Overestimates of social media addiction are common but costly

Abstract

Overuse of social media is commonly termed a behavioral addiction. However, there is reason to believe that users overestimate their addiction; instead, frequent use largely builds habits to automatically activate, scroll, post, and react on social media. The present research (N = 1204) explored the implications of overusing the addiction label. In Study 1, a national quota sample of Instagram users (N = 380) overestimated their addiction to the app, and those who perceived themselves as more addicted (but not more habitual) experienced less ability to control their use. We show that the perception of addiction likely arises from popular media’s frequent labeling of social media as addictive (vs. habit forming). Study 2 (N = 824) demonstrated experimentally that framing frequent Instagram use as an addiction has deleterious consequences for user self-efficacy, including reducing perceived control over social media use and increasing self-blame for overuse. In addition, misperceiving excessive social media use as addictive potentially diverts users from effective strategies that could be used to curb overuse habits.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-27053-2.

Overestimates of social media addiction are common but costly

Excessive social media use has been likened to an addiction, with the U.S. Surgeon General claiming that excessive use has neurological associations similar to substance abuse¹; but see^2,3. Although several studies have addressed the prevalence of social media addictions, most have relied on college student convenience samples (see⁴, meta-analysis) or on participants experiencing other addictions or psychiatric disorders (e.g.,⁵).

There is good reason to question whether overuse of this technology is necessarily a clinically relevant pathology^6,7. Social media research has often used the term addiction to describe heavy use without further assessment of any pathological symptoms^6,8. In clinical studies, addiction, or substance use disorder, is diagnosed from a cluster of symptoms that include impaired control, craving and dependence, withdrawal when not using, conflict with other activities, and hazardous or risky use⁹. Wantings and cravings for a substance are also central to incentive-sensitization theory, which locates the essence of addiction in compulsive motivations to use a substance without necessarily liking it¹⁰.

Another reason to question the validity of the addiction label is that frequent use of social media is tied to positive as well as negative outcomes. Indeed, Facebook users who self-reported greater life conflict and lack of control (two symptoms of addiction) also reported greater positive impacts of Facebook use¹¹. Although some meta-analyses have concluded that social media promotes lower well-being (e.g., reduced life satisfaction¹², and greater ill-being (e.g., increased depression symptoms¹³, these effects are inconsistent and typically weak (e.g.,^14–16). The varied effects reflect in part the heterogeneous experiences of users, with each user seeing different content when browsing^17,18. As a result, many effects appear to be person, group, and behavior-specific. Furthermore, adverse outcomes may be overrepresented in research given that some studies focus a priori on the troublesome aspects of social media by assessing the effects solely of problematic use (see^18,19).

The present research estimates the incidence of addictive pathology among Instagram (and TikTok, supplemental materials) users and compares this to the number of people who believe they are addicted. We also estimate the number of people who believe they are habitual users and test whether habits are distinct from addiction. We go on to consider whether popular news media promotes ideas of addiction that might influence users’ beliefs about their own use. Finally, we evaluate whether labeling frequent use as addiction has negative consequences in (a) limiting users’ perceived control over social media and (b) driving self-blame and other negative feelings related to frequent use.

Habits in social media use

In contrast with addiction, considerable evidence has shown that frequent social media use promotes habit formation (e.g.,^20–24). Habits are commonly defined as cognitive associations between stimuli and responses that develop as people repeat rewarded actions in stable contexts^25,26. Once a habit has formed, perception of the context automatically activates the response in memory. Then, through ideomotor processes, people often act on the response in mind. On social media, habits build through repeated use, and these habits then automatically drive how people post, share, and scroll^20,27.

Confusion between social media habits and addiction is understandable, given that some components of habit play a role in the development of substance abuse disorders (e.g.,^28,29). That is, habit learning can be a building block of addiction, with the transition to compulsion progressing through excessive use that comes to trigger additional physiological and psychological mechanisms (e.g., increasing doses required for response). Habits are also similar to addiction in that both can produce unwanted or unintended consequences. In the case of habits, this often occurs because habits are activated automatically by perceiving familiar context cues and thus persist even when people’s goals change (e.g., action slips³⁰). The present research builds on distinctions between habitual and more problematic social media and technology use (e.g.,^31,32) by identifying the incidence of habits and addiction among active users and clarifying the negative outcomes that arise from applying a label of addiction, whether appropriate or not.

