Elderly Mental Health During COVID-19: The Role of Technology Use

Abstract

This paper studies the mental health impacts of the COVID-19 pandemic on older adults with a particular focus on the role of technology use. We use nationally representative individual-level survey data from a random sample of 3257 community-dwelling older Americans (65+) from the National Health and Aging Trends Study (NHATS) COVID-19 supplement (2020). We empirically estimate the impact of COVID-19 (under three alternative specifications) on the five mental health outcomes while controlling other confounding factors. The mental health indicators include feelings of loneliness, anxiety, feeling sad or depressed, poor sleep quality, and not feeling hopeful about the future. The confounding control factors in our empirical model are age, gender, race, financial difficulty, physical health, area of residence, and living arrangement. We further examine whether the mental health effects differ between the technology users and the non-users of technology. Our findings indicate a lower likelihood of adverse health outcomes among technology users (compared to the non-users of technology) during the pandemic. These results validate the potential use of technology for healthcare delivery, mental health therapy, and other lifestyle interventions to improve the quality of life in the older population.

1. Introduction

The novel coronavirus (SARS-CoV-2, commonly called COVID-19) pandemic emerged in late 2019 and ravaged the world during 2020. According to the Centers for Disease Control and Prevention (CDC) data (Centers for Disease Control and Prevention, “Provisional COVID-19 Death Counts by Sex, Age, and State” (originally published, 23 September 2020)), in the United States, the pandemic created disproportionate adverse effects for the elderly (above 65 years of age) in severe symptoms, higher hospitalization rates, and more significant mortality risks. The CDC estimates of the COVID-19 disease burden (February 2020 to March 2021) reflect the following: 49% of infections were in the age group of 18–49 years, with a hospitalization rate of 24%. In contrast, there were 10% infections among the 65+ individuals with a hospitalization rate of 47% (Estimated Disease Burden of COVID-19 | CDC). Pre-COVID-19, older individuals already suffered more from psycho-social and environmental vulnerability issues [1]. During COVID-19, these issues were exasperated. The pandemic’s scale, uncertainty, and fear, coupled with stay-at-home orders, limited social contact, and public information overload, set the stage for a mental health crisis among the older population.

Pre-COVID-19 studies suggest older adults are more resilient to anxiety, depression, and other stress-related mental health conditions [2]. Similar results have been reported among community-dwelling older adults (age 60–80) during the initial phase of COVID-19 in several high-income countries like Spain, Canada, and the Netherlands [3,4]. A survey-based study in Austria during COVID-19 reported the highest mental health problems among adults of less than 35 years of age [5,6]. In a survey report published in August 2020, among 5412 community-dwelling American adults across different age groups, the 933 respondents aged 65+ older reported significantly lower percentages of an anxiety disorder (6.2%), depressive disorder (5.8%), or a trauma or stress-related disorder (TSRD) (9.2%) than participants in younger age groups [7,8]. However, these results are based on studies conducted the first few months into the pandemic. The higher mental resilience of older individuals pre-COVID-19 are attributed to several psycho-social factors that are supported by some theories and evidence. The socioemotional selectivity theory posits that older individuals are selective in their social interactions and in their emotional investment in relationships [9]. This selective nature of interactions helps them to balance their emotions better and build mental fortitude. Research suggests that older individuals encounter additional stressful situations over their lifespans and cultivate stronger coping mechanisms, such as adhering to routines or engaging in stress-releasing hobbies [10]. The cumulative effect of such coping measures strengthens their mental toughness. Additional factors, such as stronger social networks, participation in community or religious activities, volunteering, and staying in touch with friends and family, often lead to better mental health outcomes in older adults [11]. Positive lifestyle factors like a healthy diet and adequate sleep quality in older adults are also linked to positive mental health outcomes in older individuals [5].

But the long-term effects of COVID-19 on elderly mental health, especially in countries like the U.S. (with extremely high disease incidence and deaths) remain unclear. In our empirical research, we look at nationally representative survey data (collected through January 2021) to study the impact of COVID-19 on the mental health of the elderly, with a particular focus on the role of technology use. Our paper makes a significant potential contribution to the existing literature on the mental health impact of the pandemic. Although some studies have established the mental health resilience of older adults in the early phase of COVID-19, it is vital to investigate the mental health effects in the later stages. COVID-19 pandemic outcomes in the United States have evolved; hence it is crucial to analyze individual-level data as they become available. Also, it is essential to understand the impact for different subgroups of the older population (based on gender, race, financial status, or physical health status) because some population-based results may not be generalized. Finally, the initial concern about adverse mental health effects of the pandemic on older individuals stems from the isolation and limited social contact imposed by stay-at-home orders (or quarantines). But older individuals may have countered some negative impacts due to several factors (like a healthier lifestyle, preventive measures, or the use of technology). Technology use could be a great way to maintain social connectedness or access healthcare (via telehealth) during the pandemic. Our research mainly focuses on the technology used by the elderly. We empirically examine in a cross-sectional random sample (hence, our results are not plagued by sampling biases that may arise in self-selecting samples of older adults who are recruited via convenience sampling and participant members of community groups) of community-dwelling older individuals from a nationally representative survey data whether the mental health effects of the pandemic are different for the technology users versus the non-users of technology.

2. Literature Review

Many factors affect an individual’s general physical and mental health. Factors like social isolation and loneliness play a significant role in the overall health status of older people. Pre-COVID-19 evidence already suggests that social isolation is detrimental to health and wellbeing. Approximately ¼ of community-dwelling adults are considered to suffer from social isolation, and 43% feel lonely [12]. A pre-COVID-19 study reported that respondents with higher levels of social isolation were at higher risk for health risks, such as increased systolic blood pressure, infection, impaired cognitive function, depression, and mortality. High feelings of loneliness were associated with a higher risk of elevated blood pressure, increased hypothalamic–pituitary–adrenocortical activity, diminished immunity, the progression of Alzheimer’s disease, depressive symptoms, sleep issues, and mortality [13]. Before the COVID-19 pandemic, mental health awareness steadily increased for all age demographics. The age and gender lags in initial technology adoption continued to converge across all demographic groups, as noted in several studies [14,15].

Although there had been significant strides before the COVID-19 pandemic to make healthcare delivery readily available through differing forms of technology, the ease, convenience, comprehension, and acceptance (especially from the older population) were still not universal. The need for this improved technological avenue to relieve the adverse mental health effects of loneliness, social isolation, and other variables on the elderly is now evident in part due to the COVID-19 pandemic. Online technology seems to be more needed than ever before to combat the social disconnectedness imposed by COVID-19 contagion control measures. Recent research [16,17,18] suggests that older adults may gain significant mental health benefits from available health resources through emerging modern technologies, mainly because this population is becoming more Internet savvy. In response to increased demands for mental health interventions that are accessible and affordable, there has been a recent rise in the number of digital mental health interventions (DMHIs) that are employed for mental health treatments of older adults [19]. In addition, technology-enhanced interventions for older adults are helpful for general wellness activities (i.e., exercise) and specifically enhance mental health. Although there are widespread mental health benefits associated with modern technology use, older adults face several barriers to using technology [20]. These barriers include limited internet access, unaffordability, unavailability, inability to use smartphones [21] and needing extensive assistance from others [22].

However, the role of technology in mitigating adverse mental health effects during the pandemic cannot be denied. More specifically, only one-tenth of patients in the U.S. used a telehealth option; 75% of patients did not know that a telehealth option existed. By March 2020, telehealth video care had increased by 80%. In March 2020, the United States government waived HIPAA regulations to improve the ease of access to telehealth due to the restrictions put in place, allowing for the emergence of new forms of digital therapy. A study found that telehealth appointments yielded more benefits for some individuals with depressive symptoms than traditional office visits [23]. These programs proved to be effective due to the scalability available to each patient. Other effective forms of technology use included teleconferencing and video conferencing primarily to interact with family members, friends, or cognitive behavioral therapies [24].

Motivated by the existing literature, in this paper, we have used “use of telehealth” during the pandemic as a definition of technology use. We have also used “learning new online technology” during the pandemic to define technology use alternatively. Our results indicate that the users of such technology experience a lower probability of adverse mental health outcomes during the pandemic than the non-users of technology. These results highlight the need for government health policies to focus on eliminating the barriers faced by the elderly in adopting technology-based interventions in healthcare delivery or medical treatments.

