Social Networking Service Usage Behaviors and Loneliness Among Adults in Seoul, Korea: A Population-based Cross-sectional Study

Abstract

Objectives

Although social network services (SNS) are prevalent globally, the relationship between SNS use and loneliness remains complex, particularly regarding qualitative behavioral patterns. This study examined associations between specific SNS behaviors and loneliness among Seoul adults and explored differences by age and digital literacy.

Methods

Data from the 2024 Seoul Survey were analyzed (n=34 142; mean age, 52.3 years). Loneliness was assessed with 2 items measuring isolation in family and non-family relationships. Five SNS behavioral domains were examined: active use, information-seeking use, communication priority, social comparison, and exclusionary feeling. Digital literacy was dichotomized using an 8-item scale, with scores ≥6 considered high. Multivariable logistic regression estimated adjusted associations.

Results

Overall SNS use was associated with lower odds of loneliness than non-use (adjusted odds ratio [aOR], 0.90; 95% confidence interval [CI], 0.83 to 0.98). However, higher odds were observed for specific SNS behaviors compared with non-use: active use (aOR, 1.24), communication priority (aOR, 1.26), social comparison (aOR, 1.50), and exclusionary feeling (aOR, 1.66); information-seeking use was not significant (aOR, 1.11). A dose–response pattern peaked at 3 problematic behaviors (aOR, 3.21; 95% CI, 2.82 to 3.65). Associations between SNS behaviors and loneliness were stronger among adults aged ≥60 years and participants with high digital literacy.

Conclusions

Although overall SNS use was associated with modestly lower odds of loneliness, qualitative SNS engagement—particularly problematic behaviors—was strongly associated with higher odds. Older adults and participants with high digital literacy appeared particularly vulnerable when performing problematic SNS behaviors, underscoring the need for behavior-focused, population-tailored education and interventions.

GRAPHICAL ABSTRACT

INTRODUCTION

Loneliness has emerged as a major global public health concern, affecting 1 in 6 people worldwide [1]. It is prevalent across age groups and is strongly associated with serious health outcomes, including suicidal ideation, depression, anxiety [2], cardiovascular diseases [3], and increased mortality [4]. Loneliness is generally defined as a subjective, unpleasant feeling that arises from a perceived discrepancy between an individual’s actual and desired social relationships and often involves perceived deficiencies in social connection [5].

In contemporary society, social network services (SNS) have become integral to daily life and have undergone rapid growth and widespread adoption. Korea has exceptionally high adoption rates; 89.3% of the population used SNS in 2021, substantially exceeding global averages [6].

Despite their widespread adoption, SNS use has been associated with negative mental health outcomes, including increased depressive symptoms, anxiety, and loneliness [7]. However, the relationship between SNS use and loneliness is complex, with mixed or contradictory findings across studies and populations [8,9]. Some research suggests that active SNS engagement can foster social support and reduce loneliness by strengthening social connections [10], whereas other work suggests that the benefits of social media for mitigating loneliness may be limited because online engagement can displace offline interactions [11]. Digital literacy has also been identified as important for leveraging SNS to alleviate loneliness [12,13].

These inconsistent findings indicate a need for research that focuses on qualitative patterns of SNS engagement rather than simple use frequency. This study aimed to identify specific SNS behaviors associated with loneliness using a large, population-based sample of Seoul residents. It also explored whether these associations varied by digital literacy and age.

METHODS

Study Population

This study analyzed data from the 2024 Seoul Survey, collected by the Seoul Metropolitan Government and the Seoul Institute. The survey used a citywide representative sampling design across all 25 districts of Seoul [14]. Participants were selected as follows. Of 36 280 participants, 1040 were excluded because they gave inconsistent SNS behavior responses, answering “not using SNS” to one item but not selecting “not using SNS” for another item. Non-Korean nationals, including marriage immigrants and naturalized citizens, were then excluded (n=30), as were individuals younger than 20 years (n=1068). The final analytical sample included 34 142 Korean adults: 25 134 SNS users and 9008 SNS non-users.

