Self-esteem only goes so far: the moderating effect of social media screen time on self-esteem and depressive symptoms

Abstract

This study assessed the independent association of self-esteem and social media screen time on depressive symptoms, as well as the moderating role of social media screen time in the relationship between self-esteem and depressive symptoms. The Mobile Screen Time Project was a cross-sectional, web-based survey conducted from March to May of 2019. 437 U.S. college students were recruited via social networks from two institutions of higher education. Multivariable logistic regression assessed the associations between self-esteem and average daily social media time with depressive symptoms; an interaction effect was explored. Self-esteem had an inverse association (AOR = 0.87, 95% CI: 0.80–0.94) and daily social media time had a significant association with depressive symptoms (AOR = 1.11, 95% CI: 1.02–1.22) after adjusting for sexual and gender status, race/ethnicity, age, social status, and insomnia. We found a significant moderating effect (p = 0.016) of daily social media time. The more time spent on social media, the less protective self-esteem was against depressive symptoms. Those suffering from depressive symptoms or low self-esteem may benefit from reducing their social media use, intentionally exposing themselves to positive content and leveraging peer-to-peer social support through social media to create a sense of belonging.

Introduction

Social media use continues to increase among the U.S. population, with 84% of young adults aged 18–29 using at least one social media site as of February, 2021.¹ Some of this increased use can be attributed to the COVID-19 pandemic, with usage peaking during stay-at-home mandates.² Because social media has become such an integral part of so many people’s lives, there has been a significant increase in research examining its impact on mental health.³ There is a general consensus in the literature that social media use increases risk of both depressive symptoms and anxiety and has negative effects on well-being and life satisfaction.^3–7 Experimental studies have shown that reductions in time spend on social media per day can result in reduced depressive symptoms and increased life satisfaction.^8,9 In the past decade, there has been a nationwide trend of increasing rates of depression in college students, and in some cases rates have doubled.¹⁰ According to the American College Health Assessment, 23.45% of undergraduate students received a depressive disorder diagnosis in Spring 2021.¹¹

Another contributor to depressive symptoms is low self-esteem, or the subjective evaluation of an individual’s self-worth as a person.¹² Self-esteem has significant impacts on social and emotional outcomes in life, with high self-esteem linked to positive outcomes such as healthy relationships, improved coping skills, and occupational success.¹³ Low self-esteem, however, is linked to many negative outcomes such as depression, physical health problems and antisocial behavior.^14–16 There are a few theoretical models that propose the association between low self-esteem and depression, the most supported of which being the vulnerability model. This model states that low self-esteem is not only correlated with depression but can also be considered a vulnerability factor, where negative beliefs of oneself play a fundamental role in the etiology of depression. This model finds low self-esteem as a causal risk factor for depression, predisposing an individual to this condition.¹² Self-esteem is important to humans because it serves as a subjective monitor that measures the extent to which an individual feels valued as a member of relationships, as proposed by the sociometer theory.¹⁷

While social media has increased society’s connectedness, it has also greatly increased opportunities for social comparisons, specifically appearance-related comparisons. These social comparisons can have a cumulative effect on the satisfaction of individuals with themselves and their lives.¹⁸ Users can engage in selective self-presentation on social media, only showing aspects of themselves and their lives that they want to be seen by others. This type of behavior is considered a method of reinforcing the connection between self-worth and physical appearance.¹⁹ Those with higher tendencies to compare themselves to others exhibit lower self-esteem and poorer self-perception than those with less of a tendency to compare.²⁰ Social comparisons can play a medicating role in the relationship between self-esteem and social media addiction, whereby social media addiction can result in more upward comparisons with others and therefore a lower level of self-esteem.²¹ Therefore, high levels of social media use have been associated with low self-esteem.²² Despite the association of increased social media use and low self-esteem with depressive symptoms, no study has assessed whether social media use and self-esteem are independent risk factors for depressive symptoms, nor have they examined the potential moderating role of social media use in the relationship between self-esteem and depressive symptoms. Therefore, this study aims to address the association that self-esteem and social media screen time independently have on depressive symptoms along with the potential moderating role that social media screen time plays in the relationship between self-esteem and depressive symptoms. We hypothesize that self-esteem will be inversely and independently associated with depressive symptoms, social media screen time will be positively and independently associated with depressive symptoms, and that increased social media screen time likely attenuates the protective relationship between self-esteem and depressive symptoms.