3. Data Description

The empirical analysis for our study is based on publicly available data from the National Health and Aging Trends Study (NHATS) COVID-19 supplement survey, 2020. The NHATS (NHATS is supported by the National Institute on Aging under a cooperative agreement with the Johns Hopkins University Bloomberg School of Public Health (U01AG032947), with data collection by Westat) is a rich source of annual information for later-life functioning among Medicare beneficiaries ages 65 and older. In 2020, the National Health and Aging Trends Study (NHATS) conducted a supplemental mail study about participants’ experiences during the COVID-19 outbreak. NHATS participants who completed a Sample Person (S.P.) interview in Round 10 (2020) were eligible to receive the self-administered COVID-19 questionnaire. The questionnaires were mailed from the end of June through October 2020, and data collection continued through mid-January 2021. Most questionnaires were completed during July–August 2020. The response rate for the COVID-19 supplement was 83.5%. Our sample consisted of 3257 participants (1887 females and 1370 males) above 70 years of age. Out of these 1493 individuals (i.e., 45.8%) were in the range of 70–79 years at the survey time. In addition, 92.8% of the participants were community-dwelling, and 75.9% comprised non-Hispanic whites. In our sample, about 80% of the individuals reported residing in a metropolitan area, and 12.8% of individuals lived in a retirement community. Altogether, 10% of the sample reported experiencing financial difficulty during the pandemic (either in terms of the depletion of savings, reduced income, an inability to pay bills, taking loans, or receiving help from others) (this is the only indicator of economic status that is available in the dataset. There is no information available for real income), and 20.4% self-reported poor physical health status during the pandemic.

This survey dataset is appropriate for our study because it contains a nationally representative sample of older Americans. Besides the demographic information, the data also includes self-reported information on some mental health indicators during the pandemic—like loneliness, anxiety, quality of sleep, feeling sad or depressed, and not feeling hopeful about the future. Additionally, it contains information about technology use during the pandemic—emails/texts, video calls, telehealth, online attendance of religious services, or group activities.

Table 1. Variable definitions and descriptive statistics.

	VARIABLE DEFINITION	Mean	Std. Dev	Max	Min
MENTAL HEALTH VARIABLES
Feeling Lonely	1 if the individual reports feeling lonely at least somedays in a typical week AND 0 otherwise.	0.318	0.466	1	0
Have Poor Quality of Sleep	1 if the individual reports fair or poor quality of sleep during a typical week AND 0 otherwise.	0.405	0.491	1	0
Feeling Worried or Anxious	1 if the individual reports being moderate or severely worried/anxious during a typical week AND 0 otherwise.	0.276	0.447	1	0
Feeling Sad or Depressed	1 if the individual reports being moderate or severely sad/depressed during a typical week AND 0 otherwise.	0.221	0.415	1	0
Not Hopeful about the Future	1 if the individual reports not feeling hopeful at least somedays during a typical week AND 0 otherwise.	0.654	0.476	1	0
EXPLANATORY VARIABLES
Male	1 if the individual is Male AND 0 otherwise	0.421	0.494	1	0
White	1 if the individual is Non-Hispanic, White AND 0 otherwise	0.759	0.428	1	0
Black	1 if the individual is Non-Hispanic Black AND 0 otherwise	0.167	0.373	1	0
Lives in Metropolitan Area	1 if the individual lives in Metropolitan Area AND 0 otherwise	0.801	0.399	1	0
Lives in Retirement Community	1 if the individual lives in Retirement Community AND 0 otherwise	0.128	0.335	1	0
Age (70–79 years)	1 if the individual is in the Age Group (70-79 years) AND 0 otherwise	0.458	0.498	1	0
Financial Difficulty	1 if the individual reports facing Financial Difficulty AND 0 otherwise	0.100	0.300	1	0
Poor Health	1 if the individual reports Poor or Fair health status AND 0 otherwise	0.204	0.403	1	0
COVID_19 Incidence	1 if the individual reports being tested positive or diagnosed having COVID-19 symptoms AND 0 otherwise	0.052	0.223	1	0
COVID_19 Restrictions	1 if the individual experienced restrictions imposed by COVID-19 and 0 otherwise	0.657	0.247	1	0
COVID_19 Continues to Affect Life	1 if the individual reports COVID-19 pandemic continues to affect life and 0 otherwise.	0.925	0.262	1	0
TECHNOLOGY USE VARIABLES
Used Telehealth	1 if the individual used Telehealth or Video for doctor appointments during the pandemic AND 0 otherwise	0.299	0.420	1	0
Learned New Technology	1 if the individual Learned New Technology during the pandemic AND 0 otherwise	0.229	0.458	1	0
Attended Religious or Group Activities Online	1 if the individual attended Religious or Group Activities Online during the pandemic AND 0 otherwise.	0.242	0.428	1	0

represents the definitions and descriptive statistics of the variables used in our analysis.

Mental health Outcomes: We used five adverse mental health indicators as our study’s outcome variables. These indicators are—feeling lonely, having poor sleep quality, feeling worried or anxious, feeling sad or depressed, and not feeling hopeful about the future during the COVID-19 pandemic. In our sample, 31.8% reported feeling lonely every day, most days or at least some days in a typical week. 40.5% of individuals reported having poor sleep quality during a typical week, i.e., sleeping little for a short time or taking more than 30 min to fall asleep and having a hard time getting back to sleep on waking up. A total of 27.6% of individuals reported feeling worried or anxious almost every day (during the day and night) or some of the time on more than half the days during a typical week. Similarly, 22.1% of our sample reported feeling sad or depressed almost every day (during the day and night) or some of the time on more than half the days during a typical week. Lastly, about 65% of individuals reported never or rarely feeling hopeful about the future in a typical week during the pandemic.

COVID-19-related Indicators: We used three indicators to identify the impact of the COVID-19 pandemic and named them: COVID-19 Incidence, Covid_19 Restrictions, and Covid_19 Continues to Affect Life. A total of 171 individuals (i.e., 5.2%) in our sample had had a COVID-19 infection (as diagnosed by a doctor, or a positive test result), 214 individuals (i.e., 6.6%) reported experiencing COVID-19-imposed restrictions (like quarantine, no visitation with family or friends, no group meals or activities) at the place where they live. In contrast, 2956 (i.e., 92.7%) reported that the pandemic continues to affect their lives. These indicators were self-reported by the respondent through a mailed in survey conducted between June 2020 and mid-January 2021, where most people responded during July–August 2020. We use the three alternative indicators to investigate a broad range of impacts of the pandemic on individuals irrespective of whether they got infected or not.

Technology-use Indicators: Again, we used three different definitions of technology use to stratify our sample into technology users and non-users. Using these multiple ways of defining technology use helped us check the robustness of our results. The first indicator refers to learning a new online technology during the pandemic, either by oneself or with help from others—about 23% of the individuals in our sample reported learning a new online technology during the pandemic (however, the survey does not provide details about the new technology learned. The survey questionnaire asks respondents whether they learned a new technology either by themselves or were taught by others). The second indicator measures using telehealth or video for doctor appointments during the pandemic—about 30% of our sample reported using telehealth during the pandemic. The third indicator defines technology use through attendance of religious and group activities online—nearly 24% of the individuals reported participating in religious or group activities online during the pandemic.

Besides looking at these outcomes and indicators during the pandemic, the survey throws some light on how they compare to before the outbreak of COVID-19 (however, the survey does not provide precise numbers for the adverse mental health outcomes or technology use indicators before the outbreak of COVID-19). As for the mental health outcomes, 590 individuals (i.e., 18.11%) reported feeling lonely “MORE OFTEN” in a typical week before the COVID-19 outbreak started, 216 (i.e., 6.63%) reported sleeping “WORSE THAN” a typical week before the onset of COVID-19, and 401 individuals (i.e., 12.31%) felt hopeful “LESS OFTEN” in a typical week before the start of COVID-19 episode. This indicates a toll on the mental health of the elderly due to the COVID-19-related life changes. In terms of technology use, before the outbreak of COVID-19, the data reflects a significant surge in technology use, as represented in

Table 2. Technology use comparison before and during the COVID-19 pandemic.