Outcome Measurements

Loneliness was measured using responses to the following 2 items [15]: “I sometimes feel lonely in my relationship with my family” and “I sometimes feel lonely in my relationship with people other than my family.” This brief measurement approach is supported by established research showing that short loneliness scales can maintain adequate psychometric properties in large-scale surveys [16]. Participants responded using a 5-point Likert scale: 1=strongly disagree, 2=disagree, 3=neither agree nor disagree, 4=agree, and 5=strongly agree. Respondents who answered “agree” or “strongly agree” were categorized as “yes,” whereas those who answered “strongly disagree,” “disagree,” or “neither agree nor disagree” were categorized as “no.” For the primary analysis, overall loneliness was dichotomized as “yes” if participants were categorized as “yes” for either family or non-family relationships and as “no” if they reported no loneliness in both domains. This binary coding was consistent with previous studies [15,17] and allowed logistic regression to be used with a binary loneliness outcome, facilitating clear interpretation of the odds ratios.

Exposure Measurements

SNS behavioral patterns were assessed with 5 items measured on 5-point Likert scales: 1=strongly disagree, 2=disagree, 3=neither agree nor disagree, 4=agree, 5=strongly agree, and 9=not using SNS. The 5 behavioral domains were as follows: (1) active use, assessed with “I tend to share my daily life on SNS”; (2) information-seeking use, assessed with “I consider information obtained through SNS to be trustworthy”; (3) communication priority, assessed with “I consider communication via SNS more important than offline interaction”; (4) social comparison, assessed with “Viewing other people’s posts on SNS makes me feel that they are happier than I am”; and (5) exclusionary feeling, assessed with “I often feel socially excluded when viewing other people’s posts on SNS.” This approach is grounded in established practices in social media research that recognize the multidimensional nature of SNS use and the importance of measuring both positive and negative aspects of social media engagement [18].

Participants who responded “not using SNS” (code 9) to all 5 items were classified as SNS non-users, whereas those who provided responses on the 1–5 scale for all items were classified as SNS users. Among SNS users, individual item scores were dichotomized: “agree” or “strongly agree” was coded as “yes,” and “strongly disagree,” “disagree,” or “neither agree nor disagree” was coded as “no.” For each behavioral domain, 3 mutually exclusive categories were defined: SNS non-users, SNS users without the behavior (“no”), and SNS users with the behavior (“yes”). The SNS behavior score (SBS) was calculated only among SNS users as the sum of “yes” classifications across the 5 behavioral domains, with possible scores ranging from 0 to 5.

Covariates

Potential confounders included sex, age, education level, monthly income, marital status, occupation, social support, sleep duration, physical activity, and digital literacy.

Demographic factors included sex (male or female), age (20–29, 30–39, 40–49, 50–59, or ≥60 years), education level (middle school or below, high school, or college or above), and monthly income in units of 10 000 Korean won (<200, 200–399, 400–599, 600–799, or ≥800).

Social factors included marital status (married, never married, or other marital status, including divorced, separated, or bereaved), occupation (managerial/professional, white-collar, blue-collar, or other), and social support, defined as a yes/no response to the question “Do you have someone to help you when you are in difficulty?”

Health-related factors included sleep duration (<6, 6–9, or >9 hours) and physical activity frequency (none, once per week or less, or more than once per week).

Digital literacy was assessed using 8 items measuring functional digital competencies: the ability to (1) make phone calls and send text messages, (2) use messenger applications (e.g., KakaoTalk), (3) search for news and information (e.g., YouTube), (4) access transportation information and maps, (5) use reservation and booking services, (6) use payment methods (e.g., online shopping, delivery orders), (7) order food, and (8) issue official documents. Each item was rated on a 4-point Likert scale: 1=unable, 2=poor, 3=fair, and 4=good. Individual item scores were dichotomized as 0 for scores of 1–2 and 1 for scores of 3–4, yielding a digital literacy score ranging from 0 to 8. Participants with scores ≥6 were classified as having high digital literacy, whereas those with scores ≤5 were classified as having low digital literacy.

Statistical Analysis

Descriptive statistics were presented as frequencies and percentages for categorical variables, and chi-square tests were used to characterize participants according to loneliness status. Baseline characteristics were summarized for the full sample, and SNS behaviors were summarized among SNS users. To examine associations between SNS use or behaviors and loneliness, multivariable logistic regression models were used to estimate crude odds ratios and adjusted odds ratios (aORs) with 95% confidence intervals (CIs). Analyses comparing overall SNS use or SNS behaviors with non-use were conducted in the full sample, whereas SBS analyses were restricted to SNS users. For each behavioral domain, SNS non-users served as the reference group, and separate aORs were estimated for SNS users with and without the corresponding behavior. Dose–response patterns were evaluated using the SBS, analyzed as both a continuous and a categorical variable. Stratified analyses were conducted by age and digital literacy to explore potential effect modification. All analyses incorporated sampling weights representative of Seoul residents to support population-level generalizability and were performed using R version 4.5.1 (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was set at p-value <0.05.

