People often talk about social connection as if its value were obvious. A new study suggests that the picture is more individual than that. When 292 adults repeatedly chose between hypothetical social and non-social activities, the group showed only a modest average preference for social experiences. Yet that average concealed much larger differences associated with perceived social support and self-reported autistic traits.
The peer-reviewed study, published in Scientific Reports on 28 September 2026, used a willingness-to-pay task to turn an abstract question about social motivation into a series of concrete decisions. Participants chose between activities such as spending time with other people and solitary alternatives, while each option carried a small monetary cost. On average, they selected the social option on 52% of trials and the non-social option on 48%. They also spent an average of $0.76 on social experiences compared with $0.69 on non-social experiences.
Those differences were statistically significant, but small. The more revealing result was not that people, as a group, slightly preferred social experiences. It was that social valuation varied systematically with the social resources and socio-affective characteristics people reported bringing to the decision.
Putting a price on hypothetical social connection
Researchers Melanie Ruiz and colleagues set out to quantify how people value social experiences relative to non-social ones. Social connection is widely linked to well-being, but knowing that relationships matter does not tell researchers how strongly a person values a specific opportunity to engage socially at the moment of choice.
Data collection ran from September 2020 to April 2022. The researchers initially collected behavioural data from 289 online participants and 33 people tested in the laboratory. Twenty-nine online participants were excluded because they failed to respond to more than 15% of task trials or gave the same response across all trials of the computerised tasks. One laboratory participant was excluded after failing attention checks. The final analysed sample contained 292 people. There were 165 women and 127 men, and the mean age was 25.73 years.
A preregistered power analysis indicated that at least 193 participants would be needed to detect a correlation of about r = 0.20 with 80% statistical power, so the final sample exceeded that target.
Participants first completed self-report measures covering social resources, social motivation, affective characteristics and psychological well-being. They then completed a 24-trial card-guessing task that provided a performance-based monetary endowment of up to $6. That endowment created the budget used in the main willingness-to-pay task.
The central task consisted of 50 trials. On each trial, participants saw one hypothetical social experience and one hypothetical non-social experience. Examples included going to brunch with friends versus reading a book alone. The two options were presented with costs ranging from $0.01 to $0.05, and participants had up to 10 seconds to choose. The side on which social options appeared was balanced across trials.
This design matters because it moves beyond simply asking whether someone likes socialising. Every trial required a relative choice between two experiences while attaching a small cost to that preference. The researchers could therefore examine both how often people chose socially and how much of their task endowment they were willing to allocate to social experiences.
The average social preference was real, but modest
Across the sample, participants chose social experiences on an average of 52% of trials. The 95% confidence interval for the social-choice proportion ran from 50.3% to 53.2%. A one-sample test against the 50% chance level produced t(291) = 2.023 and p = 0.04.
The monetary measure pointed in the same direction. Participants spent an average of $0.76 on social experiences and $0.69 on non-social experiences. The difference was statistically significant, t(291) = 2.623, p = 0.009, with an effect size of d = 0.275.
That effect size is important for interpreting the headline result. The study does not show an overwhelming or universal human preference for social activity. Instead, it identifies a small average tilt toward social experiences in this sample. Some participants preferred social options much more strongly, while others preferred solitary alternatives. Understanding that variation became the more informative part of the analysis.
Social support tracked how much people valued social options
Perceived social support was positively related to both measures of social valuation. People reporting stronger social support were more likely to choose social rather than non-social experiences, with a correlation of r = 0.273 and p < 0.001. Perceived support was also positively related to the amount spent on social experiences, r = 0.286, p < 0.001.
These are associations, not evidence that supportive relationships directly cause people to spend more on social experiences. Several explanations remain possible. People who already find social interactions rewarding may build stronger support networks. Supportive networks may make future interactions feel safer or more rewarding. Both processes may operate together, or other characteristics may influence both perceived support and social valuation.
Still, the finding gives behavioural substance to the idea that social resources and social motivation are connected. The relationship was visible not only in a questionnaire about attitudes but in repeated choices carrying a small monetary consequence within the experiment.
Autistic traits showed a larger negative association
Self-reported autistic traits were negatively associated with social valuation. Higher trait scores correlated with a lower likelihood of choosing social experiences, r = -0.415, p < 0.001, and with lower spending on social experiences, r = -0.385, p < 0.001.
These relationships were stronger in magnitude than those observed for perceived social support. They should nevertheless be interpreted carefully. The study measured variation in autistic traits rather than establishing that every autistic person values social connection less. Social preferences can also reflect sensory demands, prior experiences, the particular activities presented, comfort with specific forms of interaction and many other factors that a single task cannot isolate.
The result is therefore better understood as evidence of heterogeneity in social valuation than as a simple deficit account. A lower willingness to pay for the study’s predefined social options does not mean that a participant lacks relationships, social needs or meaningful forms of connection.
