Social Isolation in America? A 20-Year Snapshot

Social isolation is a concept that has received growing attention in recent decades (Fischer 2011; Murthy 2023; Parigi and Henson 2014). Yet sociologists rarely agree on the degree to which Americans are socially isolated from one another (Fischer 2009; McPherson, Smith-Lovin, and Brashears 2006). Indeed, social isolation, which broadly refers to “the objective lack of social contact with others” (Umberson and Donnelly 2023:2), is a surprisingly difficult concept to measure given its relatively clear definition.

In this data visualization, I plot two parallel measures of social isolation using 20 years of data from the American Time Use Survey (ATUS), a nationally representative repository of 24-hour time diaries. The ATUS has gathered data on approximately 10,000 noninstitutionalized Americans every year since 2003. Each ATUS respondent provides a detailed account of their previous day, including all social interactions in which they engaged over the past 24 hours. My first measure of social isolation includes respondents who did not engage in a single face-to-face interaction during their diary day (i.e., physical isolation). My second measure of social isolation (i.e., total isolation) includes face-to-face interactions as well as telecommunication interactions (e.g., phone or video calls, text messages). ATUS respondents who report zero face-to-face interactions and zero telecommunication interactions during their diary days are coded as being totally isolated.

Figure 1 shows the percentage of respondents who were socially isolated from the inaugural ATUS year (2003) to the most recently completed ATUS year (2022). All values are nationally representative estimations that were derived using the ATUS-supplied survey weights. The mean value of physical isolation was 7.26% (95 percent confidence interval = 7.13%–7.38%) across the 20-year period. In other words, approximately 1 in every 14 ATUS respondents did not engage in a single face-to-face interaction during their diary days from 2003 to 2022. This value drops to 6.04% (95 percent confidence interval = 5.93%–6.16%) for total isolation. The difference between these two values refers to the percentage of respondents who did not engage in any face-to-face interactions but contacted someone via telecommunication. The main purpose of plotting the two types of social isolation is to show broader trends in isolation over time rather than highlight differences in physical and total isolation.

OPEN IN VIEWER

Figure 1.

Percentage of American Time Use Survey (ATUS) respondents who were socially isolated within a 24-hour period from 2003 to 2022.

As seen in Figure 1, there was sudden spike in isolation during the coronavirus pandemic (2020–2022). Although an increase in physical isolation (i.e., blue circles) is expected during these years, the simultaneous increase in total isolation (i.e., red squares) suggests that Americans were not necessarily compensating their absence of face-to-face interactions with telecommunication. It is further worth noting that social isolation was not rising in the decades prior to the pandemic, with a minor exception during the Great Recession (2008–2009).

The two measures shown in Figure 1 present a longitudinal view of social isolation in America, an issue that has been hotly debated in recent years. Although far from comprehensive, using data from daily time diaries as an indicator of isolation allows a direct assessment of the objective lack of contact that Americans face in everyday life. This micro perspective of daily life offers a nuanced view that broader assessments of social integration (e.g., social participation, social network modules) may fail to detect. Indeed, the focus on social interactions speaks to issues such as the exchange of support, resources, and feelings of belongingness. Collectively, the 20 years of ATUS diaries show that aside from the pandemic, there is minimal evidence that Americans are becoming increasingly socially isolated in recent years.