Ethics Statement

The Institutional Review Board of Seoul National University Hospital determined that this study was exempt from review and approved a waiver of written informed consent because anonymous secondary data were used (IRB No. E-2410-105-1577).

RESULTS

Study Population

The final analysis included 34 142 participants, of whom 28.2% reported loneliness (Table 1). Loneliness did not differ significantly by sex (28.0% in male vs. 28.4% in female; p=0.369) but was significantly associated with all other characteristics (all p<0.001). The prevalence of loneliness increased with age, from 22.8% among participants aged 20–29 years to 33.5% among those aged ≥60 years, and showed inverse gradients with education and income: 39.5% among participants with middle-school-or-below education versus 21.8% among those with a college degree or higher, and 39.8% in the lowest income group versus 24.0% in the highest income group. Participants with a marital status of “other,” including divorced, separated, or bereaved, reported loneliness more frequently than married participants (42.2 vs. 25.6%). Social support (27.6 vs. 31.3%), longer sleep duration (24.4 for >9 hours vs. 32.6% for <6 hours), and frequent physical activity (26.6% for more than once per week vs. 29.9% for none) were associated with lower loneliness prevalence. Higher digital literacy (25.2 vs. 34.7%) and SNS use (26.3 vs. 33.5%) were also associated with lower loneliness prevalence.

Table 1.

Characteristics of participants according to loneliness

Characteristics	Loneliness		p-value1
	No	Yes
	Total (n=34 142)	24 520 (71.8)		9622 (28.2)
Sex			0.369
Male	11 729 (72.0)	4550 (28.0)
Female	12 791 (71.6)	5072 (28.4)
Age (y)			<0.001
20–29	1976 (77.2)	584 (22.8)
30–39	5227 (75.8)	1670 (24.2)
40–49	4738 (75.1)	1572 (24.9)
50–59	4686 (71.9)	1828 (28.1)
≥60	7893 (66.5)	3968 (33.5)
Education			<0.001
Middle school or below	1530 (60.5)	999 (39.5)
High school	7403 (68.2)	3444 (31.8)
College or above	15 587 (75.1)	5179 (24.9)
Monthly income2			<0.001
<200	1106 (60.2)	730 (39.8)
200–399	5776 (67.7)	2757 (32.3)
400–599	7207 (73.4)	2617 (26.6)
600–799	6824 (74.1)	2380 (25.9)
≥800	3607 (76.0)	1138 (24.0)
Marital status			<0.001
Married	16 886 (74.4)	5809 (25.6)
Never married	4895 (73.0)	1813 (27.0)
Other marital status3	2739 (57.8)	2000 (42.2)
Occupation			<0.001
Managerial/professional	2016 (77.2)	596 (22.8)
White-collar	9618 (73.4)	3484 (26.6)
Blue-collar	6228 (70.4)	2614 (29.6)
Others	6658 (69.5)	2928 (30.5)
Social support			<0.001
No	3466 (68.7)	1582 (31.3)
Yes	21 054 (72.4)	8040 (27.6)
Sleep duration (hr)			<0.001
<6	875 (67.4)	424 (32.6)
6–9	22 965 (71.9)	8979 (28.1)
>9	680 (75.6)	219 (24.4)
Physical activity			<0.001
None	8033 (70.1)	3424 (29.9)
Once per week or less	4468 (70.8)	1841 (29.2)
More than once per week	12 019 (73.4)	4357 (26.6)
Digital literacy			<0.001
Low	7059 (65.3)	3747 (34.7)
High	17 461 (74.8)	5875 (25.2)
SNS usage			<0.001
SNS non-users	5987 (66.5)	3021 (33.5)
SNS users	18 533 (73.7)	6601 (26.3)

Values are presented as number (%).

SNS, social network service.

¹Determined using the chi-square test.