The present research

Study 1 estimated the prevalence of addictive symptoms among active, at least monthly Instagram users with a quota-based sample designed to be representative of the U.S. adult population (percentage-matched on gender, age, political views, and race; Prolific, www.prolific.com) and an often-used measure of social media addiction, the Bergen Social Media Addiction Scale³³; adapted as the Bergen Instagram Addiction Scale, BIAS, in Study 1 and for TikTok in the supplementary materials. This scale assesses addiction as a multicomponent syndrome with symptoms such as urges to use and withdrawal when not using³⁴. Given that frequent social media use has not been clearly tied to pathological symptoms, with effects ranging from positive to negative, we predicted that only a limited number of (10% or less) of users would be classified as at risk for addiction. We also assessed users’ self-reports of addicted and habitual use with single-item questions assessing the degree to which they label their Instagram use as an addiction/a habit. We predicted that fewer participants would be classified as at risk on the BIAS symptom scale than would self-report addiction.

Second, to understand the implications of applying the addiction label to social media use, we assessed a variety of concomitants of self-reported addiction. Self-labeling of clinical conditions (e.g., I think I’m depressed) has proved to be associated with maladaptive responses, including lowered self-efficacy and perceived control over the pathology^35–37. If labeling frequent social media use as an addiction has similar effects, then self-reports of addiction should be tied to negative outcomes. Thus, we predicted that users who indicated stronger beliefs that they were addicted would report lowered control over their behavior, increased self-blame for excessive use, and more unsuccessful attempts to change behavior in the past. Given that habits are not necessarily tied to pathology, these outcomes should be less closely related to self-reported habit.

Finally, we predicted that the urge and withdrawal symptoms of addiction (BIAS symptom scale items) would be less closely associated with self-reported habit than self-reported addiction (see AsPredicted #184851 at https://aspredicted.org/m4f4-6hzr.pdf, for full Study 1 preregistration).

Study 1: addiction among a quota sample of U.S. Instagram users

Method

All protocols, measures, and procedures were approved by the IRB at the University of Southern California (Protocol ID: UP-23–00585). Data collection took place in 2024. All experiments were performed in accordance with the rules of these approved protocols and relevant guidelines and regulations. Informed consent was obtained from all participants involved in this research, and anyone who did not give informed consent was not allowed to proceed beyond the consent form. Materials, data, and code are available via OSF  https://osf.io/3tcqm/overview?view_only=6f3c94fcedd0475ca08638993440b20a.

Participants

A U.S. quota sample of 380 active adult Instagram users (percentage-matched based on gender, age, political views, and race; Prolific, www.prolific.com) was recruited from Prolific. Inclusion criteria included > 90% prior approval rate, U.S.-based, English-speaking, adult, active Instagram account holders in the national sample pool. Participants were compensated at a rate corresponding to California minimum wage, $16.00/hr, at the time of data collection. Excluded from analyses were (a) one duplicate response and (b) one participant who exceeded our preregistered maximum of 100 h per week of Instagram use (see AsPredicted #184851). Two participants who indicated 0 h per week of use were retained because their elimination did not alter the results. See supplementary materials for demographics.

Measures

BIAS (Bergen Instagram Addiction Scale; adapted from³³ symptom scale. On a 5-point scale from 1 (very rarely), 2 (rarely), 3 (sometimes), 4 (often), to 5 (very often), participants indicated, “How often during the last six months have you:” (a) salience: “spent a lot of time thinking about Instagram or planned use of Instagram?”; (b) urge: “felt an urge to use Instagram more and more?”; (c) mood modification: “used Instagram to forget about personal problems?”; (d) failure to cut down on use: “tried to cut down on the use of Instagram without success?”; (e) withdrawal: “become restless or troubled if you have been prohibited from using Instagram?”; and (f) life conflict: “used Instagram so much that it had a negative impact on your job/studies?”. Total scores ranged from 6 to 30 (alpha = 0.85).

Habit Automaticity (Self Report Behavioral Automaticity Index³⁸. On 4-item scales ranging from 1 (strongly disagree) to 7 (strongly agree), participants rated items like, “Right now, using Instagram feels like something…” “I do automatically.” “.I start doing before I realize I’m doing it.” Total scores ranged from 1 to 7 (alpha =.95).

Prior Instagram use frequency

Participants indicated how many hours per week they used Instagram (on a sliding scale of 0 to 100 h). If unsure about the number of hours, they were directed to check their screentime (iPhone) or Digital Wellbeing Tools (Android). 102 participants reported checking. Participants reported a range from 0 to 74 h per week (M = 10.47).

As expected, frequency correlated with the SRBAI habit strength measure, r(378) = 0.32, p <.001, indicating the reliability and consistency of these two measures of habit strength (see^20,21.

Self-Reported addiction

Participants rated, “Right now, I feel addicted to Instagram,” on a scale from 1 (strongly disagree) to 7 (strongly agree). Responses ranged from 1 to 7.

Self-Reported habit

Participants rated, “Right now, I feel like I have a habit to use Instagram,” on a scale from 1 (strongly disagree) to 7 (strongly agree). Responses ranged from 1 to 7.

