Research
Research on human values
What the Values Institute has learned from its own data, how the work is done, the standards every claim is held to, and what comes next.
Published
From our own data
Report · first edition
Global Values Report 2026: The Values-Living Gap
A survey of 9,656 people across 114 countries measuring the distance between what people value and their ability to live those values. Family, Respect, Kindness, Empathy and Honesty lead worldwide; a quarter of people who are clear on their values are too stressed to live them.
Read the reportBenchmarks
Which values people choose most
The share of people who placed each of the 48 assessment values in their top five, from 12,420 completed assessments. Family leads at 39%. Browse it as cards, or sort the full list by most chosen.
Open the values cardsModel
The eight values archetypes
The patterns that top-five values form together, drawn from the assessment's scoring model. Each archetype page shows its characteristic values and nearest neighbors. Prevalence figures will follow once the new data has a sample size worth stating.
See the archetypesReference
149 core values, with meanings
One master list: the 48 assessment values plus 101 more, each with a one-line meaning and classified intrinsic or extrinsic, means or end. 127 are intrinsic, 9 extrinsic, 13 both.
Open the listHow the work is done
Ranked, not rated
The assessment presents 48 values as statements and ranks them through forced choices in pairs. We have used pairwise comparison since the early 2000s because it produces a real hierarchy rather than a list of everything a person admires. The 48 values sit in four dimensions, Belonging, Growth, Impact and Fulfillment, from the research behind Start With Values.
Every statistic we publish states its sample size and date range. Groups too small to protect anonymity are combined or left out. We never publish the model’s internal weightings, and we never compare results across instruments as if they were one series.
Research uses de-identified responses only: no names, email addresses, account details or written answers. Anyone can opt out from their account. The full rules are in section 4 of the privacy policy.
How we label what we publish
Four kinds of claim, labeled as such.
Credibility comes from saying which is which. Every article carries one of these labels.
Original research
Findings from our own data: the Global Values Report, the assessment benchmarks and the studies that follow. Sample sizes and dates are stated.
Research-informed
Guides and articles built on peer-reviewed work by others, cited so you can check it.
Perspective
Interpretation and commentary. Our reading of the evidence, offered as a point of view.
Exploratory
Hypotheses and early thinking, including our work on values and artificial intelligence. Interesting, unproven, and labeled that way.
What comes next
Three lines of work
- A new edition of the Global Values Report, built on the rebuilt assessment's data and reported by age group and region, with sample sizes stated.
- Archetype prevalence: how common each of the eight patterns is, published once the sample supports it.
- Values in the age of intelligence, moving from editorial inquiry toward research questions we can test.
Research with us
We welcome academic collaboration. If you are a researcher with a question our data could help answer, write to us. De-identified data is shared under the research rules in our privacy policy, and we publish what we find whether or not it flatters our assumptions.
hello@values.instituteJournalists: the media kit has the facts, figures and bios in one place.
Put the research to work.
The free Values Assessment reveals your top five in about three minutes. Every completed assessment, de-identified, also makes the next report better.