²Measured in 10 000 Korean won/mo.

³Divorced, separated, or bereaved.

Table 2 presents the prevalence of loneliness according to SNS behavior among SNS users (n=25 134). In contrast to overall SNS use, each SNS behavior was associated with higher loneliness prevalence (all p<0.001). The largest differences were observed for social comparison (34.4 vs. 22.9%) and exclusionary feeling (37.5 vs. 22.4%). A cumulative pattern was evident: loneliness prevalence increased from an SBS of 0 (18.9%) to an SBS of 3 (38.4%) and then did not increase further at higher cumulative levels (SBS=4: 34.2%; SBS=5: 26.7%).

Table 2.

Prevalence of loneliness according to SNS behavior among SNS users

SNS behaviors	Loneliness		p-value1
	No	Yes
	SNS users (n=25 134)	18 533 (73.7)		6601 (26.3)
Active use			<0.001
No	10 864 (76.7)	3304 (23.3)
Yes	7669 (69.9)	3297 (30.1)
Information-seeking use			<0.001
No	12 103 (74.6)	4113 (25.4)
Yes	6430 (72.1)	2488 (27.9)
Communication priority			<0.001
No	13 856 (75.8)	4432 (24.2)
Yes	4677 (68.3)	2169 (31.7)
Social comparison			<0.001
No	13 766 (77.1)	4100 (22.9)
Yes	4767 (65.6)	2501 (34.4)
Exclusionary feeling			<0.001
No	14 499 (77.6)	4185 (22.4)
Yes	4034 (62.5)	2416 (37.5)
SBS
Continuous analysis	1.49±1.63	1.95±1.60	<0.001
Categorical analysis			<0.001
0	7533 (81.1)	1751 (18.9)
1	3567 (76.1)	1119 (23.9)
2	2570 (67.8)	1220 (32.2)
3	2014 (61.6)	1253 (38.4)
4	1417 (65.8)	737 (34.2)
5	1432 (73.3)	521 (26.7)

Values are presented as number (%) for categorical variables and mean±standard deviation for continuous variables.

SNS, social network service; SBS, SNS behavior score.

¹Determined using the chi-square test for categorical variables and the t-test for continuous variables.

Social Network Service Behavior and Loneliness

As shown in Table 3, multivariable logistic regression revealed that overall SNS use was associated with lower odds of loneliness than non-use (aOR, 0.90; 95% CI, 0.83 to 0.98) after adjustment for sex, age, education, income, marital status, occupation, social support, sleep duration, physical activity, and digital literacy. However, specific SNS behaviors showed significant positive associations with loneliness. Compared with SNS non-users, SNS users who did not endorse each behavior had lower odds of loneliness, whereas users who endorsed the behavior generally had higher odds. For active use, the aOR was 0.78 (95% CI, 0.71 to 0.85) for “no” and 1.24 (1.12 to 1.37) for “yes”. Corresponding aORs were 0.84 (95% CI, 0.77 to 0.92) and 1.11 (95% CI, 0.99 to 1.23) for information-seeking use, 0.82 (95% CI, 0.75 to 0.89) and 1.26 (95% CI, 1.13 to 1.41) for communication priority, 0.76 (95% CI, 0.69 to 0.83) and 1.50 (95% CI, 1.35 to 1.68) for social comparison, and 0.75 (95% CI, 0.69 to 0.82) and 1.66 (95% CI, 1.49 to 1.86) for exclusionary feeling. All associations were significant except the association for information-seeking use, which was marginal (p=0.066).

Table 3.