VARIABLES	PRE-COVID-19		DURING COVID-19
TECHNOLOGY USE	Frequency	Percentage	Frequency	Percentage
Used Email/Text Portal to Communicate with Doctor	519	15.93	642	19.71
Used Telehealth/Video to Communicate with Doctor	137	4.20	974	29.9
Attended ANY Activities Online	258	7.92	1017	31.22
Shopped Groceries Online (Self)	88	2.70	170	5.21
TOTAL	1002	30.76	2803	86.06

Note: The numbers are reported based on the self-reports of the same 3257 community-dwelling survey participants. The participants responded to the mailed self-administered survey questionnaire during Round 10 of NHATS in 2020.

. Given in Table 2, we see an over 25% rise in the use of telehealth and an above 23% increase in online activity involvement. Please note that these comparisons are based on the self-reports of the 3257 survey participants (the survey comparisons qualitatively match up to nationally representative data collected through the Household Pulse Survey (reported by Kaiser Family Foundation) during the peak of the pandemic, 2020–2021, where one in four (24%) older adults (65+) reported anxiety and depression, which was significantly higher than the one in ten (11%) who reported this before the onset of the pandemic (2018) “https://www.kff.org/mental-health/one-in-four-older-adults-report-anxiety-or-depression-amid-the-covid-19-pandemic (accessed on 25 January 2026)”. Similarly, the American Association of Retired Persons (AARP), who used nationally representative data to report technology use trends for older Americans in 2021, reported an increase from 6–24% in ordering groceries online and an increase of 20–40% for telehealth or ordering prescriptions as compared to pre-pandemic levels in 2019 “https://www.aarp.org/press/releases/2021-4-21-tech-usage-among-older-adults-skyrockets-during-pandemic.html (accessed on 25 January 2026)”.

A comparison of these measures over time triggered our interest to explore the role of technology use, specifically in the mental health of older individuals.

3.1. Empirical Method

The empirical strategy involves estimating five separate limited dependent variable regressions (probit model) for the different mental health outcomes during the pandemic. Each regression estimates the predicted probability of a mental health issue—(1) feeling lonely, (2) poor sleep quality, (3) feeling worried or anxious, (4) feeling sad or depressed, and (5) not feeling hopeful about the future—controlling for several explanatory variables. These regressions are run for three different specifications (based on the primary COVID-19-related explanatory variable) to understand the impact of the pandemic on the mental health status of the elderly individuals in our sample. In addition, besides the COVID-19-related explanatory variable, the regression model controls for other independent variables, like age category, gender, race, physical health status, residence area, living arrangement, and financial difficulty.

The model to be estimated is given as follows:

Prob (Yij) = α + βC_i + γZ_i + ε_i

“i” represents an individual, and “j” represents a mental health issue. “j” could take a value from 1 to 5 (for each mental health outcome).

In each of the five regressions, Yij = 1 if individual reports having a mental health issue and 0 otherwise. Since the mental health outcome variable is binary in each case, estimating a probit model is appropriate.

In Specification 1: C_i is a binary variable. C_i = 1 if the individual had an incidence of COVID-19 and 0 otherwise.

In Specification 2: C_i is a binary variable. C_i = 1 if the individual experienced restrictions due to COVID-19 and 0 otherwise.

In Specification 3: C_i is a binary variable. C_i = 1 if the individual reports that COVID-19 still affects their life and 0 otherwise.

Z_i represents a vector of the control variables (listed above) that includes characteristics of the individual respondent.

And ε_i is the residual term that lumps the unobserved factors.

These regressions are estimated in a stratified sample of individuals—the technology users versus the non-users of technology. We use three different ways for stratifying the sample into the two groups (technology users and non-users) based on whether an individual (1) learned a new online technology during the pandemic, (2) used telehealth, or (3) attended religious services/group activities online. This sample stratification method helped us understand whether the mental health effects of the pandemic are any different for the two groups—the technology users versus the non-users of technology.

The marginal effects from the probit regressions in the stratified samples (of technology users vs. non-users) in three different COVID-19 specifications for the five mental health outcomes are discussed in the paper’s next section.

3.2. Discussion of Results

Before discussing results, we must caution the readers not to treat the results presented here as causal. The predicted marginal effects shown in the tables (given below) demonstrate an association between the COVID-19 pandemic and mental health outcomes of the community-dwelling elderly. However, we do not make any claims about causation due to potential endogeneity bias in our estimates (endogeneity refers to the fact that an independent variable included in the model is potentially a choice variable and thus correlated with unobservables relegated to the error term. For example, an individual’s mental health and financial difficulty could potentially be influenced by a common factor like retirement wealth. A common method for dealing with this endogeneity is to use instrumental variable (IV) estimation, where “instruments” are variables that have no direct association with the outcome. The COVID-19 supplement dataset that we are working with is extremely limited in the variables available and the number of observations available for most variables. Under these circumstances, we were unable to carry out an IV analysis and so realize that the results presented here are potentially biased). Instead, the readers should draw inferences based on the associative relationships predicted through our regression models.

We have presented results for stratified samples (segregated by technology use during the pandemic) using the Probit model because the true relationship between the binary (0–1) mental health outcome variables and the explanatory variables is inherently non-linear. This implies that the functional form of a linear probability model (using ordinary least squares, i.e., OLS) would be incorrectly specified. However, the choice of the Probit model for our research question over the linear probability model (OLS) is motivated by refining prediction and not causation.

The results in our paper focus on the predicted probability of five adverse mental health outcomes during the pandemic, i.e., (1) feeling lonely, (2) having poor sleep quality, (3) feeling worried or anxious (4) feeling sad or depressed, and (5) not feeling hopeful about the future. In all tables (both panels), Col. 1 to Col. 5 represent these adverse mental health outcomes, respectively. (Each column represents a separate regression where the dependent variable is a mental health outcome, the primary explanatory variable is a COVID-19-related indicator, and the other independent variables are the same.)

Table 3. Mental health impact of COVID-19 incidence (technology users vs. non-users of technology).

