in wireless-only households are also less likely to have health insurance coverage, less likely to have a usual place to go for medical care, less likely to have received an influenza vaccination in the

percentages of wireless-only adults include adults aged 25-29 (85.0%) and 30-34 (86.1%) (Figure), and adults renting their homes (81.9%).

NHIS Early Release

Program

This report is published as part of the NHIS Early Release Program. Twice each year, the National Center for Health Statistics (NCHS) releases selected estimates of telephone coverage for the civilian noninstitutionalized U.S. population based on data from NHIS, along with comparable estimates from NHIS for the previous 2 years. The estimates are based on in-person interviews that are conducted throughout the year to collect information on health status, health-related behaviors, and

Figure. Percentages of wireless-only adults and of children living in households with only wireless telephone service, by age group: United States, January-June 2021

Percent

100

90	79.1		85.0	86.1	79.1
80	75.9
70						65.9
60
50							40.8
40
30
20
10
0
	Under 18	18-24	25-29	30-34	35-44	45-64	65 and over

Age group (years)

NOTES: Wireless-only adults are adults who live in households with only wireless telephone service and have their own wireless telephone. In 2021, data collection procedures for the National Health Interview Survey were modified because of the COVID-19 pandemic. Estimates from 2021 may have been impacted by these changes. See text in this report for more details.

SOURCE: National Center for Health Statistics, National Health Interview Survey.

P a g e | 1

U.S. Department of Health and Human Services ● Centers for Disease Control and Prevention ● National Center for Health Statistics ● Released 11/2021

Wireless Substitution: Early Release of Estimates from the National Health Interview Survey, January-June 2021

past year, and more likely to have experienced financial barriers to care. Previous Early Release Program reports based on data from 2003-2018 include additional details about these health- related differences between people with and without landline telephones.

Because of these differences, the potential for coverage bias remains a real threat to RDD health surveys that do not include sufficient representation of households with only wireless telephones. NCHS continues to publish estimates of the size and characteristics of this population so that survey research organizations can evaluate whether they have appropriately included this population in their telephone surveys.

Methods

NHIS randomly selects one "sample adult" aged 18 years or older and one "sample child" aged 17 years or younger (if any children live in the household) from each household following a brief initial interview that identifies everyone who usually lives or stays in the household. Information about the sample adult is collected from the sample adults themselves unless they are physically or mentally unable to do so, in which case a knowledgeable proxy can answer for the sample adult. Information about the sample child is collected from a parent or adult who is knowledgeable about and responsible for the health care of the sample child. This respondent may or may not also be the sample adult.

To determine whether the sample adult or child lived in a household with a landline telephone, the respondent was asked if there was "at least one phone inside your home that is currently working and is not a cell phone." To avoid possible confusion with cordless landline telephones, the word "wireless" was not used in the survey. This question was asked only once, in whichever interview (sample adult or sample child) came first.

Sample adults are also asked whether they "have a working cell phone," and if not, whether they "live with anyone who has a working cell phone." This approach permits the identification of adults living in wireless-only households (that is, households without landlines within

which at least one household member has a working cell phone) and of wireless-only adults (that is, adults who live in a wireless-only household and have their own cell phone). Respondents for sample children are only asked if the child lives "with anyone who has a working cell phone," and only if the wireless status of the household is not yet known from the sample adult interview.

An additional question is included for sample adults who have a cell phone and live in households with landline telephones. The sample adult is asked to consider "all the telephone calls that you answer" and to report whether "all or almost all [are] on your cell phones, some [are] on your cell phone and some on your home phone, or very few or none [are] on your cell phones." This question permits the identification of "wireless-mostly"adults-defined as adults with both landline and cellular telephones who answer all or almost all calls on cell phones. Landline-mostly adults and dual- users can be similarly identified.

NHIS uses sampling weights to produce representative national estimates. The base weight is equal to the inverse of the probability of selection of the sample address. These weights are adjusted for household and person-level nonresponse using multilevel models predictive of response propensity. Nonresponse-adjusted weights are further calibrated to U.S. Census Bureau population projections and American Community Survey (ACS) one-year estimates for age, sex, race and ethnicity, educational attainment, Census division, and Metropolitan Statistical Area status.

Point estimates and 95% confidence intervals were calculated using SUDAAN software (RTI International, Research Triangle Park, NC) to account for the complex sample design of NHIS. All estimates shown meet the NCHS standards of reliability as specified in National Center for Health Statistics Data Presentation Standards for Proportions.