Materials and Methods

Sample

The Mobile Screen Time Project, a cross-sectional, web-based survey, was conducted from March to May of 2019. The goal of the project was to examine the relationship between how college students use their smartphones and various physical and mental health outcomes. Individuals were eligible for the study if they were aged 18 years or older, spoke English, and were enrolled in a U.S.-based institution of higher education. A convenience sample was recruited through craigslist.org, Amazon Mechanical Turk, Facebook, and Twitter. Invitations to participate were also distributed via email listservs from two colleges located in the northeast U.S. More detailed methods of the Mobile Screen Time Project have been published previously.²³ As an incentive, six participants were randomly selected to receive a $25 electronic gift card. A total of 565 individuals initiated the survey and 437 (77.3%) met study eligibility criteria and were included in this study. Sample recruitment and study protocols were approved by two university Institutional Review Boards. All participants provided electronic informed consent.

Measures

The primary outcome was depressive symptoms. Depressive symptoms were measured using the Center for Epidemiological Study Short Depression Scale (CES-D10).²⁴ The CES-D10 contains 10 items regarding past week experiences of symptoms related to the development of depression (α = 0.83). For example, items include “I felt hopeful about the future” and “I felt lonely.” The response options ranged from were “rarely or none of the time,” which was coded as 0, to “all of the time,” coded as 3. Reverse scoring was implemented for two of the items. Total continuous depressive symptom scores could range from 0 to 30 with higher scores suggesting greater severity of symptoms. Consistent with the literature, a cut-off of 10 or higher was indicative of having depressive symptoms.²⁵The CES-D10 has shown strong test-retest reliability and convergent validity in youth and adult populations.^26–28 In this sample, the interitem correlation according to Cronbach’s alpha was α = 0.84.

The primary exposure was self-esteem. Self-esteem was measured via the self-reported Rosenberg Self-Esteem Scale.²⁹ This is a measure of global self-esteem including 10 items about both positive and negative feelings of self worth. All items are assessed on a 4-point Likert scale from strongly agree to strongly disagree (e.g. 1 to 4). Five of the items, assessing negative feelings, were reverse scored according to scoring instructions. The final continuous summary score potentially ranges from 10 to 40, with higher scores indicating higher self-esteem. This measure is widely used among the adult population and is the standard measure for self-esteem in psychological research.³⁰ Studies suggest strong test-retest reliability of 0.82 and high internal consistency (α = 0.88).³¹ In this sample, the interitem correlation according to Cronbach’s alpha was α = 0.90.

Self-reported daily social media time in hours was used both as an exposure and examined as a moderating variable in the study. This was assessed via the question, “On average, how much time do you think you spend on your mobile phone on social networking each day (in hours)?” Response validation was implemented for quality assurance requiring responses ranging between 0 and 24 hours.

