Does comparing yourself to others make you unhappy? The best current answer is that looking up to people doing better is the comparison most consistently linked to lower wellbeing. The evidence is mostly correlational, and the newest headline numbers come from a preprint that has not been confirmed as peer-reviewed.
What the new 2026 studies found
Three 2026 publications have renewed the debate. The first is a systematic review and meta-analysis posted to medRxiv as a preprint, with a Zenodo mirror. It reports that a tendency to compare upward correlated with poorer mental health, and that experimental upward comparisons adversely affected mental health. It found no evidence that downward comparisons were linked to mental health. The Zenodo record gives 15,404 participants across 77 experimental studies, but that figure has not yet been checked against the medRxiv text.
The second is a Frontiers in Psychology meta-analysis pooling 54 independent samples and 36,583 participants. It links higher upward comparison on social media with lower self-esteem and wellbeing. Its authors describe these associations as correlational, not causal. The third is a 109-nation study in Nature, covered below.
Several syntheses landed close together, and social media is where many people meet comparison targets every day. That explanation is my reading, not something these papers tested.
Upward comparison: the best-supported link
Looking up is the comparison type with the most consistent link to lower wellbeing. A 2023 meta-analysis of experiments in Cyberpsychology, Behavior, and Social Networking found that exposure to upward comparison targets on social media had an overall negative effect on self-evaluations and emotions. The subjective wellbeing effect was g = −0.19, according to the Taylor & Francis abstract; I did not review the full text. Downward comparisons did not show the same pattern, and the effects were small and short-term.
Correlation versus experiments
Most of the larger evidence is correlational. People who compare upward tend to report lower wellbeing, but that pattern does not show the comparison produced the gap. Experiments are stronger on direction, but they test brief exposures, not years of scrolling.
Trait and state comparison
Comparing is not one thing. Habitually measuring yourself against others, a trait, differs from a single comparison in a single moment, a state. The Frontiers meta-analysis draws this line. It matters because a momentary scroll habit and a long-standing tendency call for different responses.
Why the wellbeing numbers disagree
The reported correlations do not match. A ResearchGate record of a meta-analysis on online upward comparison and psychological maladjustment gives r = −0.31 across 15 studies. The PMC version of the 2026 Frontiers meta-analysis reports an average wellbeing correlation of r = 0.268 after reverse-coding positive outcomes. These may reflect different subsets or analyses, so avoid quoting one figure as settled.
Uncertainty runs deeper in one place. One prediction interval in the medRxiv preprint includes zero, which means a future study could find an effect of nothing. Headline direction is not the whole story.
Downward comparison: contested ground
Whether looking down helps is unsettled. Older work argues downward comparison can lift mood and self-concept. A volunteering study in Social Indicators Research found life satisfaction gains after helping a worse-off target, but only among people high in social-comparison orientation. Against this, the 2026 preprint reports no detectable mental-health link. A 2025 review argues downward comparison can also induce fear, but it appears in a lower-profile journal and should be treated as secondary.
Income and wealth comparison
Comparing money has the strongest cross-national evidence. The Nature study of 109 nations found income rank more strongly and consistently associated with subjective wellbeing than absolute income. Absolute income was often not significant once rank was included. It is an association across countries, not a test of cause.
A 2014 US county study of more than 1.7 million people found that higher income inequality was associated with stronger relative-income effects, and that this moderation was larger for low-income individuals. It is cross-sectional.
A 2025 Frontiers paper reports that wealth comparisons were most harmful with friends, then family, then online groups, and links the effect to stress. It is cross-sectional and rests on one paper, so treat the ordering as a lead.
Does social media make it worse?
Social media as a whole shows small and mixed associations with wellbeing. Upward comparison on social media is more consistently linked to lower self-esteem and wellbeing. The 2026 Frontiers paper argues that broad usage measures blur these results, because they combine different kinds of engagement. Many meta-analyses find only a small negative average effect of social media on adolescent wellbeing, and its practical significance is disputed.
A youth study found that the link between daily social media use and wellbeing was mediated by upward comparisons. Mediation analyses do not establish causal direction.
Not all online comparison looks harmful. A 2026 ecological momentary study in Cyberpsychology found that extrinsic comparisons, such as wealth, were not associated with poorer wellbeing in the contemporaneous network. It also found that online extrinsic comparisons were linked to more positive affect and less loneliness over time. It is a single study with an unusual design and a small sample.
Why some people are hurt more, and what the evidence does not show
Individual differences are the least settled area. Secondary sources point to social-comparison orientation, neuroticism, target similarity and culture as moderators. A 2025 review describes a dual pathway in which upward comparison usually lowers satisfaction through contrast but can sometimes raise it through inspiration. Confidence in these specifics is low.
On reducing harm, the evidence is thin. An Indonesian path analysis found social comparison did not directly affect life satisfaction or happiness, but affected them through subjective financial wellbeing. It is cross-sectional, so it is a lead rather than a tested remedy. The sources here do not rigorously test interventions.
What the evidence supports acting on is modest. If a particular feed or group reliably leaves you feeling behind, that fits the upward-comparison findings. Reducing that exposure is a reasonable personal experiment, but judge it by how you feel over several weeks, not by assuming a long-term result.