Odds ratios for loneliness by SNS behavior and SBS

Variables	Crude	Model 11
SNS behavior2
SNS usage
SNS non-users	1.00 (reference)	1.00 (reference)
SNS users	0.66 (0.61, 0.72)***	0.90 (0.83, 0.98)*
Active use
SNS non-users	1.00 (reference)	1.00 (reference)
No	0.59 (0.54, 0.64)***	0.78 (0.71, 0.85)***
Yes	0.75 (0.69, 0.82)***	1.24 (1.12, 1.37)***
Information-seeking use
SNS non-users	1.00 (reference)	1.00 (reference)
No	0.64 (0.59, 0.69)***	0.84 (0.77, 0.92)***
Yes	0.71 (0.64, 0.78)***	1.11 (0.99, 1.23)
Communication priority
SNS non-users	1.00 (reference)	1.00 (reference)
No	0.60 (0.56, 0.66)***	0.82 (0.75, 0.89)***
Yes	0.82 (0.74, 0.91)***	1.26 (1.13, 1.41)***
Social comparison
SNS non-users	1.00 (reference)	1.00 (reference)
No	0.55 (0.51, 0.60)***	0.76 (0.69, 0.83)***
Yes	0.96 (0.87, 1.06)	1.50 (1.35, 1.68)***
Exclusionary feeling
SNS non-users	1.00 (reference)	1.00 (reference)
No	0.54 (0.50, 0.59)***	0.75 (0.69, 0.82)***
Yes	1.07 (0.97, 1.18)	1.66 (1.49, 1.86)***
SBS3
Continuous analysis	1.16 (1.13, 1.19)***	1.25 (1.22, 1.28)***
Categorical analysis
0	1.00 (reference)	1.00 (reference)
1	1.31 (1.16, 1.48)***	1.82 (1.60, 2.06)***
2	1.77 (1.57, 1.99)***	2.60 (2.29, 2.96)***
3	2.19 (1.93, 2.47)***	3.21 (2.82, 3.65)***
4	2.06 (1.79, 2.38)***	3.08 (2.65, 3.57)***
5	1.60 (1.34, 1.91)***	2.52 (2.10, 3.03)***

Values are presented as odds ratio (95% confidence interval).

SNS, social network service; SBS, SNS behavior score.

¹Model 1: adjusted for sex, age, education, income, marital status, occupation, social support, sleep duration, physical activity, and digital literacy.

²Full sample (n=34 142).

³SNS users (n=25 134).

^*p<0.05,

^***p<0.001.

Analysis of SBS revealed a non-linear dose–response pattern among SNS users. When SBS was analyzed continuously, each additional SNS behavior was associated with 25% higher odds of loneliness (aOR, 1.25; 95% CI, 1.22 to 1.28). In categorical analysis, the odds of loneliness increased progressively through SBS=3: SBS=1 (aOR, 1.82), SBS=2 (aOR, 2.60), and SBS=3 (aOR, 3.21). The associations attenuated at higher cumulative levels, with aORs of 3.08 for SBS=4 and 2.52 for SBS=5, suggesting a non-linear relationship.

Stratification Analysis

Table 4 presents associations between SNS use or behaviors and loneliness stratified by age and digital literacy, after adjustment for all covariates except the stratification variable.

Table 4.

Adjusted odds ratios for loneliness by SNS behavior stratified by age and digital literacy

Variables	Age (y)		Digital literacy
	20–59	≥60	Low	High
	SNS behavior1
SNS usage
SNS non-users	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
SNS users	0.84 (0.75, 0.95)**	0.95 (0.85, 1.07)	0.88 (0.78, 1.00)*	0.92 (0.82, 1.04)
Active use
SNS non-users	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
No	0.71 (0.62, 0.81)***	0.79 (0.69, 0.90)***	0.80 (0.70, 0.91)***	0.75 (0.66, 0.85)***
Yes	1.04 (0.91, 1.18)	1.83 (1.57, 2.15)***	1.27 (1.08, 1.50)**	1.23 (1.07, 1.40)**
Information-seeking use
SNS non-users	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
No	0.78 (0.69, 0.89)***	0.84 (0.74, 0.95)**	0.86 (0.76, 0.98)*	0.82 (0.72, 0.92)**
Yes	0.97 (0.84, 1.11)	1.58 (1.32, 1.88)***	1.01 (0.82, 1.25)	1.15 (1.00, 1.32)*
Communication priority
SNS non-users	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
No	0.76 (0.67, 0.86)***	0.82 (0.72, 0.93)**	0.83 (0.73, 0.94)**	0.81 (0.72, 0.92)***
Yes	1.08 (0.94, 1.25)	1.89 (1.60, 2.22)***	1.24 (1.04, 1.49)*	1.29 (1.11, 1.48)***
Social comparison
SNS non-users	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
No	0.68 (0.60, 0.77)***	0.81 (0.72, 0.93)**	0.81 (0.71, 0.93)**	0.73 (0.64, 0.83)***
Yes	1.32 (1.15, 1.52)***	1.81 (1.53, 2.14)***	1.24 (1.04, 1.48)*	1.57 (1.37, 1.81)***
Exclusionary feeling
SNS non-users	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
No	0.67 (0.59, 0.76)***	0.81 (0.71, 0.92)**	0.80 (0.70, 0.91)***	0.73 (0.64, 0.83)***
Yes	1.44 (1.25, 1.66)***	2.08 (1.75, 2.48)***	1.40 (1.17, 1.68)***	1.71 (1.48, 1.97)***
SBS2
Continuous analysis	1.23 (1.19, 1.26)***	1.37 (1.31, 1.43)***	1.19 (1.13, 1.24)***	1.28 (1.24, 1.32)***
Categorical analysis
0	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
1	1.67 (1.44, 1.93)***	1.89 (1.46, 2.44)***	2.33 (1.84, 2.94)***	1.62 (1.39, 1.89)***
2	2.39 (2.06, 2.77)***	2.85 (2.25, 3.61)***	2.35 (1.86, 2.98)***	2.63 (2.26, 3.06)***
3	2.81 (2.41, 3.28)***	4.62 (3.66, 5.85)***	2.65 (2.05, 3.42)***	3.36 (2.89, 3.91)***
4	2.84 (2.38, 3.37)***	3.52 (2.66, 4.65)***	2.38 (1.77, 3.19)***	3.30 (2.78, 3.91)***
5	2.49 (2.02, 3.06)***	2.27 (1.64, 3.14)***	1.05 (0.70, 1.58)	2.96 (2.42, 3.62)***