(A)
PANEL A: DID NOT LEARN A NEW ONLINE TECHNOLOGY						PANEL B: LEARNED A NEW ONLINE TECHNOLOGY
Explanatory Variables	(1)	(2)	(3)	(4)	(5)	Explanatory Variables	(1)	(2)	(3)	(4)	(5)
	Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future		Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future
Male	−0.143 ***	−0.0701 ***	−0.0750 ***	−0.114 ***	0.0415 **	Male	−0.166 ***	−0.0835 **	−0.0600 *	−0.0743 **	0.0753 **
	(0.0183)	(0.0200)	(0.0174)	(0.0160)	(0.0192)		(0.0347)	(0.0368)	(0.0354)	(0.0310)	(0.0360)
White	0.0533	−0.0225	−0.0555	0.00869	−0.00502	White	−0.0135	0.0275	0.0655	0.0440	0.00270
	(0.0340)	(0.0376)	(0.0339)	(0.0305)	(0.0362)		(0.0735)	(0.0756)	(0.0706)	(0.0621)	(0.0738)
Black	−0.0345	−0.0533	−0.0539	−0.00568	−0.0465	Black	−0.213 ***	−0.0698	0.0119	−0.0154	0.0866
	(0.0394)	(0.0415)	(0.0338)	(0.0347)	(0.0423)		(0.0619)	(0.0826)	(0.0845)	(0.0727)	(0.0781)
Lives in Metropolitan Area	0.0273	0.0193	−0.00976	−0.0144	−0.00910	Lives in Metropolitan Area	0.00933	0.0207	0.0821 *	0.0782 *	−0.0104
	(0.0226)	(0.0244)	(0.0218)	(0.0206)	(0.0235)		(0.0502)	(0.0518)	(0.0472)	(0.0400)	(0.0508)
Lives in Retirement Community	0.0881 ***	−0.00235	0.0578 **	0.0872 ***	−0.0817 ***	Lives in Retirement Community	−0.0914 *	−0.136 ***	−0.0635	−0.0507	−0.0377
	(0.0288)	(0.0293)	(0.0272)	(0.0265)	(0.0291)		(0.0515)	(0.0524)	(0.0529)	(0.0449)	(0.0571)
Age (70–79 years)	0.0157	−0.00404	0.0235	0.00263	0.0614 ***	Age (70–79 years)	−0.0298	−0.0226	0.0128	−0.0601 *	0.0327
	(0.0192)	(0.0203)	(0.0180)	(0.0168)	(0.0193)		(0.0366)	(0.0372)	(0.0357)	(0.0322)	(0.0365)
Financial Difficulty	0.193 ***	0.0487	0.150 ***	0.135 ***	−0.0884 **	Financial Difficulty	0.101 *	0.151 ***	0.0618	0.0654	−0.0882 *
	(0.0358)	(0.0357)	(0.0346)	(0.0335)	(0.0353)		(0.0536)	(0.0534)	(0.0518)	(0.0483)	(0.0527)
Poor Health	0.0927 ***	0.179 ***	0.0664 ***	0.0898 ***	−0.0998 ***	Poor Health	0.275 ***	0.208 ***	0.171 ***	0.137 ***	−0.0326
	(0.0234)	(0.0240)	(0.0221)	(0.0213)	(0.0237)		(0.0554)	(0.0548)	(0.0551)	(0.0524)	(0.0541)
COVID-19 Incidence	0.140 ***	0.0500	0.136 ***	0.0910 **	−0.0876 *	COVID-19 Incidence	−0.0658 **	−0.0898	−0.0850 **	−0.0543 **	−0.0097
	(0.0472)	(0.0475)	(0.0459)	(0.0431)	(0.0470)		(0.0666)	(0.0728)	(0.0710)	(0.0651)	(0.0694)
Observations	2509	2509	2509	2509	2509	Observations	748	748	748	748	748
(B)
PANEL A: DID NOT USE TELEHEALTH						PANEL B: USED TELEHEALTH
Explanatory Variables	(1)	(2)	(3)	(4)	(5)	Explanatory Variables	(1)	(2)	(3)	(4)	(5)
	Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future		Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future
Male	−0.153 ***	−0.0708 ***	−0.0851 ***	−0.109 ***	0.0577 ***	Male	−0.141 ***	−0.0797 **	−0.0572 *	−0.0992 ***	0.0492
	(0.0191)	(0.0210)	(0.0183)	(0.0168)	(0.0200)		(0.0301)	(0.0319)	(0.0301)	(0.0267)	(0.0316)
White	0.0747 **	−0.0330	−0.0481	−0.0173	−0.0430	White	−0.0332	0.0214	0.0101	0.0822 *	0.0525
	(0.0374)	(0.0418)	(0.0375)	(0.0345)	(0.0393)		(0.0548)	(0.0568)	(0.0533)	(0.0444)	(0.0571)
Black	−0.0152	−0.0889 **	−0.0197	−0.00780	−0.0820 *	Black	−0.192 ***	0.00989	−0.110 *	−0.0305	0.0993
	(0.0443)	(0.0440)	(0.0388)	(0.0370)	(0.0472)		(0.0492)	(0.0687)	(0.0569)	(0.0580)	(0.0625)
Lives in Metropolitan Area	0.0329	0.0102	0.0132	0.000995	−0.0195	Lives in Metropolitan Area	−0.0129	0.0464	−0.0266	−0.00220	0.0561
	(0.0234)	(0.0253)	(0.0223)	(0.0211)	(0.0240)		(0.0447)	(0.0457)	(0.0446)	(0.0393)	(0.0471)
Lives in Retirement Community	0.0615 **	0.0125	0.0550 *	0.0789 ***	−0.0940 ***	Lives in Retirement Community	0.0108	−0.133 ***	−0.0186	0.0133	−0.0213
	(0.0304)	(0.0313)	(0.0289)	(0.0281)	(0.0309)		(0.0462)	(0.0449)	(0.0449)	(0.0407)	(0.0481)
Age (70–79 years)	0.00256	−0.00522	0.0238	−0.000314	0.0414 **	Age (70–79 years)	0.0139	−0.0221	0.00704	−0.0453	0.108 ***
	(0.0203)	(0.0215)	(0.0191)	(0.0178)	(0.0203)		(0.0314)	(0.0328)	(0.0308)	(0.0280)	(0.0324)
Financial Difficulty	0.218 ***	0.0958 **	0.153 ***	0.125 ***	−0.130 ***	Financial Difficulty	0.0769	0.0532	0.0797 *	0.0976 **	−0.00802
	(0.0372)	(0.0374)	(0.0360)	(0.0346)	(0.0369)		(0.0490)	(0.0492)	(0.0479)	(0.0458)	(0.0485)
Poor Health	0.0899 ***	0.186 ***	0.0903 ***	0.106 ***	−0.0974 ***	Poor Health	0.194 ***	0.189 ***	0.0429	0.0718 *	−0.0568
	(0.0255)	(0.0261)	(0.0243)	(0.0235)	(0.0255)		(0.0413)	(0.0410)	(0.0395)	(0.0368)	(0.0410)
COVID-19 Incidence	0.144 ***	0.101 **	0.106 **	0.0452	−0.00682	COVID-19 Incidence	−0.0420	−0.0132	−0.144 **	−0.136 **	−0.147 **
	(0.0492)	(0.0497)	(0.0470)	(0.0426)	(0.0467)		(0.0606)	(0.0655)	(0.0658)	(0.0636)	(0.0668)
Observations	2280	2280	2280	2280	2280	Observations	977	977	977	977	977
(C)
PANEL A: DID NOT ATTEND RELIGIOUS OR GROUP ACTIVITIES ONLINE						PANEL B: ATTENDED RELIGIOUS OR GROUP ACTIVITIES ONLINE
Explanatory Variables	(1)	(2)	(3)	(4)	(5)	Explanatory Variables	(1)	(2)	(3)	(4)	(5)
	Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future		Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future
Male	−0.143 ***	−0.0709 ***	−0.0802 ***	−0.112 ***	0.0462 **	Male	−0.162 ***	−0.0839 **	−0.0445	−0.0747 **	0.0579 *
	(0.0186)	(0.0201)	(0.0178)	(0.0163)	(0.0194)		(0.0328)	(0.0356)	(0.0332)	(0.0296)	(0.0348)
White	0.0382	0.00356	−0.0256	0.00926	0.0115	White	0.0448	−0.0786	−0.0555	0.0408	−0.0649
	(0.0345)	(0.0374)	(0.0339)	(0.0307)	(0.0364)		(0.0703)	(0.0775)	(0.0714)	(0.0616)	(0.0718)
Black	−0.0705 *	−0.0400	−0.0308	0.00574	−0.0158	Black	−0.0870	−0.119	−0.0886	−0.0409	−0.0366
	(0.0382)	(0.0424)	(0.0366)	(0.0360)	(0.0418)		(0.0736)	(0.0755)	(0.0679)	(0.0672)	(0.0842)
Lives in Metropolitan Area	0.0307	0.0266	0.0185	0.00752	0.000318	Lives in Metropolitan Area	0.00701	0.0119	−0.00978	−0.00895	−0.0521
	(0.0230)	(0.0247)	(0.0220)	(0.0205)	(0.0240)		(0.0461)	(0.0483)	(0.0458)	(0.0414)	(0.0459)
Lives in Retirement Community	0.0633 **	−0.00494	0.0261	0.0639 **	−0.0784 ***	Lives in Retirement Community	0.0116	−0.116 **	0.0618	0.0376	−0.0523
	(0.0290)	(0.0297)	(0.0272)	(0.0263)	(0.0295)		(0.0526)	(0.0508)	(0.0534)	(0.0483)	(0.0544)
Age (70–79 years)	0.0195	−0.00284	0.0253	−0.00918	0.0581 ***	Age (70–79 years)	−0.0223	−0.0251	0.0368	−0.0125	0.0283
	(0.0194)	(0.0204)	(0.0184)	(0.0170)	(0.0195)		(0.0343)	(0.0358)	(0.0331)	(0.0303)	(0.0348)
Financial Difficulty	0.187 ***	0.0305	0.131 ***	0.110 ***	−0.0810 **	Financial Difficulty	0.105 *	0.227 ***	0.121 **	0.131 **	−0.116 **
	(0.0347)	(0.0345)	(0.0334)	(0.0320)	(0.0341)		(0.0571)	(0.0573)	(0.0557)	(0.0539)	(0.0569)
Poor Health	0.110 ***	0.166 ***	0.0772 ***	0.0949 ***	−0.0937 ***	Poor Health	0.176 ***	0.254 ***	0.0900 *	0.104 **	−0.0671
	(0.0239)	(0.0243)	(0.0227)	(0.0217)	(0.0239)		(0.0529)	(0.0518)	(0.0507)	(0.0486)	(0.0516)
COVID-19 Incidence	0.109 **	0.0438	0.125 ***	0.0961 **	0.0489	COVID-19 Incidence	−0.0214	−0.116	−0.119	−0.0172	−0.0955
	(0.0449)	(0.0453)	(0.0440)	(0.0415)	(0.0445)		(0.0753)	(0.0831)	(0.0794)	(0.0695)	(0.0802)
Observations	2469	2469	2469	2469	2469	Observations	788	788	788	788	788

Note: Marginal effects from Probit regressions are reported. Standard errors are given in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1.