Differences between percentages were evaluated using two-sided significance tests at the 0.05 level. All differences discussed are statistically significant unless otherwise noted. Lack of comment regarding the difference between any two estimates does not necessarily mean that the difference was

tested and found to be not significant. Because of small sample sizes, estimates based on less than 1 year of data may have large confidence intervals, and caution should be used in interpreting such estimates.

Impact of the COVID-19 Pandemic

Additional caution is warranted when interpreting telephone status estimates from 2020. Due to the COVID- 19 pandemic, NHIS data collection switched to a telephone-only mode beginning on March 19, 2020. Personal visits to households resumed in selected areas in July 2020 and in all areas of the country in September 2020. However, contact with households was still attempted by telephone first, and a majority of interviews were completed by telephone. Additionally, starting in August and continuing through the end of December, a subsample of adult respondents who completed an NHIS interview in 2019 were recontacted by telephone and asked to participate again, completing the 2020 NHIS questionnaire. Estimates for 2020 in Table 1 are based on data from both samples.

Survey weights were adjusted to account for changes in respondent characteristics due to these changes in the 2020 data collection. An evaluation of nonresponse bias following survey weighting is available online. It revealed that the weighted 2020 sample underrepresented adults living alone and adults with family income below the federal poverty level. The sample also underrepresented wireless-only adults, whereas adults living in households with both landline and wireless telephones were overrepresented. Moreover, phoneless households (those with neither wireless nor landline telephones) generally could not be interviewed in Quarter 2, 2020 or as part of the reinterviewed sample. For these reasons, caution should be used in interpreting differences observed in estimates between 2020 and other time periods, particularly for estimates of people living in phoneless households.

The "telephone first" data collection approach that began in July 2020

P a g e | 2

U.S. Department of Health and Human Services ● Centers for Disease Control and Prevention ● National Center for Health Statistics ● Released 11/2021

Wireless Substitution: Early Release of Estimates from the National Health Interview Survey, January-June 2021

continued through April 2021. As a result, a majority of completed NHIS interviews in January-June 2021 were conducted by telephone. However, preliminary analyses of weighted estimates suggest that biases observed in 2020 did not persist in 2021.

Household Telephone Status

From January through June 2021, information on household telephone status was obtained for 13,901 civilian adults aged 18 and over and 4,016 children under age 18. In the first six months of 2021, 68.0% of adults (about 172 million) and 79.1% of children (over 57 million) lived in households that did not have a landline telephone but did have at least one wireless telephone (Table 1).

The percentages of adults and children living in wireless-only households have been generally increasing since 2003. Consistent with that trend, there was also a statistically significant increase (5.5 percentage points) in the past 12 months-that is, from the first 6 months of 2020 to the first 6 months of 2021-in the percentages of adults and children living in wireless-only households.

or non-Hispanic Asian (68.7%) adults to be wireless-only.

• Men (69.3%) were more likely than women (65.1%) to be wireless-only.
• Adults with family incomes below the federal poverty threshold (73.9%) were more likely than adults with family incomes of 100% to less than 200% of the federal poverty threshold (68.3%) and adults with higher family incomes (67.1%) to be wireless-only.
• Adults living in the Midwest (71.1%), South (69.1%), and West (70.6%) were more likely than those living in the Northeast (53.6%) to be wireless- only.
• Four in five adults living in rented homes (81.9%) were wireless-only. This percentage is higher than the percentage for adults living in homes owned by a household member (60.7%).

Table 2 also includes estimates of the percentage of adults who were wireless- mostly, landline-mostly, dual users, landline-only, and phoneless, by selected demographic characteristics. Confidence intervals for these percentages are shown in Table 3.

In addition to these products, preliminary microdata files containing selected NHIS variables are produced as part of the Early Release Program. The telephone service use variables presented in this report have been included in those microdata files. Analysts can access these files through the NCHS Research Data Centers (https://www.cdc.gov/rdc/) without having to wait for the final annual NHIS microdata files to be released.

For more information about NHIS and the NHIS Early Release Program, or to find other Early Release Program products, see

• NHIS home page athttps://www.cdc.gov/nchs/nhis.htm.
• Early Release Program home page athttps://www.cdc.gov/nchs/nhis/relea ses.htm.

Suggested Citation

Blumberg SJ, Luke JV. Wireless substitution: Early release of estimates from the National Health Interview Survey, January-June 2021. National Center for Health Statistics. November 2021. DOI:

https://doi.org/10.15620/cdc:111171.