Positive/Negative feelings about Instagram use

On three semantic differential items, participants rated: “My Instagram use is…” 1 (Harmful/Bad/unproductive), 2 (Neither harmful/bad/unproductive nor beneficial/good/productive), and 3 (Beneficial/Good/Productive). Aggregated scores ranged from 3 to 9 (alpha = 0.81).

Perceived control

Participants indicated how much control they have over their Instagram use on a scale from 1 (None at all) to 5 (Great control). Scores ranged from 1 to 5.

Number of past control attempts

Participants indicated how often they had tried in the past to quit or control their Instagram use on a sliding scale from 0 to 100. Responses ranged from 0 to 100.

Blame for overuse

On two scales ranging from 1 (Strongly disagree) to 7 (Strongly agree), participants rated, “When I use Instagram too often, it is because… (a) I have failed to control my use and (b) of the way the Instagram app is designed.”

Results

Prevalence of Instagram addicts and habitual users

Figure 1 contrasts participants’ scores on the BIAS symptom scale, which assesses specific clinical symptoms, with a measure of the frequency with which participants self-reported addiction. As anticipated, a very small percentage of users met the criteria based on the addiction symptoms in the BIAS symptom scale. That is, 2% of active users scored in the range “generally considered a warning of potential addiction,” according to guidelines published by the American Psychiatric Association^39,40. This means 9 participants scored 24 or above on the scale, a score that reflects the experience of multiple symptoms (i.e., no ratings below the scale midpoint and at least half of the symptoms experienced very often). Comparable findings emerged when social media addiction was defined from the four symptoms most strongly associated with psychopathology (i.e., mood modification, relapse, withdrawal, conflict³⁴).

Fig. 1.

Distributions of Measures of Addiction to Instagram from Study 1. Note. Frequencies of self-reported Instagram addiction and BIAS symptom addiction.

Participants were more likely to self-report addiction, with 18% of users agreeing at least somewhat that they were addicted to Instagram and 5% indicating substantial agreement (6 or 7 on the 7-point scale, see Fig. 1). Contrasting the 5% indicating substantial agreement with the 2% of users who scored as at risk to addiction implies that over half of Instagram users who self-reported being addicted were not basing their assessments on the set of pathological symptoms clinically used to define addiction. Contrasts between these measures of self-reported addiction and addiction symptoms were similar using other BIAS symptom scale cutoffs (see supplemental materials; BIAS symptom cutoff section).

The most frequently experienced symptom of addiction to Instagram on the BIAS was salience, with 20% of participants indicating that they often or very often spent time thinking about Instagram. The least frequently experienced symptoms were withdrawal (only 4% indicated this occurred often or very often) and life conflict (6%, see Supplemental Figure S1 for BIAS measure distributions and self-reported addiction and BIAS histograms).

In contrast with the addiction measures, participants were more likely to recognize their habits to use Instagram, with about half (49%) reporting that they were habitual Instagram users (5 or greater on the 7-point scale). Self-reports of habit and addiction were substantially correlated (see Table 1), which is understandable given that habit is often considered a component of or a precursor to addictive and compulsive use (e.g.,²⁹).

Table 1.

Means, standard deviations, and correlations from Study 1.

Variables	M	SD	Salience	Urge	Mood modification	Failure to cut down	Withdrawal	Life conflict	Prior Frequency of Instagram Use	Perceived control	Past control attempts	Positive/negative feelings	Self-reported addiction	SRBAI
Salience	2.52	1.14
Urge	2.47	1.15	0.68***
Mood Modification	2.38	1.19	0.51***	0.55***
Failure to cut down	2.02	1.14	0.38***	0.44***	0.51***
Withdrawal	1.55	0.88	0.38***	0.43***	0.48***	0.59***
Life conflict	1.57	0.96	0.34***	0.36***	0.42***	0.62***	0.62***
Prior frequency of Instagram Use	10.47	13.65	0.48***	0.44***	0.42***	0.36***	0.38***	0.38***
Perceived control	1.65	0.88	0.31***	0.38***	0.46***	0.58***	0.52***	0.57***	0.41***
Past control attempts	4.48	10.15	0.18***	0.23***	0.28***	0.43***	0.42***	0.39***	0.36***	0.46***
Positive/negative feelings	6.31	1.64	0.08	0.01	− 0.12*	− 0.30***	− 0.14**	− 0.27***	− 0.02	− 0.27***	− 0.16**
Self-reported addiction	2.42	1.64	0.50***	0.53***	0.55***	0.64***	0.58***	0.62***	0.48***	0.64***	0.40***	− 0.24***
SRBAI	3.49	1.86	0.42***	0.48***	0.53***	0.51***	0.42***	0.44***	0.46***	0.48***	0.29***	− 0.25***	0.67***
Self-reported habit	3.83	1.88	0.54***	0.51***	0.50***	0.47***	0.40***	0.41***	0.51***	0.43***	0.25***	− 0.13*	0.65***	0.68***

Note. All correlations were significant at p <.001. Urge, failure to control, withdrawal, and life conflict are components of the BIAS scale. Correlations were computed using the dataset mean-centered and divided by each measure’s standard deviation (standardized).

