May 8, 2026·Journal of Personality and Social Psychology, 1995
The Six-Factor Model Replicated on 1,108 Americans (Ryff & Keyes, 1995)
Read the paperA national US sample of 1,108 adults confirmed Ryff's six dimensions of well-being held together, and gave us the 18-item short form that became the workhorse of eudaimonic research.
The Question
Does Ryff's six-dimension model of psychological well-being actually replicate at national scale, and can it be condensed into a short form usable in large surveys?
What They Found
- Confirmatory factor analysis on the national sample showed the six-factor model fit the data better than a single well-being factor or any reduced-factor alternative
- The team produced a compact 18-item short form (3 items per dimension), the version now used in the Ryff PWB and most modern research
- Subscale means clustered between 4.55 and 4.83 out of 6, with a composite around 4.65; these are still the reference norms used today
- Personal growth and self-acceptance varied most by age and education, while purpose in life declined gently in older respondents, painting a more nuanced picture than a single happiness number ever could
- Eudaimonic well-being correlated only moderately with life satisfaction, reinforcing Ryff's claim that feeling good and living well are different things
How They Tested It
- National telephone sample of 1,108 American adults drawn for the MIDUS study, spanning ages 25 to 74
- Tested the original six-factor structure against several reduced models using confirmatory factor analysis on the full battery
- Selected the strongest 3 items per dimension based on factor loadings to build the 18-item short form, then re-ran the model on the short version
- Compared subscale scores across age, gender, and education groups to examine how the six dimensions shift across the lifespan
Caveats
- The fit indices for the strict six-factor model were acceptable but not ideal; several later samples have argued that autonomy and environmental mastery share more variance than the model assumes
- The 18-item short form trades range for brevity: with only three items per subscale, individual dimensions are more sensitive to a single odd answer than the original 20-item versions
- The MIDUS sample is American and predominantly white; cross-cultural adaptations of the structure came later and sometimes need adjusted thresholds
How Reborn Helps
Reborn uses this exact 18-item short form and the MIDUS subscale norms so you can see where you sit against the same US reference population the paper established.

Reviewed by
Rodrigue Buisson
Rod has spent the last five years reading the well-being literature so most people don't have to. Reborn is the app he wished existed when he started, built around feeling positive emotions and grounded in peer-reviewed research.