Wireless-only Adults

Two-thirds of adults were wireless- only (67.2%, 170 million); that is, they personally had a wireless telephone and lived in a household that did not have a landline. The percentage of adults who were wireless-only is shown, by selected demographic characteristics, in Table 2. Confidence intervals for these percentages are shown in Table 3. For January-June 2021:

• Four in five adults aged 25-29 (85.0%) and aged 30-34 (86.1%) were wireless-only(Figure). The percentage of adults who were wireless-only decreased as age increased beyond 35 years: 79.1% for those 35-44; 65.9% for those 45-64; and 40.8% for those 65 and over.
• Hispanic adults (77.4%) were more likely than non-Hispanic White (64.5%), non-Hispanic Black (64.3%),

Other NHIS Early Release Program Products

This report is published as part of the NHIS Early Release Program. Earlier reports on wireless substitution are at https://www.cdc.gov/nchs/nhis/erwireless subs.htm.

The prevalence of adults and children living in wireless-only households varies across states. For more information about prevalence estimates at the state level, see

• NCHS. Modeled estimates (with standard errors) of the percent distributions of personal telephone status for adults aged 18 and over and household telephone status for children under 18, by state: United States, 2019. August 2021. Available from:https://www.cdc.gov/nchs/data/nhis/ earlyrelease/Wireless_state_202108- 508.pdf.

P a g e | 3

U.S. Department of Health and Human Services ● Centers for Disease Control and Prevention ● National Center for Health Statistics ● Released 11/2021

Wireless Substitution: Early Release of Estimates from the National Health Interview Survey, January-June 2021

Table 1. Percent distribution of household telephone status for adults and children: United States, 2019-2021

	Number of people	Wireless-only	Landline with	Landline-only		Landline with	Wireless with
Date of interview	(unweighted)	household	wireless	household1	Phoneless	unknown wireless	unknown landline	Total
Adults
January-June 20192	16,577	59.2	37.2	2.5	1.0	0.0	0.1	100.0
95% confidence interval	…	57.97-60.37	35.99-38.43	2.30-2.79	0.78-1.18	0.02-0.09	0.03-0.12	…
July-December 20192	14,725	61.3	35.2	2.4	1.1	0.0	0.1	100.0
95% confidence interval	…	59.66-62.85	33.71-36.62	2.05-2.69	0.86-1.28	0.01-0.10	0.06-0.31	…
January-June 20203	13,602	62.5	34.3	2.3	0.7	0.1	0.1	100.0
95% confidence interval	…	61.21-63.78	33.03-35.54	2.01-2.62	0.54-0.96	0.03-0.28	0.04-0.25	…
July-December 20203	17,028	65.8	31.9	1.8	0.4	0.1	0.1	100.0
95% confidence interval	…	64.69-66.81	30.88-32.95	1.60-2.08	0.28-0.52	0.03-0.12	0.04-0.14	…
January-June 2021	13,901	68.0	29.3	1.9	0.5	0.0	0.1	100.0
95% confidence interval	…	66.70-69.34	28.04-30.66	1.66-2.23	0.44-0.69	0.01-0.05	0.09-0.22	…
Children
January-June 20192	4,800	70.5	28.0	0.6	1.0	-	-	100.0
95% confidence interval	…	68.67-72.19	26.33-29.68	0.34-1.04	0.68-1.39	-	-	…
July-December 20192	4,375	70.3	27.7	0.6	1.3	-	0.1	100.0
95% confidence interval	…	68.10-72.37	25.76-29.64	0.31-1.29	0.95-1.82	-	0.02-0.38	…
January-June 20203	3,808	73.6	25.1	0.3	0.9	-	0.0	100.0
95% confidence interval	…	71.43-75.73	23.12-27.29	0.15-0.54	0.58-1.38	-	0.00-0.23	…
July-December 20203	1,984	75.5	23.1	0.6	0.6	-	-	100.0
95% confidence interval	…	72.90-77.91	20.84-25.55	0.29-1.37	0.23-1.80	-	-	…
January-June 2021	4,016	79.1	19.8	0.4	0.7	0.0	0.0	100.0
95% confidence interval	…	77.27-80.78	18.12-21.61	0.17-0.91	0.42-1.05	0.00-0.04	0.01-0.23	…

0.0 Quantity more than zero but less than 0.05. … Category not applicable.

- Quantity zero.

1Landline-only refers to households with a landline telephone in which no residents have a working cell phone. In reports based on 2003-2018 data, this category was labeled as "landline without wireless."