Several covariates and potential confounders were measured including insomnia, age, sexual and gender identity (heterosexual cisgender males, heterosexual cisgender females, sexual and gender minorities), race/ethnicity, and relative socioeconomic status. A growing body of literature suggests that insomnia is a significant risk factor for depressive symptoms.^32–34 Insomnia was measured via the Insomnia Severity Index (ISI), a 7-item self-report questionnaire of insomnia symptoms with 5-point Likert-scale responses (e.g. 0 to 4) ranging from “None” to “Very Severe.” For example, items include the severity of your insomnia problem over the last two weeks: “difficulty falling asleep” and “problems waking up too early.” Total summary insomnia symptom scores potentially ranged from 0 to 28, with higher scores reflecting greater severity of insomnia symptoms. Summary scores of 15 or above indicated either moderate or severe clinical insomnia. The ISI is a valid and reliable instrument and holds excellent internal consistency and with a Cronbach alpha of about 0.90.³⁵ In this sample, the interitem correlation according to Cronbach’s alpha was α = 0.86. Race/ethnicity was dichotomized as non-Hispanic white or Black, Indigenous and People of Color (BIPOC). Relative socioeconomic status (SES) was measured using the Macarthur Scale of Subjective Social Status, which assesses a person’s perceived rank relative to others in their community, where 1 indicates being the “worst off” and 10 indicates being the “best off.”³⁶

Statistical Analysis

Descriptive statistics were used to describe the full sample. We then compared the characteristics between participants who do and do not have depressive symptoms. Characteristics were reported as counts with percentages for categorical variables and means with standard deviations for continuous variables. Comparisons between groups were evaluated either by chi-square tests for categorical variables or t-tests for continuous variables. Simple logistic regressions were conducted to assess the crude odds ratios for each included characteristic. A multivariable logistic regression model controlling for both exposures and covariates was used to assess the association between self-esteem and depressive symptoms as well as between daily social media time and depressive symptoms. We calculated adjusted odds ratios (ORs) with their 95% confidence intervals (CIs) from the model. An interaction term of self-esteem score and average daily social media time in hours was then created and added to the model to test if social media time was an effect modifier. Finally, we plotted the fully adjusted predicted probabilities of depressive symptoms from self-esteem score depending on the average daily social media time in hours. Statistical significance was assessed at α = 0.05. All analyses were conducted in Stata 15.0.³⁷

Results

Table 1 presents the characteristics of the entire study sample (n = 437) and by depressive symptoms. The sample was predominantly heterosexual cisgender female (50%) and non-Hispanic white (51%). Two hundred twenty-two (51%) young adults met the criteria for depressive symptoms and young adults reported average daily social media time of 3.11 hours. Overall, sexual and gender minorities (p = 0.013), BIPOC young adults (p = 0.005), people with lower social status (p < 0.001), and people with insomnia (p < 0.001) were more likely to have depressive symptoms. The mean self-esteem score was significantly lower (p < 0.001) and the mean average daily social media time was significantly higher (p < 0.001) among participants with depressive symptoms. However, no significant difference was found by age of the two groups (p = 0.155).

TABLE 1.

Characteristics of participants by depressive symptoms

	Total	Depressive Symptoms
	N = 437 N (%)	N = 222 N (%)	P value*
Sexual and Gender Status			0.013
Heterosexual Female	218 (50)	104 (47)
Heterosexual Male	131 (30)	61 (27)
Sexual or Gender Minority	88 (20)	57 (26)
Race/Ethnicity			0.005
White, non-Hispanic	224 (51)	99 (45)
BIPOC	213 (49)	123 (55)
Age [mean (SD)]	21.2 (0.10)	21.4 (0.13)	0.155
Social Status [mean (SD)]	6.2 (0.09)	5.7 (0.13)	<0.001
Insomnia			<0.001
Yes	95 (22)	80 (36)
No	342 (78)	142 (64)
Self-Esteem [mean (SD)]	23.0 (0.13)	22.4 (0.20)	<0.001
Average Daily Social Media Time in hours [mean (SD)]	3.11 (0.13)	3.56 (0.20)	<0.001

^*P values were generated using chi-square tests or t-tests.

