The 2026 research changes the comparison more than the conclusion: it asks whether people describe themselves, a peer, or the typical person as different across social media and offline life. Across five studies, participants' Big Five profiles were more similar across contexts for themselves and peers than for the typical person. Average trait ratings also varied by target. This does not show that people have separate online and offline personalities. It shows why an apparent difference needs a clear subject, a specific behavior, and a matched context before it becomes a claim about a lasting tendency.
What did the 2026 study change about the comparison?
The central change is that “personality online versus offline” is not one comparison. It can mean how I see myself in each setting, how I see a particular peer in each setting, or what I assume a typical person is like in each setting. Cameron J. Bunker's five-study 2026 paper measured all three kinds of perception, rather than treating them as interchangeable. Across the five studies, 1,749 participants rated the Big Five traits in social-media and offline contexts. The full paper’s sample breakdown shows four student samples and one Prolific online-panel sample. That mix gives the findings evidence from both student respondents and an online panel, but it is not a representative cross-section of all social-media users; student samples are narrower, and online-panel participation may not reflect people who do not join such panels. The results should therefore be read as reported perceptions in these study samples, not a population-wide estimate. The Big Five are broad dimensions commonly called openness, conscientiousness, extraversion, agreeableness, and neuroticism; the last refers to a tendency toward negative emotional experience, not a diagnosis.
The paper reports two kinds of comparison. Profile agreement means how similar the pattern across the five traits looks between contexts. Mean-level comparison asks whether ratings tend to be higher or lower in one context. A person could see the same relative pattern in both places while rating every trait a little lower online. Similarity and level are therefore different questions, not competing results.
Profile agreement was strong for self-perceptions and perceptions of peers, and moderate for perceptions of the typical person. In plain terms, participants generally saw the broad shape of their own and peers' traits as carrying across contexts. Their image of a generic person was less consistent. This is a refinement to the question, not evidence that social media reveals a hidden self or produces a second personality.
The finding is about reported perceptions. Participants were asked to describe traits in broad social-media and offline settings; the abstract does not establish that researchers observed each person across specific platforms, conversations, work settings, or relationships. Nor does the summary identify a single cause for any difference. The sound conclusion is bounded: who is being described changes the pattern people report.
Why can self and typical-person descriptions point in different directions?
In the 2026 study, participants generally rated themselves and peers lower on social media than offline. Their ratings of the typical person went the other way, or showed a smaller difference, for openness, extraversion, and neuroticism. Agreeableness was rated lower online across all three targets. Conscientiousness did not show a conclusive pattern across studies. These are mean-level perceptions, not measurements of how much people actually talked, explored ideas, cooperated, or felt distress in each setting.
That target difference matters in ordinary reflection. Someone may think, “I write more openly than I speak,” while also assuming that social-media users in general are louder or more outgoing. The first statement is a self-description; the second is a belief about a group. One does not verify the other. A like count, a short post, or a lively group chat also shows only a slice of behavior shaped by the audience, platform, purpose, and available time.
Earlier work by Bunker and Kwan helps explain why continuity and difference can both be part of a person's account. In two studies with 1,741 participants, people described their social-media and offline selves as similar on average, but not identical. Their second study compared early adults in Generation Z with Baby Boomers in late adulthood; the older group reported greater perceived similarity. Associations between perceived similarity and well-being were small and varied by group, so they do not support a rule that feeling more or less like oneself online is inherently better.
The studies do not settle why people rate the typical person differently. One plausible interpretation is that generic judgments draw on a stereotype of visible online behavior, whereas self- and peer-ratings draw on more specific knowledge. That is an interpretation, not a mechanism tested by the reported results. The practical lesson is to name the target before interpreting the trait: “my own posting,” “my friend in a group chat,” or “people I imagine on social media” are distinct observations.
Sources: Self-Perceptions on Social Media vs. Offline Contrast With Those Perceived in Generalized but Not Close Others; Similarity between perceived selves on social media and offline and its relationship with psychological well-being in early and late adulthood
How can you check what your own difference means?
