Meta Spirituality · The Meta Self Scale
The paradigm that dissolves the ego and increases agency at the same time
Nina needed a questionnaire that did not exist, so she built one.
Spiritual traditions have always said the same thing about losing the ego. You lose your desires along with it. No ambition, no drive, no wanting anything, and very little interest in running an ordinary life. Nina Verkoeyen kept seeing the opposite in the people she worked with. Their ego-self lost its grip and they became more decisive, more capable and more in charge of their own lives at the same time.
She wanted to show that with numbers instead of stories. Some questionnaires measure how far the ego has dissolved. Others measure how much control a person feels over their life. Not one of them asked about both, so there was no way to show that the two arrive together. The Meta Self Scale was built to ask both questions in one place.
PREREGISTRATION
Convergent and discriminant validity
Related to the right things, and redundant with none of them
Four preregistered hypotheses tested whether the dimensions relate to established measures in the predicted direction. All four were confirmed under Bonferroni correction, with correlations large enough to share content and modest enough to prove the scale is not a restatement of an existing one.
Hypothesis | Tested against | r | Result |
|---|---|---|---|
H3 · Convergent | Dissolution × EDS Ego-Loss | .53 | Confirmed |
H4a · Convergent | Unity × EDS Unity | .66 | Confirmed |
H4b · Discriminant | Unity × existential loneliness | −.49 | Confirmed |
H5 · Convergent | Agency × Pearlin Mastery | .64 | Confirmed |
Not grandiosity | Narcissistic rivalry | −.32 | Higher scores, less rivalry |
The obvious worry about any expanded-identity construct is that it captures narcissism in spiritual language. Higher scores predicted consistently less narcissistic rivalry and were essentially unrelated to narcissistic admiration, which means the configuration arrives with more connection to other people rather than less.
The key finding
Transcendence and authorship rise together
This is the hypothesis the scale was built to test, and it is where the received wisdom is most likely to be wrong. Two preregistered regressions put it directly to the data, entering an external measure of ordinary self-attenuation alongside all three transcendence dimensions at once.
Ego loss on its own did not predict greater agency. The integrated configuration did, which is the finding the whole instrument existed to make possible.
Simultaneous regression with EDS Ego-Loss and the three MSS transcendence dimensions as predictors, both models significant at p < .001, all variance inflation factors at or below 2.5, and agency coefficients identical under HC3 robust standard errors. The ego-loss coefficient was small and not robustly distinguishable from zero.
What the numbers actually show
The ego did loosen. The authorship did grow. And it was not the ego’s kind of authorship
Three separate things had to be true for the claim to hold, and each one was tested against instruments other researchers built, so none of it rests on our own questions agreeing with themselves.
r = .53
The dissolution was real
r = .64
The agency was real
r = −.32
And it was not the ego growing back
Two qualities that were supposed to cancel each other out went up together, and what came up was not a bigger ego. It was a person who can act without needing to be somebody first.
The dissolution items line up with an established external measure of ego loss at r = .53, which means this is the same thing other researchers have been measuring for years and not a private definition. In the retreat trial those items moved further than anything else on the scale. Agreement with no longer identifying as a separate individual rose by 28%, and with experiencing the dissolution of the ego-self by 24%.
Agency tracks the Pearlin Mastery Scale at r = .64, the standard measure of how much control a person feels over their own life, so the authorship being described is the kind psychologists already know how to measure. In the retreat trial mastery itself rose with a controlled effect size of 1.28, which is among the largest movements recorded in the whole research program.
This is the part that had to be checked, because self-assertion can look like authorship from the outside. If what rose were the ego, higher scores would arrive with more rivalry and more grandiosity. They arrive with less. Narcissistic rivalry runs in the opposite direction at r = −.32, narcissistic admiration is unrelated, and the people scoring high on all four dimensions had the lowest rivalry of any group in the sample. In both regression models the strongest predictor of agency was unity, the sense of belonging to other people and to life, which means this authorship arrives through connection rather than through separation.
Spirituality got this backwards for two thousand years
Why it had to be built
Everyone says you have to choose one or the other
Ask almost anybody what happens when the ego dissolves and you get the same picture back. A person who wants nothing, decides nothing and withdraws from ordinary life. Traditions call it renunciation, popular spirituality repeats it, and psychology built its instruments on the same idea. The scales that measure self-transcendence ask how much of the ordinary self has loosened and stop there. Whether a person can still run their life is measured by other researchers, in a different field, with different questionnaires.