Symptoms of Self-Reported addiction vs. Self-Reported habit

Supporting our prediction that habits are a separate determinant of social media use from addiction, the symptom of withdrawal was more highly correlated with participants’ self-reports of addiction than with self-reports of habit (see Table 1, Steiger’s Z = 3.67, p <.001). In addition, life conflict and past failures to cut down use were more strongly related to self-reported addiction than habit (Steiger’s Zs = 4.17 and 3.58, ps < 0.001). Thus, these three addiction symptoms distinguished self-reported feelings of Instagram addiction from habit.

However, contrary to our hypotheses, the urge to use Instagram was not more closely tied to self-reported addiction than to habit. We had assumed that urges refer to addictive cravings and compulsions. However, participants might also have considered the response automaticity of habit to be a kind of urge. If so, then it makes sense that both addiction and habit were tied to urges to use Instagram (see Supplemental Table S1).

Symptoms of addiction but not habit

To further distinguish the symptomatology of addiction from habits, we computed three separate regression models in which these self-reports simultaneously predicted withdrawal, life conflict, and failure to cut down. As anticipated, for withdrawal, self-reported addiction was a strong predictor, = 0.55, p <.001, 95% CI: [0.44, 0.66], df = 377, and habit was not significant (see Supplemental Table S2). Analyses on life conflict and failure to cut down showed similar patterns, with self-reported addiction remaining a strong predictor of these symptoms (, ps < 0.001) and habit having a limited impact (see Supplemental Tables S3 and S4). Thus, these three symptoms were uniquely related to perceived addiction to Instagram but had little relation to ratings of habitual use after controlling for addiction.

Negative implications of addiction but not habit

Perceived control and number of attempts to control

As shown in Table 1, participants’ self-reports of Instagram addiction (vs. habit) were more strongly associated with lower ability to control their Instagram use and had more past attempts to control it (Steiger’s Zs = 4.55 and 3.20, ps < 0.003, for ability to control and control attempts, respectively).

We computed additional analyses to differentiate the effects of self-reported addiction and habit on perceived control. In a regression model predicting control simultaneously from self-reported addiction and habit, addiction proved to be a significant predictor of control, = 0.63, p <.001, 95% CI: [0.53, 0.73], df = 377, and after accounting for this effect, habit had little impact, = 0.02, p =.70, 95% CI: [−0.08, 0.12], df = 377 (see Supplemental Table S5). Similarly, the number of past attempts to control Instagram use was predicted by addiction and not habit when both were entered into a regression model (see Supplemental Table S6). Thus, self-reported addiction but not habit was positively related to both participants’ present feelings of control as well as recall of their past ability to control their use.

Attributions for overuse of Instagram

Table 1 shows that self-reported Instagram addiction was more strongly associated with blaming oneself for using the app too often than was self-reported habit (Steiger’s Z = 2.52, p =.01). However, blaming overuse on the app’s design was only slightly (not significantly) more closely associated with self-reported addiction than habit.

Feelings about Instagram use

As shown in Table 1, self-reported Instagram addiction was more negatively associated with feelings about the app than self-reported habit (Steiger’s Z = −2.14, p =.03).

Discussion

Based on the BIAS symptom scale, only 2% of our quota-based sample of American adults who are active users of Instagram are at risk of being addicted. Nonetheless, on a self-report scale, 18% of participants agreed that they were at least somewhat addicted to Instagram. Even among the 46% of participants using Instagram for about an hour or more daily, only 5% were at risk of being addicted based on their clinical symptoms.

It is fascinating that so many Instagram users believe that they are addicted when, according to clinical criteria, the risk of addiction is relatively rare. Indicating that the overestimation of addiction is not limited to Instagram, we obtained even more extreme results in a convenience sample of college student TikTok users, with only 9% at risk of addiction compared with 59% estimating that they were at risk (see Supplemental Figure S2). These TikTok results are comparable to the estimates in a recent meta-analysis (i.e., 4–12% addicted) that was also largely based on convenience samples of college students⁴. The variation in results highlights the importance of using more representative samples to better estimate prevalence.

In contrast with addiction, most participants in Study 1 reported habits of using Instagram, with 50% of participants who used Instagram an hour or more daily indicating they were habitual, automatic users (i.e., at least somewhat agreed). This result is consistent with research showing that frequent use is largely habitual, reflecting mental associations between cues and responses that automatically drive accessing, scrolling, and responding on social media^20,21,23,24.