A–C show the predicted marginal effects of COVID_19 Incidence (when an individual had a positive COVID-19 test result, diagnosis, or symptoms) on the predicted probability of the five mental health outcomes, controlling for other individual attributes. Similarly,

Table 4. Mental health impact of COVID-19 restrictions (technology users vs. non-users of technology).

(A)
PANEL A: DID NOT LEARN A NEW TECHNOLOGY						PANEL B: LEARNED NEW TECHNOLOGY
Explanatory Variables	(1)	(2)	(3)	(4)	(5)	Explanatory Variables	(1)	(2)	(3)	(4)	(5)
	Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future		Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future
Male	−0.145 ***	−0.0722 ***	−0.0754 ***	−0.113 ***	0.0319 *	Male	−0.166 ***	−0.0828 **	−0.0595 *	−0.0744 **	0.0742 **
	(0.0183)	(0.0200)	(0.0174)	(0.0160)	(0.0194)		(0.0347)	(0.0368)	(0.0354)	(0.0310)	(0.0360)
White	0.0529	−0.0207	−0.0602 *	0.00273	0.0148	White	−0.00622	0.0201	0.0585	0.0393	0.00412
	(0.0341)	(0.0376)	(0.0339)	(0.0308)	(0.0366)		(0.0732)	(0.0753)	(0.0706)	(0.0622)	(0.0739)
Black	−0.0384	−0.0548	−0.0581 *	−0.00855	−0.0437	Black	−0.211 ***	−0.0714	0.00808	−0.0190	0.0825
	(0.0391)	(0.0415)	(0.0335)	(0.0344)	(0.0422)		(0.0622)	(0.0821)	(0.0839)	(0.0720)	(0.0786)
Lives in Metropolitan Area	0.0277	0.0195	−0.00929	−0.0143	−0.0101	Lives in Metropolitan Area	0.00635	0.0231	0.0839 *	0.0787 **	−0.0123
	(0.0226)	(0.0243)	(0.0218)	(0.0206)	(0.0237)		(0.0503)	(0.0517)	(0.0471)	(0.0399)	(0.0508)
Lives in Retirement Community	0.117 ***	0.0181	0.0625 **	0.0728 **	0.0194	Lives in Retirement Community	−0.0684	−0.149 ***	−0.0667	−0.0497	0.00837
	(0.0323)	(0.0323)	(0.0300)	(0.0288)	(0.0306)		(0.0572)	(0.0558)	(0.0566)	(0.0487)	(0.0603)
Age (70–79 years)	0.0105	−0.00831	0.0234	0.00684	0.0377 *	Age (70–79 years)	−0.0368	−0.0175	0.0154	−0.0590 *	0.0225
	(0.0194)	(0.0205)	(0.0182)	(0.0171)	(0.0197)		(0.0369)	(0.0375)	(0.0360)	(0.0325)	(0.0369)
Financial Difficulty	0.194 ***	0.0492	0.151 ***	0.135 ***	−0.0898 **	Financial Difficulty	0.0965 *	0.155 ***	0.0653	0.0681	−0.0915 *
	(0.0358)	(0.0358)	(0.0346)	(0.0335)	(0.0355)		(0.0535)	(0.0532)	(0.0518)	(0.0483)	(0.0528)
Poor Health	0.0996 ***	0.183 ***	0.0718 ***	0.0920 ***	−0.0958 ***	Poor Health	0.280 ***	0.204 ***	0.169 ***	0.136 ***	−0.0276
	(0.0235)	(0.0240)	(0.0221)	(0.0213)	(0.0238)		(0.0555)	(0.0549)	(0.0551)	(0.0524)	(0.0541)
COVID-19 Restrictions	0.0778 **	0.0654	0.0106	0.0442	0.340 ***	COVID-19 Restrictions	−0.101 **	−0.0646	−0.0201	−0.00126	−0.185 *
	(0.0359)	(0.0415)	(0.0369)	(0.0371)	(0.0423)		(0.0802)	(0.0968)	(0.0916)	(0.0796)	(0.0945)
Observations	2509	2509	2509	2509	2509	Observations	748	748	748	748	748
(B)
PANEL A: DID NOT USE TELEHEALTH						PANEL B: USED TELEHEALTH
Explanatory Variables	(1)	(2)	(3)	(4)	(5)	Explanatory Variables	(1)	(2)	(3)	(4)	(5)
	Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future		Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future
Male	−0.154 ***	−0.0721 ***	−0.0848 ***	−0.108 ***	0.0481 **	Male	−0.144 ***	−0.0801 **	−0.0590 **	−0.101 ***	0.0472
	(0.0192)	(0.0211)	(0.0183)	(0.0168)	(0.0202)		(0.0301)	(0.0319)	(0.0300)	(0.0267)	(0.0317)
White	0.0719 *	−0.0332	−0.0518	−0.0219	−0.0267	White	−0.0251	0.0231	0.00214	0.0741 *	0.0672
	(0.0375)	(0.0418)	(0.0376)	(0.0347)	(0.0399)		(0.0545)	(0.0567)	(0.0533)	(0.0448)	(0.0570)
Black	−0.0184	−0.0902 **	−0.0218	−0.00862	−0.0837 *	Black	−0.191 ***	0.0108	−0.118 **	−0.0394	0.109 *
	(0.0441)	(0.0440)	(0.0387)	(0.0369)	(0.0473)		(0.0492)	(0.0687)	(0.0558)	(0.0566)	(0.0617)
Lives in Metropolitan Area	0.0343	0.0113	0.0141	0.00169	−0.0218	Lives in Metropolitan Area	−0.0135	0.0463	−0.0275	−0.00488	0.0569
	(0.0233)	(0.0252)	(0.0222)	(0.0210)	(0.0241)		(0.0448)	(0.0457)	(0.0446)	(0.0394)	(0.0471)
Lives in Retirement Community	0.0791 **	0.0304	0.0576 *	0.0622 **	0.0132	Lives in Retirement Community	0.0574	−0.125 ***	−0.0155	0.00977	0.0324
	(0.0341)	(0.0346)	(0.0320)	(0.0306)	(0.0325)		(0.0511)	(0.0483)	(0.0479)	(0.0433)	(0.0501)
Age (70–79 years)	0.00112	−0.00746	0.0249	0.00390	0.0184	Age (70–79 years)	0.00122	−0.0244	0.0058	−0.0438	0.0936 ***
	(0.0205)	(0.0217)	(0.0193)	(0.0181)	(0.0207)		(0.0318)	(0.0332)	(0.0312)	(0.0284)	(0.0328)
Financial Difficulty	0.216 ***	0.0949 **	0.153 ***	0.125 ***	−0.134 ***	Financial Difficulty	0.0733	0.0523	0.0880 *	0.107 **	−0.0178
	(0.0372)	(0.0374)	(0.0360)	(0.0346)	(0.0371)		(0.0487)	(0.0491)	(0.0479)	(0.0459)	(0.0486)
Poor Health	0.0953 ***	0.190 ***	0.0938 ***	0.106 ***	−0.0890 ***	Poor Health	0.201 ***	0.189 ***	0.0456	0.0731 **	−0.0537
	(0.0256)	(0.0260)	(0.0243)	(0.0235)	(0.0257)		(0.0415)	(0.0410)	(0.0395)	(0.0369)	(0.0411)
COVID-19 Restrictions	0.0424	0.0520	0.00227	0.0501	0.346 ***	COVID-19 Restrictions	−0.170 ***	−0.0337	−0.0165	−0.0126	−0.220 ***
	(0.0406)	(0.0447)	(0.0400)	(0.0402)	(0.0456)		(0.053)	(0.0729)	(0.0683)	(0.0618)	(0.0736)
Observations	2280	2280	2280	2280	2280	Observations	977	977	977	977	977
(C)
PANEL A: DID NOT ATTEND RELIGIOUS OR GROUP ACTIVITIES ONLINE						PANEL B: ATTENDED RELIGIOUS OR GROUP ACTIVITIES ONLINE
Explanatory Variables	(1)	(2)	(3)	(4)	(5)	Explanatory Variables	(1)	(2)	(3)	(4)	(5)
	Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future		Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future
Male	−0.148 ***	−0.0739 ***	−0.0819 ***	−0.111 ***	0.0351 *	Male	−0.164 ***	−0.0834 **	−0.0445	−0.0753 **	0.0580 *
	(0.0186)	(0.0202)	(0.0179)	(0.0163)	(0.0196)		(0.0328)	(0.0356)	(0.0332)	(0.0296)	(0.0348)
White	0.0389	0.00500	−0.0299	0.00271	0.0286	White	0.0394	−0.0850	−0.0655	0.0368	−0.0567
	(0.0344)	(0.0374)	(0.0339)	(0.0309)	(0.0367)		(0.0706)	(0.0775)	(0.0717)	(0.0620)	(0.0723)
Black	−0.0757 **	−0.0421	−0.0364	0.00131	−0.0159	Black	−0.0876	−0.119	−0.0891	−0.0412	−0.0362
	(0.0378)	(0.0424)	(0.0362)	(0.0356)	(0.0419)		(0.0734)	(0.0753)	(0.0676)	(0.0671)	(0.0842)
Lives in Metropolitan Area	0.0314	0.0270	0.0195	0.00822	−0.000641	Lives in Metropolitan Area	0.00921	0.0133	−0.00876	−0.00821	−0.0536
	(0.0230)	(0.0247)	(0.0219)	(0.0205)	(0.0242)		(0.0462)	(0.0483)	(0.0457)	(0.0414)	(0.0458)
Lives in Retirement Community	0.111 ***	0.0198	0.0419	0.0562 *	0.0287	Lives in Retirement Community	−0.0193	−0.136 **	0.0243	0.0176	−0.0167
	(0.0329)	(0.0329)	(0.0304)	(0.0289)	(0.0311)		(0.0548)	(0.0527)	(0.0552)	(0.0501)	(0.0568)
Age (70–79 years)	0.0111	−0.00759	0.0234	−0.00602	0.0344 *	Age (70–79 years)	−0.0145	−0.0187	0.0479	−0.00703	0.0179
	(0.0196)	(0.0206)	(0.0186)	(0.0172)	(0.0199)		(0.0347)	(0.0362)	(0.0335)	(0.0307)	(0.0352)
Financial Difficulty	0.188 ***	0.0309	0.133 ***	0.112 ***	−0.0835 **	Financial Difficulty	0.107 *	0.233 ***	0.129 **	0.134 **	−0.122 **
	(0.0346)	(0.0345)	(0.0334)	(0.0320)	(0.0343)		(0.0570)	(0.0570)	(0.0557)	(0.0539)	(0.0569)
Poor Health	0.118 ***	0.170 ***	0.0828 ***	0.0974 ***	−0.0871 ***	Poor Health	0.176 ***	0.252 ***	0.0881 *	0.103 **	−0.0649
	(0.0240)	(0.0243)	(0.0228)	(0.0217)	(0.0241)		(0.0530)	(0.0518)	(0.0506)	(0.0486)	(0.0516)
COVID-19 Restrictions	0.124 ***	0.0769 *	0.0424	0.0258	0.348 ***	COVID-19 Restrictions	−0.140	−0.0986	−0.166 *	−0.0878	−0.160 *
	(0.0337)	(0.0416)	(0.0363)	(0.0366)	(0.0423)		(0.0929)	(0.0946)	(0.0916)	(0.0841)	(0.0915)
Observations	2469	2469	2469	2469	2469	Observations	788	788	788	788	788