Values are presented as adjusted odds ratio (95% confidence interval), adjusted for sex, age, education, income, marital status, occupation, social support, sleep duration, physical activity, and digital literacy (excluding the stratification variable).

SNS, social network service; SBS, SNS behavior score.

¹Full sample. Stratum number: age 20–59 years, 22 281; ≥60 years, 11 861; low digital literacy, 10 806; high digital literacy, 23 336.

²Among SNS users. Stratum number: age 20–59 years, 18 816; ≥60 years, 6318; low digital literacy, 5921; high digital literacy, 19 213.

^*p<0.05,

^**p<0.01,

^***p<0.001.

Participants aged ≥60 years had generally higher aORs for loneliness across SNS behaviors than adults aged 20–59 years. The largest age difference was observed for communication priority: the aOR for “yes” was 1.08 (95% CI, 0.94 to 1.25) in adults aged 20–59 years and 1.89 (95% CI, 1.60 to 2.22) in adults aged ≥60 years. The inverse association between overall SNS use and loneliness, which was significant among adults aged 20–59 years (aOR, 0.84; 95% CI, 0.75 to 0.95), was attenuated and non-significant among adults aged ≥60 years (aOR, 0.95; 95% CI, 0.85 to 1.07). Positive associations for endorsement of active use, information-seeking use, and communication priority were non-significant in adults aged 20–59 years but significant in adults aged ≥60 years (aORs for “yes”: 1.04 vs. 1.83, 0.97 vs. 1.58, and 1.08 vs. 1.89, respectively). For cumulative SNS behaviors, adults aged ≥60 years had higher aORs at most SBS levels, with the difference particularly pronounced at SBS=3 (aOR: 2.81 vs. 4.62).

Higher digital literacy was associated with stronger positive associations between SNS behaviors and loneliness. This contrast was especially prominent for social comparison (aOR for “yes”: 1.24 for low digital literacy vs. 1.57 for high digital literacy) and exclusionary feeling (aOR for “yes”: 1.40 vs. 1.71). For SBS, each additional behavior was associated with 19% higher odds of loneliness in the low-digital-literacy group (aOR, 1.19; 95% CI, 1.13 to 1.24) and 28% higher odds in the high-digital-literacy group (aOR, 1.28; 95% CI, 1.24 to 1.32).

DISCUSSION

Using a large, representative dataset from the 2024 Seoul Survey, this study examined associations between SNS behaviors and loneliness in an urban population. The findings clarify how SNS use and specific SNS behavioral patterns are associated with perceived loneliness among Seoul residents.