Symptoms of addiction

Two clinical symptoms proved to be the strongest signals of social media addiction: withdrawal or feelings of restlessness and trouble if use is prohibited, and life conflict or negative impact on job or studies. Both were experienced at least sometimes by all 9 of our participants at risk of addiction. These symptoms also differentiated addiction from habit; they were more strongly correlated with self-reported addiction than with habit and proved to be unrelated to habit after controlling for addiction.

It is interesting to speculate why people seem to believe that they are addicted to Instagram despite minimal evidence of pathology. Given that Facebook users who were more frequently exposed to posts about social media addiction rated themselves higher on pathological symptoms such as life conflict and loss of control¹¹, one possible explanation is that addiction is a common trope in popular media. Thus, we examined how frequently the term “social media addiction” appeared in news articles and the engagement these articles received from social media users. We suspected that, following the U.S. Surgeon General’s claims of social media addiction, the news media may have encouraged people to believe that the term addiction is the normative descriptor for heavy social media use. If so, users might adopt it even with limited evidence of pathology.

To assess the popular use of addiction and habit to describe social media, we used Buzzsumo, a media engagement tracking tool, to estimate the number of articles published across all U.S.-based media outlets that contained the phrases “social media addiction” or “social media habit” from November 2021 to November 2024. This tool also tracks the amount of social media engagement on Facebook, Twitter, Reddit, and Pinterest attached to each identified article.

Addiction proved to be a far more common descriptor than habit, based on the raw number of articles published and the amount of engagement those articles received online (see Table 2). In addition, over the 36 assessment months, the number of articles discussing “social media habits” never approached the number of articles including the term “social media addiction” (see Fig. 2). The stories driving these effects were often lawsuits. For example, the May 2022 and October 2024 peaks for “social media addiction” related to news reporting on multiple lawsuits against Meta (owners of Instagram). In addition, the May 2023 Surgeon General’s warning about social media addiction seems to have contributed to the steady drumbeat of new articles during the April-June 2023 period for “social media addiction.” For “social media habit,” the slight increases come from articles related to tips and tricks for altering one’s social media habits.

Table 2.

Number of media mentions of addiction vs. habit.

Search term	Social media addiction	Social media habit
Total articles	4,383	50
Total social media engagement	71,981	464
Total links to other web domains	1,759	4

Note. Engagements are the total number of reactions from other social media users to all published articles on a given topic across Facebook, Twitter, Reddit, and Pinterest (e.g., like, upvote, downvote, pin, reply, share, retweet, etc.). Data range from November 2021 to November 2024. Two search terms: “social media habit” and “social media addiction” were used. The search included only U.S.-based, English-language media outlets.

Fig. 2.

Results of Content Tracking Analysis from Buzzsumo. Note. Engagement metrics represent all interactions (likes, shares, replies) from Twitter/X, Facebook, Reddit, and Pinterest. Article quantities may also include YouTube videos.

As the engagement numbers in Table 2 imply, these articles are also spread by social media users and accumulate engagement across multiple social media platforms, suggesting that addiction is part of more social media discourse than is habit. This discourse could expose more social media users to the idea of being addicted, as more such articles are disseminated (either by algorithmic feed pickups or by manual user re-sharing) to more users. Study 2’s manipulation was designed to examine what happens to users when they adopt this addiction narrative.

In sum, these data suggest that, in the present U.S. media, addiction is the normative term to describe heavy social media use. This norm may drive our participants’ tendency to over-attribute their own Instagram use to addiction.

Consequences of self-reported addiction

In Study 1, the labeling of frequent Instagram use as an addiction or habit was associated with deficits in self-efficacy. That is, greater self-reported addiction to Instagram was tied to lower perceived control, more past control failures, less success cutting down on use, and greater self-blame for excessive use, suggesting that addiction feelings are connected with users’ sense of control. Notably, habit was not associated with these adverse outcomes in models that controlled for addiction, indicating that reduced control is specific to feelings of addiction.

Although Study 1 provides initial evidence of the negative effects of labeling social media use as an addiction, its correlational design cannot reveal whether perceived addiction is responsible for and, therefore, drives these negative experiences. Past literature has also been primarily correlational, revealing positive associations between exposure to Facebook posts about addiction and negative outcomes of usage¹¹.

Study 2: framing Instagram use as addiction can drive negative user outcomes

To test the idea that a belief in addiction yields negative consequences, Study 2 experimentally manipulated participants’ perceptions that they were addicted to Instagram. Specifically, we drew on language from the Surgeon General’s 2023 report on social media to replicate the experience of seeing authoritative sources frame social media use as an addiction. Finally, we assessed the consequences of this belief (see preregistration with AsPredicted #192351, https://aspredicted.org/t5bf-vcn5.pdf ).