Note: Marginal effects from Probit regressions are reported. Standard errors are given in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1.

A–C show the predicted marginal effects of COVID_19 Restrictions (when an individual experiences quarantine, stay at home, visitation limitations, limitation on group activities or meals) on the predicted probability of the same five mental health outcomes, controlling for the same individual attributes. While

Table 5. Mental health impact of COVID-19’s continued effect on life (technology users vs. non-users of technology).

(A)
PANELA: DID NOT LEARN A NEW TECHNOLOGY						PANELB: LEARNED A NEW TECHNOLOGY
Explanatory Variables	(1)	(2)	(3)	(4)	(5)	Explanatory Variables	(1)	(2)	(3)	(4)	(5)
	Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future		Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future
Male	−0.143 ***	−0.0713 ***	−0.0761 ***	−0.114 ***	0.0417 **	Male	−0.164 ***	−0.0831 **	−0.0593 *	−0.0740 **	0.0756 **
	(0.0183)	(0.0200)	(0.0174)	(0.0160)	(0.0192)		(0.0347)	(0.0367)	(0.0355)	(0.0310)	(0.0360)
White	0.0430	−0.0339	−0.0717 **	0.00697	−0.00278	White	−0.0171	0.0200	0.0471	0.0311	0.0119
	(0.0346)	(0.0380)	(0.0345)	(0.0308)	(0.0364)		(0.0734)	(0.0755)	(0.0713)	(0.0629)	(0.0741)
Black	−0.0401	−0.0575	−0.0605 *	−0.00791	−0.0440	Black	−0.212 ***	−0.0727	0.00491	−0.0226	0.0882
	(0.0391)	(0.0415)	(0.0333)	(0.0346)	(0.0422)		(0.0616)	(0.0820)	(0.0835)	(0.0712)	(0.0779)
Lives in Metropolitan Area	0.0272	0.0180	−0.0104	−0.0144	−0.00950	Lives in Metropolitan Area	0.0120	0.0228	0.0885 *	0.0809 **	−0.0146
	(0.0226)	(0.0244)	(0.0219)	(0.0206)	(0.0235)		(0.0501)	(0.0518)	(0.0468)	(0.0397)	(0.0507)
Lives in Retirement Community	0.0896 ***	−0.00212	0.0596 **	0.0882 ***	−0.0826 ***	Lives in Retirement Community	−0.0914 *	−0.135 **	−0.0605	−0.0482	−0.0395
	(0.0288)	(0.0294)	(0.0272)	(0.0266)	(0.0291)		(0.0516)	(0.0525)	(0.0530)	(0.0451)	(0.0572)
Age (70–79 years)	0.0157	−0.00518	0.0231	0.00365	0.0607 ***	Age (70–79 years)	−0.0310	−0.0212	0.0135	−0.0599 *	0.0338
	(0.0192)	(0.0203)	(0.0180)	(0.0168)	(0.0193)		(0.0366)	(0.0372)	(0.0357)	(0.0322)	(0.0365)
Financial Difficulty	0.194 ***	0.0500	0.152 ***	0.135 ***	−0.0892 **	Financial Difficulty	0.101 *	0.155 ***	0.0672	0.0691	−0.0908 *
	(0.0358)	(0.0358)	(0.0346)	(0.0335)	(0.0353)		(0.0536)	(0.0532)	(0.0520)	(0.0484)	(0.0528)
Poor Health	0.0965 ***	0.180 ***	0.0700 ***	0.0938 ***	−0.103 ***	Poor Health	0.280 ***	0.206 ***	0.176 ***	0.139 ***	−0.0367
	(0.0234)	(0.0240)	(0.0220)	(0.0213)	(0.0236)		(0.0555)	(0.0548)	(0.0553)	(0.0525)	(0.0543)
COVID-19 Continues to Affect Life	0.0450	0.0799 **	0.0783 ***	0.0124	0.00509	COVID-19 Continues to Affect Life	−0.134 *	−0.0206	−0.168 **	−0.0917	−0.150 *
	(0.0324)	(0.0346)	(0.0283)	(0.0302)	(0.0346)		(0.0785)	(0.0914)	(0.0706)	(0.0645)	(0.0779)
Observations	2509	2509	2509	2509	2509	Observations	748	748	748	748	748
(B)
PANEL A: DID NOT USE TELEHEALTH						PANEL B: USED TELEHEALTH
Explanatory Variables	(1)	(2)	(3)	(4)	(5)	Explanatory Variables	(1)	(2)	(3)	(4)	(5)
	Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future		Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future
Male	−0.153 ***	−0.0706 ***	−0.0848 ***	−0.109 ***	0.0577 ***	Male	−0.142 ***	−0.0805 **	−0.0596 **	−0.0997 ***	0.0514
	(0.0191)	(0.0210)	(0.0182)	(0.0168)	(0.0200)		(0.0301)	(0.0319)	(0.0301)	(0.0267)	(0.0316)
White	0.0635 *	−0.0452	−0.0658 *	−0.0205	−0.0422	White	−0.0411	0.0179	−0.00257	0.0795 *	0.0624
	(0.0380)	(0.0422)	(0.0382)	(0.0349)	(0.0395)		(0.0555)	(0.0573)	(0.0539)	(0.0452)	(0.0574)
Black	−0.0206	−0.0929 **	−0.0256	−0.00930	−0.0817 *	Black	−0.195 ***	0.00863	−0.119 **	−0.0359	0.109 *
	(0.0440)	(0.0440)	(0.0385)	(0.0369)	(0.0472)		(0.0489)	(0.0687)	(0.0557)	(0.0574)	(0.0617)
Lives in Metropolitan Area	0.0343	0.0110	0.0138	0.00122	−0.0195	Lives in Metropolitan Area	−0.0134	0.0461	−0.0275	−0.00465	0.0568
	(0.0233)	(0.0252)	(0.0222)	(0.0210)	(0.0240)		(0.0447)	(0.0457)	(0.0446)	(0.0394)	(0.0470)
Lives in Retirement Community	0.0650 **	0.0148	0.0590 **	0.0801 ***	−0.0942 ***	Lives in Retirement Community	0.00872	−0.134 ***	−0.0199	0.0140	−0.0200
	(0.0304)	(0.0313)	(0.0290)	(0.0282)	(0.0309)		(0.0461)	(0.0449)	(0.0447)	(0.0407)	(0.0479)
Age (70–79 years)	0.00380	−0.00465	0.0244	0.000176	0.0413 **	Age (70–79 years)	0.0123	−0.0230	0.00638	−0.0440	0.108 ***
	(0.0203)	(0.0215)	(0.0191)	(0.0178)	(0.0203)		(0.0314)	(0.0329)	(0.0308)	(0.0280)	(0.0323)