Contrary to a meta-analysis suggesting a weak positive correlation between general SNS use and loneliness [19], this study found that overall SNS use among Seoul residents was associated with lower odds of loneliness than non-use. This finding does not support a simple displacement interpretation, which posits that online interactions displace more intimate offline social ties and thereby increase loneliness [20]. Instead, it is more consistent with the view that SNS can facilitate connection and mitigate feelings of isolation. Notably, this pattern may reflect the socio-cultural context of highly connected urban populations, such as Seoul [21], or characteristics specific to the survey period.

However, a more nuanced understanding emerged when the qualitative dimensions of SNS engagement were examined, suggesting that specific SNS behavioral patterns are associated with loneliness in ways not captured by overall SNS use alone.

Although the active social media use hypothesis predicts that active engagement may reduce loneliness by increasing interpersonal satisfaction [22], our study found that active SNS use for sharing daily life was associated with higher odds of loneliness. This finding is consistent with recent research suggesting that even active engagement may be associated with negative psychological outcomes, potentially through mechanisms such as social comparison or fear of missing out (FOMO) [23]. Conversely, other studies have proposed that active SNS use can foster social support and self-esteem, thereby mitigating loneliness [11,24]. These mixed findings suggest that the effect of active use may vary according to the context of SNS engagement and individual psychological vulnerability.

In a counterintuitive but statistically non-significant pattern, information-seeking use of SNS showed a marginal tendency toward higher odds of loneliness after adjustment for covariates, tempering assumptions about the benefits of instrumental social media use [25]. This finding contributes to research showing complex associations between social media use patterns and psychological outcomes and contrasts with evidence that instrumental SNS use can support relationship formation [26]. This complexity is also relevant for older adults, whose use of social media for information is often driven by social motivations and relationship needs [27]. However, the “information-seeking use” measure in this study was based on trust in SNS-acquired information and may not fully capture actual information-seeking behavior.

Prioritizing SNS communication over offline interaction was significantly associated with higher odds of loneliness. This finding partially supports the displacement hypothesis, suggesting that valuing digital over in-person ties may exacerbate loneliness, and indicates that other specific usage patterns may confound or mediate this association [28].

Across all 5 behavioral domains, SNS users who did not endorse a given behavior had lower odds of loneliness than SNS non-users (aORs for “no”: 0.78 to 0.84), whereas users who endorsed these behaviors had higher odds (aORs for “yes”: 1.11 to 1.66). This pattern suggests that SNS use is not inherently harmful; rather, restrained and selective SNS use without problematic engagement styles may be more favorable than complete non-use [11]. These findings indicate that the qualitative style of engagement, rather than SNS use per se, is important for understanding loneliness.

Consistent with the passive social media use hypothesis and a broad range of existing research [29], social comparison and feelings of exclusion were among the strongest correlates of loneliness. This finding aligns with prior work conceptualizing FOMO as a pervasive apprehension of being absent from rewarding experiences that others are having, closely linking it to feelings of exclusion and unmet belonging needs [30]. SNS environments provide ample opportunities for upward social comparison—comparing oneself with people perceived as superior—which can be associated with envy, lower self-esteem, and poorer psychological well-being [31]. Among users prone to social comparison and exclusion sensitivity, SNS interactions may amplify perceived discrepancies between desired and actual social connections, thereby intensifying loneliness [8,32].

This study also observed a dose–response pattern in which higher SBS values were associated with progressively higher odds of loneliness, similar to findings for quantitative SNS use patterns [11,33]. This finding underscores the cumulative relevance of qualitative SNS behaviors for loneliness. The apparent attenuation among participants with 4–5 problematic SNS behaviors suggests that SNS-related associations may vary by individual characteristics and use patterns, consistent with the “rich-get-richer” versus “poor-get-poorer” framework [28,34]. At moderate levels, displacement mechanisms, such as reduced offline interaction, may predominate, particularly among socially vulnerable individuals. At extreme levels, some users may already have reached a high-loneliness plateau, whereas others may engage in compensatory SNS use that does not reduce loneliness [8]. Overall, these findings are consistent with the coexistence of stimulation and displacement processes, with net outcomes depending on individual susceptibilities and contexts [34]. Importantly, the SBS can be interpreted as a cumulative index of dependency-like or problematic SNS engagement, aggregating tendencies such as prioritizing online communication, intensive self-presentation, social comparison, and feelings of exclusion [18,32], rather than as an index of the frequency or duration of SNS use.