We predicted that participants who first reflected on their own addictive use would experience negative downstream consequences, especially reduced feelings of control over their Instagram use. This impact should be evident across multiple aspects of perceived control, including participants’ past, present, and future control over app use. In addition, we expected that participants who had just self-reflected on their own addiction would blame themselves more for overusing Instagram and would like using Instagram less (less beneficial, less productive, less positive). Accordingly, they should want to use Instagram less than they currently do.

Method

All protocols, measures, and procedures were approved by the IRB at the University of Southern California (Protocol ID: UP-23–00585). Data collection took place in 2024. All experiments were performed in accordance with the rules of these approved protocols and relevant guidelines and regulations. Informed consent was obtained from all participants involved in this research, and anyone who did not give informed consent was not allowed to proceed further in the study beyond the consent form. Materials, data and code are available via OSF https://osf.io/3tcqm/overview?view_only=6f3c94fcedd0475ca08638993440b20a.

Participants

The participants were 824 US-based, English-speaking, adult daily Instagram users recruited from Prolific (N = 398 in the addiction reflection condition and N = 426 in the control condition). Prolific inclusion criteria included a 90% prior approval rate, US-based, English-speaking, adult, active Instagram account holders. Participants were compensated at a rate corresponding to California minimum wage, $16.00/hr, at the time of data collection. See the supplemental materials for justification of the sample size.

As preregistered (AsPredicted #192351), an additional set of participants were excluded who (a) gave incomplete, incomprehensible, abbreviated, or off-topic answers when self-reflecting about addiction (N = 37) or (b) were not willing or able to reflect on being addicted to Instagram (N = 133 in the addiction reflection condition, N = 112 in the comparison condition with the reflection task at the end of the study). Finally, two participants were eliminated for indicating they use Instagram 100 h per week (the scale max).

Procedure

All participants first responded to the BIAS symptom and SRBAI habit automaticity scales. Then they were randomly assigned to one of two conditions: the self-reflection condition, in which they responded to the addiction self-reflection task before completing the self-reported habit and addiction measures, followed by perceived control and self-blame measures, or the comparison condition, in which they completed the self-reported habit and addiction measures, perceived control and self-blame measures before the addiction self-reflection task.

Below is the full text of the addiction self-reflection task:

"You have just completed a series of questions that identify different types of Instagram users. Your responses indicate that you use Instagram often. The U.S. Surgeon General has warned that frequent, excessive social media use is addictive. It can be harmful to your mental health and well-being.

Please write 2–3 sentences below about times you felt addicted to Instagram. If you can’t think of any examples, then write how you imagine this experience would be for you."

Measures

Self-Reported Habit, Self-Reported Addiction, BIAS Symptom Scale³³, and SRBAI habit automaticity scale³⁸. These measures were described in Study 1.

Control over behavior

Participants reported their present ability to curb use on scales ranging from 1 (never true) to 7 (always true): “When I want to, I can curb my Instagram use and not think about or open the app”; “Right now, I feel like I have control over how often I use Instagram;” and “Right now, I feel like I am…” on a scale from 1 (Unsuccessful at reducing my Instagram use, even when I want to) to 7 (Very successful at reducing my Instagram use when I want to). Finally, they indicated, “I predict that, in the future, I will be able to reduce my Instagram use if I want to” on a scale from 1 (strongly disagree) to 7 (strongly agree). These four measures were combined into a single control scale (alpha = 0.84) that ranged from 4 to 28.

Number of past control attempts

To assess any attempts to reduce use, participants indicated, “How many times have you tried to quit or control your Instagram use in the past few months without success?” (sliding scale 0 to 100). Responses ranged from 0 to 100. Windsorizing the extreme outliers to 1.5 SDs above the mean yielded comparable results to those reported in the text.

Blame for overuse

On 7-point scales ranging from 1 (strongly disagree) to 7 (strongly agree), participants rated: “When I use Instagram too often, it is because…” “…of my own difficulties controlling my use” and “…the Instagram app is designed to keep people using it.” Obtained responses ranged from 1 to 7.

Positive/Negative feelings about Instagram

For this 3-item measure, participants indicated “My Instagram use is,” on a 5-point scale ranging from 1 (Harmful/Bad/Unproductive) to 5 (Beneficial/Good/Productive). Cronbach’s alpha was 0.84. Responses ranged from 3 to 15.

Desire to reduce Instagram use

Measured as: “Right now, I wish I used Instagram less frequently than I currently do” on a scale from 1 (Strongly Disagree) to 7 (Strongly Agree). Responses ranged from 1 to 7.