Financial Difficulty	0.216 ***	0.0955 **	0.153 ***	0.125 ***	−0.130 ***	Financial Difficulty	0.0771	0.0539	0.0893 *	0.105 **	−0.0171
	(0.0372)	(0.0374)	(0.0360)	(0.0346)	(0.0369)		(0.0490)	(0.0491)	(0.0480)	(0.0459)	(0.0486)
Poor Health	0.0934 ***	0.188 ***	0.0930 ***	0.108 ***	−0.0976 ***	Poor Health	0.193 ***	0.188 ***	0.0448	0.0739 **	−0.0590
	(0.0255)	(0.0260)	(0.0243)	(0.0235)	(0.0255)		(0.0412)	(0.0410)	(0.0395)	(0.0369)	(0.0409)
COVID-19 Continues to Affect Life	0.0459	0.0733 **	0.0945 ***	0.0101	−0.00520	COVID-19 Continues to Affect Life	−0.114	−0.0498 **	−0.0472 **	−0.0479	−0.0213
	(0.0328)	(0.0348)	(0.0277)	(0.0294)	(0.0343)		(0.0760)	(0.0878)	(0.0804)	(0.0815)	(0.0873)
Observations	2280	2280	2280	2280	2280	Observations	977	977	977	977	977
(C)
PANEL A: DID NOT ATTEND RELIGIOUS OR GROUP ACTIVITIES ONLINE						PANEL B: ATTENDED RELIGIOUS OR GROUP ACTIVITIES ONLINE
Explanatory Variables	(1)	(2)	(3)	(4)	(5)	Explanatory Variables	(1)	(2)	(3)	(4)	(5)
	Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future		Feeling Lonely	Have Poor Quality of Sleep	Feeling Worried or Anxious	Feeling Sad or Depressed	Not Hopeful About the Future
Male	−0.145 ***	−0.0726 ***	−0.0824 ***	−0.112 ***	0.0464 **	Male	−0.162 ***	−0.0821 **	−0.0401	−0.0707 **	0.0552
	(0.0185)	(0.0201)	(0.0178)	(0.0163)	(0.0194)		(0.0328)	(0.0356)	(0.0333)	(0.0297)	(0.0348)
White	0.0262	−0.00683	−0.0431	0.00680	0.0131	White	0.0394	−0.0886	−0.0722	0.0248	−0.0466
	(0.0349)	(0.0377)	(0.0345)	(0.0310)	(0.0365)		(0.0713)	(0.0782)	(0.0727)	(0.0637)	(0.0734)
Black	−0.0760 **	−0.0430	−0.0376	0.00198	−0.0140	Black	−0.0904	−0.123	−0.0951	−0.0484	−0.0269
	(0.0378)	(0.0424)	(0.0361)	(0.0357)	(0.0417)		(0.0734)	(0.0753)	(0.0674)	(0.0663)	(0.0840)
Lives in Metropolitan Area	0.0312	0.0262	0.0190	0.00822	−0.000205	Lives in Metropolitan Area	0.00666	0.0113	−0.0106	−0.0102	−0.0508
	(0.0230)	(0.0247)	(0.0220)	(0.0205)	(0.0240)		(0.0462)	(0.0483)	(0.0458)	(0.0415)	(0.0459)
Lives in Retirement Community	0.0657 **	−0.00396	0.0290	0.0651 **	−0.0790 ***	Lives in Retirement Community	0.0110	−0.118 **	0.0604	0.0343	−0.0496
	(0.0290)	(0.0297)	(0.0273)	(0.0263)	(0.0295)		(0.0526)	(0.0508)	(0.0532)	(0.0481)	(0.0543)
Age (70–79 years)	0.0194	−0.00394	0.0252	−0.00753	0.0575 ***	Age (70–79 years)	−0.0231	−0.0248	0.0368	−0.0149	0.0292
	(0.0194)	(0.0205)	(0.0184)	(0.0170)	(0.0195)		(0.0343)	(0.0358)	(0.0331)	(0.0303)	(0.0348)
Financial Difficulty	0.188 ***	0.0311	0.133 ***	0.112 ***	−0.0818 **	Financial Difficulty	0.105 *	0.233 ***	0.130 **	0.138 **	−0.123 **
	(0.0347)	(0.0345)	(0.0334)	(0.0320)	(0.0341)		(0.0571)	(0.0571)	(0.0560)	(0.0543)	(0.0572)
Poor Health	0.112 ***	0.166 ***	0.0798 ***	0.0989 ***	−0.0954 ***	Poor Health	0.177 ***	0.254 ***	0.0918 *	0.105 **	−0.0694
	(0.0239)	(0.0243)	(0.0227)	(0.0217)	(0.0239)		(0.0530)	(0.0518)	(0.0508)	(0.0489)	(0.0518)
COVID-19 Continues to Affect Life	0.0611 *	0.0802 **	0.0899 ***	0.0220 **	0.00592	COVID-19 Continues to Affect Life	−0.0428	−0.0607	−0.118 *	−0.129 ***	−0.138 **
	(0.0329)	(0.0355)	(0.0292)	(0.0315)	(0.0355)		(0.0732)	(0.0767)	(0.0625)	(0.0496)	(0.0678)
Observations	2469	2469	2469	2469	2469	Observations	788	788	788	788	788

Note: Marginal effects from Probit regressions are reported. Standard errors are given in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1.

A–C show the predicted marginal effects of COVID_19 Continues to Affect Life (when an individual reports that COVID-19 continues to affect their daily life) on the predicted probability of the same five mental health outcomes, controlling for the same individual attributes.

All the result tables have two panels—Panel A represents the sub-sample of individuals who “did not use technology” during the pandemic, while Panel B depicts the sub-sample of “technology users.” Table 3A, Table 4A and Table 5A stratify technology users based on whether they “learned a new technology during the pandemic.” Table 3B, Table 4B and Table 5B use an alternative definition of technology use to stratify the sample. In these three tables, technology users are defined by “whether they used telehealth during the pandemic.” Finally, Table 3C, Table 4C and Table 5C determine technology use based on “whether an individual attended religious or group activities online.” Three alternative ways are used to stratify the technology users versus non-users to validate the general robustness of our results.