Table 2 shows the simple and multivariable logistic regression models’ estimates of the association between participants’ characteristics and depressive symptoms. For simple logistic regressions of depressive symptoms, significant associations were observed for sexual and gender minority status, being a BIPOC young adult, social status, insomnia, self-esteem, and average daily social media time. In the full multivariable model, all characteristics remained significant except for being a BIPOC young adult. Self-esteem had an inverse association with depressive symptoms (AOR = 0.87, 95% CI: 0.80–0.94) after adjusting for sexual and gender status, race/ethnicity, age, social status, insomnia, and daily social media time. Similarly, daily social media time had a significant association with depressive symptoms (AOR = 1.11, 95% CI: 1.02–1.22) after adjusting for sexual and gender status, race/ethnicity, age, social status, insomnia, and self-esteem. When the interaction term of self-esteem and average daily social media time was added into the model, we found a significant moderating effect (p = 0.016) of daily social media time between self-esteem and depressive symptoms. Generally, for each additional hour of daily social media time, the inverse association between self-esteem and depressive symptoms became weaker (Table 2, Figure 1).

TABLE 2.

Simple and multivariable logistic regression for depressive symptoms (N=437)

	Simple Logistic Regression		Multivariable Logistic Regression
	OR	95% CI	AOR	95% CI
Sexual and Gender Status
Heterosexual Female	1	ref	1	ref
Heterosexual Male	0.96	0.62–1.47	1.33	0.81–2.20
Sexual or Gender Minority	2.02	1.21–3.36	1.82	1.02–3.25
Race/Ethnicity
White, non-Hispanic	1	ref	1	ref
BIPOC	1.73	1.18–2.52	1.45	0.94–2.25
Age	1.07	0.98–1.17	0.99	0.89–1.10
Social Status	0.7	0.63–0.79	0.73	0.64–0.83
Insomnia
Yes	7.51	4.16–13.58	5.83	3.12–10.90
No	1	ref	1	ref
Self-Esteem	0.85	0.79–0.92	0.87	0.80–0.94
Average Daily Social Media Time in hours	1.15	1.06–1.25	1.11	1.02–1.22

FIGURE 1.

The moderating effect of daily social media time between self-esteem score and risk of having depressive symptoms (N=437)

Discussion

The first two aims of this study were to analyze the relationship that self-esteem and social media screen time independently have with depressive symptoms. The literature suggests that having low self-esteem is associated with negative outcomes such as depression, while having high self-esteem is associated with more positive outcomes.^13,14 The findings of this study are consistent with existing literature, identifying an inverse association between self-esteem and depressive symptoms. This means that as self-esteem decreases, the risk of having depressive symptoms increases. Current literature also suggests that increasing social media usage is associated with a higher risk of experiencing depressive symptoms.^3–7 This study confirms the hypothesis that daily social media time is an independent risk factor for depressive symptoms, with a significant association identified between the two.

Because it has been established that self-esteem and social media time are independent risk factors for depressive symptoms, there is a potential that social media screen time could change the nature of the relationship between self-esteem and depressive symptoms. Previous literature has identified self-esteem as a moderating factor between social media use and systemic inflammation, where a decrease in self-esteem strengthened this positive association.³⁸ Inflammation has also been found to increase the risk of depression, where those with depression are more likely to have elevated levels of inflammatory markers.^39,40 This study identified daily social media time as a moderating factor between self-esteem and depressive symptoms. Each additional hour of daily social media time weakened the inverse association between self-esteem and depressive symptoms. Essentially, the protective effect that high self-esteem has against depressive symptoms weakens as daily time spent on social media increases.

Research on the directionality of the relationship between social media time and self-esteem is lacking.⁴¹ It can be hypothesized that those who spend more time on social media are more likely to have a lower self-esteem due to comparisons with others based on their appearance, lifestyle, social life, or apparent wealth.⁴² Being selectively presented with only the best aspects of the lives of others can cause one to feel inadequate, leading to lower levels of life- and self-satisfaction.¹⁸ Opposing this, it can also be argued that having lower self-esteem is associated with more time spent on social media daily.⁴³ Some may find that receiving attention such as likes or comments provides a temporary boost in self-esteem.⁴⁴