Results

Manipulation check

Demonstrating the effectiveness of the manipulation, participants who completed the self-reflection task indicated greater self-reported addiction in the addiction self-reflection first condition (M = 4.15, SD = 1.72) than in the comparison condition (M = 3.42, SD = 1.83), t(821.96) = 5.94, p <.001, 95% CI [0.49, 0.97], d = 0.41. The BIAS symptom scale, measured before the manipulation in both conditions, showed no significant difference between our participant groups (see supplemental results).

Negative experiences and Addiction-Self reflection

Number of past control attempts

Self-reflection participants reported significantly higher estimates of their prior attempts to control use (M = 6.33, SD = 12.41) than the comparison condition (M = 4.41, SD = 9.45); t(740.41) = 2.49, p =.01, 95% CI [0.41, 3.44], d = 0.18.

Feelings of control

Participants in the self-reflection condition reported significantly lower feelings of control over their Instagram use (M = 20.30, SD = 4.31) than the comparison condition (M = 20.90, SD = 4.35), t(819.19) = −1.98, p =.04, 95% CI [−1.19, −0.01], d = −0.14.

Blame for overuse

Participants blamed themselves more for overuse of Instagram in the addiction self-reflection condition (M = 4.18, SD = 1.67) than in the comparison condition (M = 3.73, SD = 1.88); t(820.78) = 3.63, p <.001, 95% CI [0.21, 0.69], d = 0.25. Participants in the addiction self-reflection condition were also more likely to blame the app for overuse (M = 5.66, SD = 1.26) than those in the control condition (M = 5.30, SD = 1.53); t(810.13) = 3.70, p <.001, 95% CI [0.17, 0.55], d = 0.26. The manipulation had comparable effects across self- and app-blame.

Positive/Negative feelings about Instagram use

Participants felt their use was slightly less beneficial, productive, and good in the addiction self-reflection condition (M = 8.69, SD = 2.52) than in the comparison condition (M = 8.95, SD = 2.28), although this difference was not significant (p =.12).

Desire to reduce Instagram use

Participants reported significantly higher intent to reduce Instagram use in the addiction self-reflection condition (M = 4.40, SD = 1.60) than in the comparison condition (M = 4.18, SD = 1.68), t(821.72) = 2.00, p =.046, 95% CI [0.01, 0.45], d = 0.14.

Discussion

The reminder of the U.S. Surgeon General’s report labeling social media use as addictive with negative effects increased daily Instagram users’ self-labeling of addiction, which in turn resulted in the experience of lower self-efficacy. Our results provide clear causal evidence that an addiction framing negatively impacts users’ perceived control over their Instagram use, impairing perceptions of their ability to manage past use and their capacity to control or reduce it currently and in the future. Additionally, the addiction self-reflection task increased users’ tendency to blame overuse on themselves as well as the Instagram app. This increase in self-blame is surprising given that the Instagram app is specifically designed to promote regular, habitual use²⁰.

It is impressive that even the two-minute exposure to addiction framing in our research was sufficient to produce a statistically significant negative impact on users. This effect is aligned with past literature showing that merely seeing addiction scales can negatively impact feelings of well-being¹⁹. Presumably, continued exposure to the broader media narrative around social media addiction has even larger and more profound effects. In conclusion, the addiction label does not empower users to regain control over their use. Instead, it hinders users by reducing feelings of control, increasing self-blame, and making the experience slightly less positive.

General discussion

In two studies, one correlational and the other experimental, we showed that labeling the use of Instagram as an addiction creates adverse outcomes, including greater perceived difficulty controlling engagement with the app and self-blame for overuse. Along with these largely overlooked harmful consequences of the addiction label, our findings raise questions about the broader construct validity of addiction as applied to problematic behaviors. Given that frequent social media use, and we suspect many other so-called behavioral addictions, do not typically involve the craving, lifestyle disruption, and withdrawal symptoms characteristic of substance abuse (see^41,42, the habit mechanisms that maintain excessive, problematic repeated use may be quite different from those that propel addictions.

Nonetheless, around 2% of Instagram users proved to be at risk of addiction based on their self-reports of psychophysiological symptoms. Given that more than 121.4¹ million Americans use Instagram, this is not an insubstantial number. Even these users, however, are not necessarily addicted by clinical diagnostic standards. Instead, their self-reports placed them in the at-risk category on a standard scale. Given the impact of reflecting on the addiction label in Study 2, exposure to popular news media mentions in daily life might have inflated our participants’ reports of addictive symptoms on the BIAS symptom scale. If so, then the actual number of social media addicts is likely much lower than our 2% estimate.