In Table 3A–C, we find the predicted marginal effects of COVID_19 Incidence are almost always positive in Panel A (technology non-users) and negative in Panel B (technology users). This implies that having an incidence of the disease (COVID-19) raises the predicted probability of an adverse mental health outcome among individuals who did not use technology during the pandemic and lowers the predicted probability of adverse mental health outcomes among the technology users. For example, in Table 3A, we see that COVID_19 Incidence among technology non-users led to a statistically significant greater predicted probability of feeling lonely, worried or anxious, and sad or depressed by 0.140, 0.136, and 0.091 percentage points, respectively. On the other hand, in the sub-sample of technology users, COVID_19 Incidence led to a statistically significant lower predicted probability of feeling lonely, worried or anxious, and sad or depressed by 0.065, 0.085, and 0.054 percentage points, respectively. The results remain qualitatively similar (with a difference in only statistical significance) in Table 3B,C (based on alternative definitions of technology use).

In Table 4A–C, we find the predicted marginal effects of COVID_19 Restrictions are all positive in Panel A and all negative in Panel B. This leads to a similar interpretation as in the previous case. We infer that the COVID_19 Restrictions generally led to a higher predicted likelihood of an adverse mental health outcome among individuals who did not learn a new technology, did not use telehealth, or did not attend religious/group activities during the pandemic. Meanwhile, the same variable, Restrictions, led to a lower predicted likelihood of experiencing adverse mental health effect during the pandemic among individuals who learned a new technology, used telehealth, or attended religious/group activities online during the same time. For example, in Table 4A, we see that COVID_19 Restrictions among individuals who did not learn a new technology (in panel A) led to a statistically significant greater predicted probability of feeling lonely and not feeling hopeful about the future by 0.077 and 0.340 percentage points, respectively. On the other hand, in the sub-sample of individuals who learned a new technology, COVID_19 Restrictions led to a statistically significant lower predicted probability of feeling lonely and not feeling hopeful about the future by 0.101 and 0.185 percentage points, respectively. Similarly, in Table 4B, among individuals who did not use telehealth during the pandemic, COVID_19 Restrictions led to a higher predicted probability of not feeling hopeful about the future by 0.346 percentage points. While among the telehealth users, the same Restrictions led to a lower predicted probability of not feeling hopeful about the future by 0.220 percentage points. Both predicted marginal effects are statistically significant at a 1% level. The results remain qualitatively similar (with a difference in only statistical significance levels) in Table 4C (where technology use is defined by participation in online religious/group activities).

Lastly, in Table 5A–C, we present the results for the explanatory variable COVID_19 Continues to Affect Life on the five adverse mental health outcomes. Our results, in general, remain similar to the previous cases. For example, in Table 5B, the Continued Effect of COVID_19 on an individual’s life leads to a statistically significant higher predicted probability of poor sleep quality and feeling worried or anxious by 0.073 and 0.0945 percentage points, respectively, in the sub-sample of non-users of technology (panel A). In comparison, it led to a statistically significant lower predicted probability of poor sleep quality and feeling worried or anxious by 0.049 and 0.047 percentage points, respectively, in the sub-sample of technology users (panel B). In Table 5C, the Continued Effect of COVID_19 on an individual’s life led to a higher predicted probability of feeling sad or depressed by 0.022 percentage points among individuals who did not attend religious or group activities online during the pandemic. In comparison, for the sub-sample of technology users who attended religious or group activities online during the pandemic, the Continued Effect of COVID_19 had a lower predicted likelihood of feeling sad or depressed by 0.129 percentage points.

As for some of the other non-COVID-19-related explanatory variables, we found that Males (in both sub-samples) had better mental resilience during the pandemic. Similarly, in general, Black People had a lower predicted probability of adverse mental health outcomes across both groups (technology users and non-users). We also found that Financial Difficulty and Poor Physical Health were associated with a higher predicted probability of unfavorable mental health outcomes irrespective of using technology or not. At the same time, technology users living in retirement communities were better off in mental health status than their counterparts who did not use technology.

The qualitative consistency of the marginal effects for the three COVID-19 specifications (Incidence, Restrictions, Continues to Affect life) across the multiple ways of stratifying the sample into technology users vs. non-users validates the robustness of the results. These results underline the role of technology use in mitigating the mental health impact of COVID-19 among the elderly. Therefore, we infer that elderly individuals who used technology during the pandemic experienced a lower probability of an adverse mental health outcome due to COVID-19. There could be several channels through which technology use dampened the adverse mental health outcomes during the pandemic (for example, improved cognitive functioning, better connectivity with family/friends, less delayed medical visits, or general awareness). But our study cannot explore these channels of causation due to the unavailability of relevant information in the survey data.

4. Conclusions

The findings of our paper underline the role of technology during the COVID-19 pandemic, particularly for older people. However, we caution the readers that the results may not be generalized across different age groups. Using a nationally representative random sample of older individuals, we find that the elderly individuals who used technology during the pandemic had a lower probability of adverse mental health outcomes than the non-users of technology. This corroborates existing evidence that suggests technology interventions yield positive health effects (both physical and mental) for older people. Older people use technology to lessen social isolation and disconnectedness during the pandemic, which lowers the adverse mental health effects imposed by the contagion control strategies during the pandemic. The results also point towards the promising potential of technology use in the overall improvement of long-term mental health issues of the elderly. An underdeveloped aspect of mental health delivery is the consideration of lifestyle interventions like physical exercise, a healthy diet, etc. [25]. The use of technology may significantly help with lifestyle interventions that positively affect the mental health status of the older population over time. However, ensuring older individuals effectively use technology in their best interests is a multi-pronged approach that requires targeted policies and well-thought-out strategic initiatives. Some policy recommendations include designing digital literacy programs for older adults, implementing a community-engaged learning approach in which younger individuals (students) engage with underserved older adults in digital literacy training to reduce the digital gap [26]. Such intergenerational learning opportunities through reverse mentoring (from young to old) promote social engagement and improve overall digital literacy. For example, platforms like Eldera that connect older individuals with younger adults in 8–12-week immersive reverse-mentoring programs significantly enhance their digital literacy [27]. The digital divide can be further reduced by incorporating digital inclusion initiatives into federal programs like AmeriCorps, as well as independent initiatives targeted at older adults in the Departments of Labor and Education. Adequate policy response is also needed to ensure access to affordable devices and better internet connectivity for older Americans. Designing devices with simple, intuitive interfaces, available at affordable list prices or subsidized by the government, along with better broadband connectivity, would be an important determinant of how seamlessly the older population can embrace digitization. Additionally, literacy and affordability programs that offer sustained technical support to help older adults navigate technology barriers are crucial. For example, services like Senior Planet from the American Association of Retired Persons (AARP), which offer free and comprehensive support to older adults, either in person or remotely, must be made broadly available to benefit a larger segment of the older population. Last but not least, policy initiatives must address the inequitable socio-economic factors that make technology adoption difficult and slow for older individuals from a disadvantaged status. Policy needs to address the income and education gaps to ensure equitable access to technology and technology-enabled services [28].

The results from the paper underline the need for including technology use an important explanatory variable in future research for older adults. The pattern and extent of technology use will significantly affect the access to information and overall management of life by older people. Hence, ignoring technology use will lead to the risk of losing important information about differences in the relative well-being of different groups of older Americans. Additionally, it is important to highlight that older adults are a heterogenous group who will differ by their birth cohorts (older olds are not similar to the younger olds). For example, individuals raised with personal computers, internet, smartphones or social media will enter old age with different levels of technology use preparation, digital behavior and expectations. Thus, it is important to capture this cohort-sensitive, evolving dimension of technology use in future research.

Unfortunately, the COVID-19 crisis is far from over. With the emergence of new variants, it continues to be a global challenge. Also, the COVID-19 outbreak may only mark the beginning of an era of pandemics that we must survive over time. Hence, it is crucial to strengthen the resources and infrastructure to safeguard the most vulnerable population groups, like the elderly. Investing in making technology-based health interventions accessible and usable by older people should be a healthcare policy priority in aging societies like America.