A broader pattern emerged from many overlapping traits
The researchers also faced a common problem in social psychology: many measures of social and emotional functioning overlap. Analysing each questionnaire independently can make it difficult to tell whether apparently separate relationships reflect a smaller number of underlying patterns.
They therefore used principal components analysis to reduce the collection of socio-affective measures into broader dimensions. In the final peer-reviewed report, the three principal components were interpreted as reflecting social dysfunction, empathy and social motives, and pandemic-related effects.
A mixed-effects logistic regression then tested whether these component scores were associated with trial-by-trial social choice while accounting for participant-level variation and including age, gender and reaction time as covariates. Higher scores on the first component were associated with a lower probability of selecting a social option, β = -0.278, z = -8.268, p < 0.0001. The second component showed a positive association, β = 0.096, z = 2.926, p = 0.00343, and the third also showed a positive association, β = 0.1039, z = 3.186, p = 0.00144.
The monetary analysis showed the same directional pattern. The first component was negatively associated with total spending on social experiences, β = -0.11, t = -7.963, p < 0.001. The second component was positively associated with spending, β = 0.042, t = 3.210, p = 0.0014, while the third was also positive, β = 0.04, t = 2.818, p = 0.005.
Taken together, the component analyses reinforce the central point. Social valuation was not simply a stable preference shared equally across participants. It varied with broader patterns of social and emotional functioning.
Why the pandemic period matters
The timing of the research deserves attention. Data were collected between September 2020 and April 2022, spanning periods in which COVID-19 altered social opportunities, restrictions and people’s experience of isolation. The researchers included pandemic-related measures, and pandemic effects emerged as one of the broader dimensions associated with social choice.
This makes the study unusually informative but also historically specific. Social opportunities during that period were not equivalent to ordinary pre-pandemic or later conditions. Restrictions, perceived infection risk, remote work and study, and changes in friendship routines may all have influenced how people imagined and valued the activities presented.
Rather than treating that context purely as noise, the analysis suggests it formed part of the structure of individual differences. Future studies could test whether the same pattern of components appears in samples recruited entirely outside the pandemic period.
The findings do not mean socialising is always better
The study should not be read as evidence that choosing social experiences is inherently healthier, more rational or more desirable. Solitude can be restorative, and people differ in how much interaction they want and in which social settings they find rewarding.
The experiment instead offers a way to quantify relative subjective value. By placing social and non-social activities side by side and requiring repeated choices, the researchers could measure how preferences varied across people and connect those differences to social resources and socio-affective functioning.
That distinction has practical implications for research on loneliness, social motivation and interventions designed to increase connection. If people differ substantially in how they value specific forms of social engagement, simply increasing the number of available social opportunities may not affect everyone equally. The perceived quality, safety and personal relevance of those opportunities may matter as much as their availability.
Important limitations remain
The willingness-to-pay choices involved hypothetical activities. A participant might choose brunch with friends because they enjoy brunch, or reject reading alone because they dislike reading, rather than because one option is social and the other solitary. The authors identify the lack of individually tailored activity choices as a limitation because preferences for the activities themselves cannot be completely separated from their social content.
The monetary stakes were also small, ranging from one to five cents per option within the task. This is useful experimentally because it creates repeated trade-offs, but it is not equivalent to deciding whether to spend substantial real-world money on a concert, dinner, trip or other social experience.
The sample was relatively young, with a mean age of 25.73 years, and combined online and laboratory participants. The results therefore should not be assumed to describe older populations or all cultural and socioeconomic groups. The study is also primarily correlational when linking social support, autistic traits and broader socio-affective dimensions to valuation. It cannot establish the direction of causation.
Finally, data collection overlapped substantially with the COVID-19 pandemic. That context is analytically interesting, but it also limits how confidently the exact size of the observed social preference can be generalised to other periods.
A small average difference can conceal large individual variation
The most useful message from the study may be the contrast between the group average and the individual-level associations. Across 292 participants, the average preference for social experiences was only slightly above chance, and the spending difference was seven cents. Yet perceived social support and autistic traits showed clear relationships with how people allocated value between social and non-social experiences.
That is a reminder that statements about humans being a social species operate at a very broad level. In actual decisions, the subjective value of connection depends on the person, their social environment and the characteristics they bring to the interaction. Quantifying that variation may help researchers understand not merely whether people seek connection, but why the same social opportunity can feel highly valuable to one person and comparatively unappealing to another.
Source Information
Study: Ruiz, M., Stasiak, J. E., Sokol-Hessner, P., Smith, D. V. & Fareri, D. S. “Social cognitive functioning relates to willingness to pay for social experiences.”
Journal: Scientific Reports
Published: 28 September 2026
DOI: 10.1038/s41598-026-69935-z
Study type: Behavioural willingness-to-pay experiment with self-report measures and correlational analyses
Sample: 292 adults