Some scholars publicly advocate treating social media use like a substance addiction (e.g., ^45,46). Although pathologizing social media use may help spur public support for regulatory policies, our findings suggest caution in applying addiction language too broadly. Broad application of the behavioral addiction framework to social media use has blurred the conceptual boundary between habits and addictions in research, misclassifying routine technology use as inherently harmful (see^28,32,47). Even clinical researchers have questioned the utility of this framework for social media, including whether two of the six Bergen scale components—salience and tolerance (referred to as urge in our study)—meaningfully distinguish pathological from non-pathological behavior outside the context of substance use³⁴. Misclassifying frequent social media and technology use as addictive has muddled public understanding of the psychology behind these behaviors and likely inhibits users’ understanding of the ways to effectively control their own behavior.

To avoid pathologizing heavy use that may not meet clinical addiction criteria, some researchers prefer the term problematic social media use (e.g.,^11,48–51). Although this circumvents the negative outcomes associated with the addiction label, problematic use describes the behavior without identifying a specific psychological mechanism underlying it. Understanding habit and other relevant psychological forces is central to controlling unwanted social media use (see^20,27).

Implications for behavior change

For the small minority whose excessive use of social media reflects an addictive response, treatment can require addressing the withdrawal symptoms and cravings that emerge when curbing use. Yet these may not be the best strategies to treat even excessive social media use that appears to be an addiction. Problematic social media use is a behavior that could be incorporated into a variety of psychosocial disorders, and these may require therapeutic approaches that differ from addiction treatment⁴².

For the majority of social media users, however, curbing excessive use involves primarily controlling habits. Like any other habit, social media habits can become misaligned with the original motivations for use (e.g., to obtain social rewards²¹), or conflict with other goals (e.g., sharing true information²³). Strong habits are notoriously difficult to control with willpower alone³⁰. For habitual social media users, the narrative of addiction and willpower-based attempts to control behavior could profitably be replaced with habit change strategies to realign their social media use with their current preferences²⁷.

Given that cues in the performance context activate habits, unwanted habits can be altered by shifting or removing cue triggers, potentially allowing more goal-directed behavior³⁰. Removing triggers might be as simple as changing social media settings to stop notifications, placing one’s phone out of sight, or using greyscale mode to create design friction (e.g., ⁵²). Unwanted habits may also be changed by practicing substitute activities, as in classic habit reversal training (see⁵³). By practicing competing activities, such as reading a book when bored or using alternative apps (e.g., language learning), people can develop new habits that better meet their current goals. In sum, changing addictions and habits requires different strategies, and inappropriate use of these terms not only leads to ineffective treatment strategies but also threatens self-efficacy to change, as we showed in the current research.

Conclusion

Our findings offer users good news: the large majority are not addicted to social media, even if they feel that way. Framing Instagram use in more realistic terms than addiction may improve self-efficacy and reduce self-blame. In addition, with a correct understanding that much excessive use is habitual, users have a path to curb use effectively.

Understanding the mechanisms of social media habit formation is essential for researchers to develop interventions that support healthy use patterns. This understanding enables testing and building actionable ways for users to adjust their habits in line with their goals and for social media sites to be redesigned to promote more socially beneficial outcomes.

While future research and replication is needed to fully generalize our work beyond the U.S. context, U.S. policymakers and media should also consider using the term, addiction, selectively for the few cases in which social media use has pathological symptoms. Continued overuse of this terminology risks stigmatizing healthy usage, wasting public health resources, and contributing to moral panic around social media use ^54,55. Furthermore, for most users, policies aimed at curbing or reforming social media platforms should address habits, not clinical approaches.

As understanding of social media use evolves, we encourage a balanced perspective–one that, on the one hand, doesn’t trivialize the real harms that social media poses (e.g., misinformation, individual bullying) or that, on the other, overextends psychiatric labels. By moving beyond stigmatizing terms for everyday behavior and fostering open, rigorous research based on a clear understanding of the habits that develop with repeated use, scientific research can better support users and identify maladaptive habits generated by social media’s design. As noted by UNICEF, “applying clinical concepts to children’s everyday behavior does not help support them in developing healthy screen time habits.”⁵⁶, p. 115). We encourage future researchers and clinicians to focus on solutions that do.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Dr. Ian Anderson co-conceptualized the study ideas and designs, collected and analyzed the data, and co-wrote the manuscript. Dr. Wendy Wood equally contributed to conceptualizing the study ideas and designs but did not participate in data analysis and/or collection. Dr. Wendy Wood author also co-wrote the manuscript.

Data availability

The anonymized data and analysis scripts for this study are available via this OSF link https://osf.io/3tcqm/overview?view_only=6f3c94fcedd0475ca08638993440b20a. All identifiable information has been removed to ensure participant anonymity, and only the cleaned dataset will be made publicly available. Researchers seeking access to the entire raw dataset can contact the corresponding author with an appropriate data-sharing request. The raw datasets and the corresponding data cleaning code are stored privately with the lead researcher and will be made available promptly for interested parties when appropriate.

Declarations

Competing interests

The authors declare no competing interests.