Start with one tendency and two comparable situations. For example, if the question is whether you are more outgoing online, note one recent instance of initiating a conversation in a social-media space and one in an offline setting. Record what you did, who was present, how familiar they were, and whether the setting invited conversation. Repeat the comparison over a short, ordinary period rather than treating one unusually quiet evening or busy thread as a stable pattern. Keep the notes descriptive: count actions or record a brief example instead of scoring yourself against an imagined ideal. This makes it easier to notice whether the same tendency recurs under similar demands.
Then keep three kinds of evidence apart. First is your own behavior: messages sent, questions asked, or time spent listening. Second is your interpretation of that behavior: “I felt more comfortable initiating online.” Third is someone else's impression or an estimate inferred from digital traces. These may inform one another, but they are not the same measurement. A 2024 systematic review and meta-analysis by Joanne Hinds and Adam Joinson found moderate convergence between self-reports and human personality impressions across 24,124 people, and moderate convergence between self-reports and computer predictions across 42 studies. The authors also reported that results varied with the digital data sources. An inferred profile can offer a partial signal; it cannot settle what a specific behavior means to you.
A useful note has four parts: the behavior, the setting, the audience, and the alternative explanation. “I replied quickly in the hobby forum but waited during the family discussion” is more informative than “I am two different people.” Perhaps the forum gave you time to compose a response; perhaps the family discussion involved a sensitive topic or a strong status difference. Those possibilities do not erase personality. They identify conditions under which a tendency may be easier or harder to observe.
The verdict is modest but useful: the 2026 paper supports a better question, not a new label. Ask whether a difference recurs in comparable settings and whether it describes your own actions rather than an image of typical users. The private Context Profile offers continuums for reflection; use it to organize questions about preferences, then compare those questions with concrete experiences. Treat it as a reflection aid, not a validated test or an explanation of why a particular interaction happened. To explore the wider pattern, compare the result with examples you have recorded and visit the personality profile library.
Sources: Similarity between perceived selves on social media and offline and its relationship with psychological well-being in early and late adulthood; Digital data and personality: A systematic review and meta-analysis of human perception and computer prediction
Questions readers ask
Did the 2026 study find that people have different online and offline personalities?
No. It measured people's perceptions of Big Five traits across broad contexts. Self- and peer-profile patterns were relatively similar across contexts, while perceptions of the typical person were less similar. It did not establish separate identities or directly observe behavior on every platform and offline setting.
Why did participants see the typical person differently from themselves?
The study found target-dependent ratings, but its reported results do not establish why. A stereotype about visible online behavior is one possible interpretation, not a tested explanation. Keep assumptions about typical users separate from observations of yourself or someone you know.
What is a useful way to compare my online and offline behavior?
Choose one observable behavior and compare it across two reasonably similar situations. Note the audience, familiarity, purpose, and time available, then repeat before drawing a trait conclusion. Separate what you did from what you infer about yourself or other users.
Sources and notes
- Self-Perceptions on Social Media vs. Offline Contrast With Those Perceived in Generalized but Not Close Others
The publisher abstract reports five studies, 1,749 participants, target-specific profile agreement and mean-level Big Five perceptions across social-media and offline contexts.
- Similarity between perceived selves on social media and offline and its relationship with psychological well-being in early and late adulthood
The two-study paper reports perceived similarity without sameness across contexts, age-group differences, and small associations with well-being.
- Digital data and personality: A systematic review and meta-analysis of human perception and computer prediction
The University of Bath record reports moderate convergence of digital-data-based human and computer inferences with self-reports, with data-source moderators.
Apply it to your own pattern
See how your everyday tendencies combine
From this guide: A broad finding cannot tell you whether your own difference reflects audience, setting, or a recurring preference across situations.
If you want to compare this pattern with other tendencies, the private Context Profile offers a set of continuums for reflection. It does not choose a career or predict how you will behave in a role. Use it to organize questions about the environments and demands you prefer, then compare those questions with concrete experiences from work and life.