01
That is not what Nina watched happen. People who went through the shift started making decisions faster, arguing less, and taking on things they had been avoiding for years. Their ego-self had loosened its grip and their sense of being the author of their own life had gone up, both at once, which is exactly what this paradigm sets out to do.
If the two rise together, one questionnaire has to ask about both. Until this one, none did.
02
Showing that requires one questionnaire that asks about both, and there was not one. A scale that only measures dissolution cannot tell the difference between a person who has come together and a person who has come apart, because the two of them answer those questions the same way. So the Meta Self Scale puts authorship in the same instrument as dissolution, expansion and unity: four dimensions, sixteen questions, one score.
The construct
Four dimensions, sixteen items
Each dimension is measured by four statements rated from strongly disagree to strongly agree. No item is reverse-worded, and the scale is scored as a single mean of all sixteen.
Dissolution
How far the ordinary ego-self has stopped being the centre of who you are.
RELIABILITY
α = .78
Expansion
Whether your sense of identity reaches past the separate individual.
RELIABILITY
α = .84
Unity
How connected you feel to other people, to life, to reality as a whole.
RELIABILITY
α = .86
Agency
How far you experience yourself as the author of your own life.
RELIABILITY
α = .83
The first three describe what loosens. The fourth describes what holds. Measuring them together in one instrument is the entire point of the scale.
The validation study
1,652 people, and every decision lodged in advance
A single-session online survey. Hypotheses, exclusion criteria and the analysis plan were all published on the Open Science Framework before any data was collected, which means the exclusions below were rules rather than choices.
Reliability
Consistent at every level, and not by inflation
Cronbach’s α with 95% confidence intervals. McDonald’s ω tracked α exactly at every level, which rules out the possibility that these figures were propped up by near-duplicate items.
INTERNAL CONSISTENCY BY SCALE
Total score .92
[.917–.928]
Unity .86
[.845–.867]
Expansion .84
[.823–.849]
Agency .83
[.817–.844]
Dissolution .78
[.759–.794]
.923
.856
.836
.831
.777
CRONBACH’S Α · 95% CONFIDENCE INTERVALS
Factor structure
The four-dimension model beat every simpler one
Confirmatory factor analysis under robust maximum likelihood, tested against the four acceptability thresholds that were preregistered before the data existed.
Model | CFI | TLI | RMSEA | SRMR | Thresholds met |
|---|---|---|---|---|---|
One factor | .812 | .783 | .122 | .072 | 0 / 4 |
Two factors | .876 | .855 | .100 | .056 | 0 / 4 |
Three factors | .898 | .879 | .091 | .051 | 0 / 4 |
Four factors | .934 | .919 | .075 | .043 | 4 / 4 |
Preregistered thresholds were CFI and TLI at or above .90, RMSEA at or below .08, and SRMR at or below .08. Every simpler model was rejected against the four-factor model by chi-square difference test, all at p < .001, and standardized factor loadings ran from .64 to .85.
The central structural finding
One dimension, four faces
A bifactor analysis showed that all four dimensions are largely expressions of one dominant underlying trait, an integrated transcendent identity, rather than four independent constructs. That has a practical consequence: the total score is the score to use, and the four facets describe what kind of integrated identity somebody has instead of four separate things to add up.
VARIANCE STRUCTURE
General Meta Self factor
Agency-specific variance
Dissolution-specific
Expansion-specific
Unity-specific
70
15
8
4
3
.70
Share of common variance carried by the general factor (ECV)
.86
Proportion of total-score variance due to the general factor (ωH)
.37
Unique variance carried by Agency, the most distinct dimension (ωHS); Unity carried the least at .12
This is not a flaw in the instrument. It is evidence that the configuration is real.
Known-groups validity
Scores tracked familiarity with the paradigm, as they should
Mean total score rose steadily with reported familiarity, and because item loadings and intercepts were equal across familiarity and age groups under formal invariance testing, this gradient reflects genuine differences in the latent construct rather than insider language.
MEAN TOTAL SCORE BY FAMILIARITY
Slightly familiar | N = 333
Moderately familiar | N = 502
Very familiar | N = 474
Deeply engaged | N = 343
3.51
3.67
3.78
4.13
d ≈ 0.8–1.2
Latent difference between the least and most familiar groups under scalar invariance
50.5%
Reported only slight or moderate familiarity with the paradigm
3.51
Mean score in the least familiar group, still well above the scale midpoint
Familiarity was by far the strongest correlate (η² = .13) ahead of years of practice (.07) and years of engagement (.02), which indicates that subjective integration with the framework rather than sheer time spent in it is what the scale tracks.