The stratified analyses further highlighted age-specific heterogeneity. Among adults aged 20–59 years, overall SNS use retained an inverse association with loneliness, and several problematic behaviors showed weaker or non-significant associations, suggesting that SNS may provide compensatory social resources in this group. In contrast, adults aged ≥60 years showed consistently stronger associations between problematic behaviors and loneliness, and the inverse association for overall SNS use was attenuated and no longer significant. These age-specific patterns indicate that older adults may be more vulnerable to adverse aspects of SNS engagement and may require more targeted guidance on healthy SNS use.

Digital literacy emerged as another key factor. Unexpectedly, participants with higher digital literacy showed stronger associations between loneliness and problematic SNS behaviors, contrasting with prior research linking higher digital literacy to reduced loneliness or improved psychological well-being [12,35]. This pattern may reflect an important distinction between functional and critical digital literacy: functional digital literacy refers to technical skills for operating platforms, whereas critical digital literacy involves self-regulation, awareness of persuasive design, and understanding of algorithmic curation. Research among Chinese adolescents demonstrated that functional digital literacy predicted internet addiction, whereas critical digital literacy—particularly self-management ability—negatively predicted addiction [36]. Since our study measured digital literacy primarily as functional competency, higher technical skills may have facilitated greater SNS exposure and more dependency-like engagement without a corresponding increase in self-regulatory capacity [37]. These findings suggest that functional digital literacy alone is insufficient to mitigate SNS-related loneliness. Furthermore, they highlight the critical need for targeted SNS-specific education that develops critical digital literacy, including platform literacy, awareness of persuasive technology, understanding of social comparison mechanisms, and self-regulation strategies [13,38]. Such comprehensive education is essential to foster positive SNS engagement and prevent loneliness from worsening.

Strengths and Limitations

A strength of this study is its use of a large, representative metropolitan sample, which provides evidence relevant to policy formulation and evaluation. The use of secondary data from the 2024 Seoul Survey, collected through face-to-face household interviews and stratified cluster sampling, may also have reduced missing responses and potential bias. A key implication is that qualitative patterns of SNS use, rather than quantity of use alone, may be more strongly associated with loneliness, emphasizing the need to move beyond simple measures of SNS usage time or frequency and to focus on the qualitative aspects of usage purpose and behavior. This study also highlights the need for tailored interventions for vulnerable populations, particularly older adults, given the stronger associations observed in this group and the finding that higher digital literacy may be associated with greater loneliness in the context of problematic SNS behaviors. Accordingly, appropriate digital education and policy support are crucial to effectively leverage SNS as a tool for loneliness alleviation.

Despite these contributions, this study has several limitations. First, because of its cross-sectional design, causal relationships between SNS behaviors and loneliness cannot be established. Longitudinal studies are needed to fully explore the dynamic and reciprocal nature of these associations. Second, the study used data from the 2024 Seoul Survey, which included only residents of Seoul. This geographic specificity may limit the direct generalizability of the findings, as patterns of SNS use, problematic behaviors, and loneliness may differ across regions and cultural contexts. Third, reliance on self-reported data may have introduced social desirability or recall bias. Future research could incorporate objective measures of SNS use, such as platform analytics or experience sampling methods, to provide a more comprehensive and accurate understanding. Fourth, loneliness was assessed using brief 5-point Likert items. This approach may introduce a risk of misclassification bias, because the validity of the selected items for measuring loneliness has not been fully established, especially compared with widely used multi-item standardized measures such as the revised UCLA Loneliness Scale [39] or the de Jong Gierveld Loneliness Scale [40]. Future research should consider adopting more robust and validated multi-item scales to enhance reliability and comparability. Finally, this study did not examine the specific characteristics and functional differences of individual SNS platforms, such as Naver Band, Facebook, and Instagram. Although these platforms share common features, including connecting users and sharing content, their primary content formats and interaction methods can differ substantially. Future research on platform-specific nuances would provide more detailed insight into the differential impacts of SNS on loneliness.

This study found that qualitative patterns of SNS engagement, rather than SNS use itself, were associated with loneliness. Although overall SNS use was modestly associated with lower odds of loneliness, specific problematic behaviors were associated with higher odds of loneliness, especially among older adults and participants with higher digital literacy. Accordingly, targeted, behavior-focused digital education appears necessary to leverage SNS for loneliness alleviation.