2To produce estimates for 2019, the NHIS sampling weights were further adjusted to account for telephone-question nonresponse. These data were missing for 2.9% of sample adults and 0.7% of sample children in the first six months of 2019, and for 3.9% of sample adults and 1.8% of sample children in the second six months of 2019. Although telephone-question nonresponse remains at similar levels, this adjustment was dropped starting in 2020 because sensitivity analyses indicated that the adjustment changed very few resulting estimates, and for those that did change, the impact was no more than two-tenths of one percent. People with both unknown landline and unknown wireless status are not included in the denominators when calculating percentages.

3Due to the COVID-19 pandemic, NHIS data collection switched to a telephone-only mode beginning on March 19, 2020. Personal visits resumed in all areas in September 2020. However, contact with households was still attempted by telephone first, and a majority of interviews were completed by telephone. Additionally, from August-December 2020, a subsample of adult respondents who completed an NHIS interview in 2019 were recontacted by telephone and asked to participate again. Response rates were lower and respondent characteristics were different in April-December 2020. Differences observed in estimates between 2020 and other time periods-particularly estimates of people living in phoneless households-may have been impacted by these differences in respondent characteristics.

NOTE: Data are based on household interviews of a sample of the civilian noninstitutionalized population.

SOURCE: National Center for Health Statistics, National Health Interview Survey.

P a g e | 4

U.S. Department of Health and Human Services ● Centers for Disease Control and Prevention ● National Center for Health Statistics ● Released 11/2021

Wireless Substitution: Early Release of Estimates from the National Health Interview Survey, January-June 2021

Table 2. Percent distribution of personal telephone status for adults, by selected demographic characteristics: United States, January-June 2021

	Wireless-only	Wireless-mostly		Landline-mostly	Landline-only	Phoneless
Demographic characteristic	adults	adults	Dual-users	adults	adults	adults	Unknown1	Total
Total	67.2	15.5	8.1	4.5	3.0	1.2	0.5	100.0
Age (years)
18-24	75.9	16.2	3.9	0.6	1.4	1.4	0.6	100.0
25-29	85.0	10.3	1.7	1.0	0.6	1.1	0.3	100.0
30-34	86.1	9.6	2.3	0.3	0.6	0.8	0.4	100.0
35-44	79.1	13.5	3.5	1.2	0.9	1.1	0.6	100.0
45-64	65.9	19.3	8.7	2.6	1.7	1.3	0.5	100.0
65 and over	40.8	15.6	17.7	15.0	9.2	1.4	0.3	100.0
Race and ethnicity
Hispanic or Latino, any race(s)	77.4	12.0	4.1	1.5	2.0	2.5	0.5	100.0
White, single race, non-Hispanic	64.5	16.4	9.2	5.7	3.1	0.7	0.4	100.0
Black, single race, non-Hispanic	64.3	17.8	7.8	3.9	3.9	1.3	1.0	100.0
Asian, single race, non-Hispanic	68.7	14.9	8.2	2.9	1.8	3.0	0.5	100.0
Other and multiple races, non-								100.0
Hispanic	74.3	9.6	8.1	2.2	3.5	0.9	1.5
Sex
Male	69.3	15.4	7.9	3.3	2.4	1.3	0.4	100.0
Female	65.1	15.6	8.3	5.7	3.5	1.2	0.6	100.0
Education
Some high school or less	67.1	7.9	8.3	5.5	6.8	3.8	0.6	100.0
High school graduate or GED2	66.3	15.0	8.1	5.7	3.2	1.2	0.5	100.0
Some post-high school, no degree	67.3	16.6	7.7	4.6	2.6	0.8	0.4	100.0
4-year college degree or higher	68.0	18.2	8.4	3.0	1.3	0.7	0.5	100.0
Family income relative to
federal poverty threshold3
Less than 100%	73.9	9.6	5.8	3.4	3.9	2.8	0.7	100.0
100% to less than 200%	68.3	12.1	6.3	5.4	5.4	2.2	0.3	100.0
200% or greater	67.1	17.6	8.4	4.2	1.9	0.6	0.3	100.0
Geographic region4
Northeast	53.6	19.8	14.2	6.3	4.5	1.2	0.4	100.0
Midwest	71.1	12.9	7.9	4.6	2.6	0.6	0.3	100.0
South	69.1	15.2	6.5	4.3	3.0	1.4	0.6	100.0
West	70.6	15.2	6.2	3.6	2.1	1.6	0.7	100.0
Metropolitan statistical area status
Metropolitan	67.5	15.8	8.1	4.2	2.7	1.3	0.5	100.0
Not metropolitan	65.2	13.7	8.2	6.7	4.8	1.0	0.5	100.0

See footnotes at end of table.

P a g e | 5

U.S. Department of Health and Human Services ● Centers for Disease Control and Prevention ● National Center for Health Statistics ● Released 11/2021