Daily social media screen time is significantly associated with depressive symptoms, where more time spent on social media results in a higher risk of experiencing depressive symptoms. Those who are suffering from depressive symptoms may benefit from reducing their daily time spent on social media. Spending more time on social media each day puts people at increased risk of experiencing depressive symptoms, whether they have low or high self-esteem. Self-esteem can only protect one from these symptoms to an extent, as shown by the main outcome of this study. There are a few methods one can take to effectively reduce their daily social media use, including reducing or eliminating notifications, removing the phone from the bedroom at night, and even turning on a grayscale setting to make the phone less aesthetically pleasing.⁴⁵ All of these interventions serve as nudges to reduce screen time by lowering motivation, ability, and prompts to use smartphones, three components that reinforce habitual behaviors.⁴⁶

The addictive nature of social media makes it difficult for many to cut back their use.⁴⁷ In order to reduce the harms associated with social media use, solutions should focus on altering the ways that users are interacting with it. The ways in which people are using social media may be more impactful on overall well-being than simply the frequency of their use.⁴⁸ Many platforms allow users to choose which accounts they want to follow, and to unfollow or block accounts that they do not want to interact with. This freedom allows users to cater their social media feeds to fit their own unique needs. Users should focus on exposing themselves to positive, uplifting content that can improve self-esteem and mental health. Social media has the potential to teach, promote, and provide ways to improve mental health in users while also providing a medium for peer-to-peer support that encourages feelings of belonging and connectedness.^49,50

Other accounts are focused on relaxation, meditation, or even reminders to put your phone down and go outside. If users can learn which content improves or worsens their mood and mental health, they can begin to use social media in a mindful fashion to reduce the harms associated with it.⁴⁸ This also includes sharing positive content with others, and reflecting on how content will make others feel. Finally, social media is an effective tool for building and maintaining real relationships with others. Creating a network of diverse, interpersonal connections can reduce the harm of social media and increase subjective well-being.⁵¹

Despite the novel and significant findings of this study, it is not without limitations. This is a cross-sectional study which limits our ability to assess causality–there may be a bidirectional effect between self-esteem and social media screen time, as well as depressive symptoms. Data were also primarily self-reported, increasing risk of measurement error. In particular, social desirability bias may have cause participants to under-report their social media screen time. However, self-reported screen time was consistent with objectively measured screen time in a similar study.²³ It is important to note that this study measured social media mobile screen time and it is possible that social media accessed by a different device such as a computer may not be captured here. Also, despite self-reports, self-esteem and depressive symptoms were assessed with valid and reliable instruments. By dichotomizing race/ethnicity in this study, there may be important race-related findings that have been missed, as prior studies suggest Black and “other” race adults may have more average social media screen time than White users,⁵² Asian Americans may have lower self-esteem than other minority racial and ethnic groups,⁵³ and BIPOC young adults in this study had a higher rate of depressive symptoms. Future studies on this topic should further examine the role of race and ethnicity. Another potential limitation is that this study includes a convenience sample and may not be generalizable to all college students in the U.S. Finally, some potential confounders were left unmeasured such as physical activity or the motives for using social media. Regardless of these limitations, these study findings add a better understanding of the role of social media screen time in the protective effect of self-esteem against depressive symptoms.

Conclusions

Researching the role of social media is a crucial step towards understanding the true scope of mental health disorders plaguing young adults. By examining the mental health outcomes of social media users through the Mobile Screen Time Project, this study established that low self-esteem and increasing social media usage are independent risk factors for depressive symptoms. This study also demonstrates that daily social media time plays a moderating role in the relationship between self-esteem and depression, where each additional hour of daily social media time weakens the inverse association between self-esteem and depressive symptoms. The results of this study allow recommendations to be made that may promote healthier social media habits such as decreasing social media time, intentionally exposing oneself to positive content, removing negative content, becoming a source for positive material, and fostering real relationships with others. Future research should study the effects of varying social media motives and behaviors, such as those aforementioned, on mental health outcomes in order to inform users on the healthiest ways to interact with this type of technology.