Identity profiles
The integrated profile had the lowest rivalry of any group
Within-sample median splits identified four theory-derived configurations. The integrated profile, above the median on all four dimensions, was also the least rivalrous.
Profile | Share of sample | n | Narcissistic rivalry (1–6) |
|---|---|---|---|
Meta Self, all four elevated | 29.7% | 490 | 1.63 |
Unity without agency | 9.3% | 153 | 2.09 |
Dissolution without integration | 8.3% | 137 | 2.01 |
Ego inflation, expansion only | 5.1% | 85 | 2.1 |
Profiles are mutually exclusive by construction but not exhaustive, and 47.6% of participants met none of the four patterns. Latent profile analysis confirmed that people vary along a common elevation continuum rather than forming distinct types.
Incremental validity
It adds something the existing scales do not
A deliberately conservative test. Three established instruments, the Pearlin Mastery Scale, EDS Unity and EDS Ego-Loss, were entered first as predictors of existential loneliness, and the Meta Self Scale was added at a second step.
The test is conservative because the baseline already contained a felt-unity measure that overlaps with one of the scale’s own dimensions. That the scale still added unique prediction indicates it captures variance the existing instruments miss.
A scale that cannot move is not measuring a process.
The second preregistered study
Can it detect real change?
A cross-sectional survey can show that a scale is reliable and that its scores relate to the right things. What it structurally cannot show is whether the scale moves when identity actually shifts, which is the property psychometricians call responsiveness and the one that matters most for an instrument built to track a process. A second preregistered study answered it directly, using the scale as the primary outcome in a randomized controlled trial of the fourteen-day online retreat.
1,204
Participants, randomized 602 to the retreat and 602 to a waitlist control
+21%
Increase in the retreat arm, mean total 3.43 to 4.16
d = 1.42
Effect size in the retreat arm, the range clinical trials describe as very large
−2.5%
Change in the waitlist control over the same period
2 in 3
Participants who crossed the threshold for reliable change
¾ SD
Size of the shift relative to the spread of the validation sample
This adds what the first study could not. The scale registers genuine movement in identity and not only stable differences between people who were always different. Taken together, one study establishing the structure and one testing responsiveness are the standard two-step sequence for an instrument intended to track change over time, and both were preregistered before data collection began.
Method note.
Study one was a single-session online survey with hypotheses, exclusion criteria and the analysis plan preregistered on the Open Science Framework. Of 1,989 responses, 337 were removed under the preregistered rules, leaving 1,652 for analysis. Confirmatory factor analysis used robust maximum likelihood, and reliability is reported as Cronbach’s α with McDonald’s ω as a check. Study two used the scale as the primary outcome in a randomized controlled trial of the fourteen-day online retreat, with 1,204 participants randomized in equal arms and preregistered separately. The effect size quoted here is the within-arm change in the retreat group; the between-group estimate adjusted for baseline is reported on the trial page. Both studies are self-report, and participants were not blind to condition.
Source.
Verkoeyen 2026, the Meta Self Scale: development and validation, N = 1,652, with responsiveness tested in a randomized controlled trial of N = 1,204.
Availability.
Preregistration at doi.org/10.17605/OSF.IO/TSH6U, preprint at doi.org/10.23668/psycharchives.22220, and archived items and materials at doi.org/10.17605/OSF.IO/3RVXD.
A paradigm for ending suffering at the level of identity, tested in four randomized controlled trials.
Meta Spirituality Institute LLC · USA
Programs, books and facilitator certification.
Meta Spirituality Foundation Inc. · Nonprofit
Research, trials and public programs.
IMPORTANT NOTICE
Meta Spirituality programs, books and protocols are educational and are not medical care, psychotherapy or a substitute for either. They are not intended to diagnose, treat, cure or prevent any condition. If you are in crisis or under the care of a clinician, please speak with a qualified professional before making changes to your treatment.
© 2026 Meta Spirituality Institute LLC. Meta Spirituality® and the Meta Self® are registered trademarks. All rights reserved.
PROGRAMS
The Meta Self
The Process
Anxiety Protocol
Relationships Protocol
Meta Self Scale
RESEARCH & RESULTS
The trials are carried out by Meta Spirituality Foundation Inc., the nonprofit, and reported in full in the linked preprints. Meta Spirituality Institute LLC delivers the paid programs and the facilitator certification. Individual results vary, and outcomes reported for a group are not a promise of an outcome for any one person.
Terms of Service
Privacy Policy
Refund Policy
Disclaimer
ORGANIZATION
Research & trials
The Institute
Get certified
The Foundation
About Nina
CONNECT
Contact
FAQ
Telegram channel
Podcast
Press & media