1. The problem this paper addresses
A gym owner completes a personality assessment. It tells her she is high in conscientiousness and low in agreeableness. The report explains that she is organised, demanding of others and comfortable with confrontation. All of this may be accurate. None of it explains why she has cancelled the same difficult conversation with her head trainer four weeks running.
The instrument measured something real and reported it correctly. It also missed the thing she actually needed to know. Her conscientiousness has not changed since Tuesday. Her capacity to act on it may have.
That gap is the subject of this paper. The claim is not that trait measurement is invalid. It is among the more robust achievements of differential psychology and we have no quarrel with it. The claim is that trait measurement on its own cannot fully answer the question owner-operators bring to assessment, which is almost never “what am I like?” and much more often “why am I not doing the thing I know I should be doing?”
2. Theoretical grounding: the state–trait decomposition
The formal apparatus here is not new. Latent state–trait theory, developed by Steyer and colleagues from the late 1980s, generalises classical test theory to account for the fact that psychological measurement does not occur in a situational vacuum. LST separates stable person effects, situation-specific state effects and measurement error.
Its central definitional move is precise and worth stating. A score on a latent state variable is the expectation of an observed variable given a person in a situation. A score on a latent trait variable is the expectation given a person. The theory further defines consistency, occasion specificity, reliability and stability coefficients, allowing the relative contribution of person and situation to be estimated rather than assumed.
The revised formulation, LST-R, sharpens this by insisting on four facts any measurement theory must accommodate. Observations are fallible. They never happen in a situational vacuum. They are always made using a specific method of observation. And there is no person without a past. The second of those is the one commercial assessment routinely ignores.
The empirical significance of the decomposition is not marginal. A construct can be genuine, measurable and reliably assessed while still showing meaningful variation across occasions and situations. The size and meaning of that variation depend on the construct, the population, the interval between measurements and the method used. It cannot be inferred from a single universal threshold.
The practical point is simpler. A one-time score may contain both enduring and transient influences. Where those influences are not separated, the report can describe an observed pattern more confidently than the design permits.
This framework descends from the person–situation debate, in which Mischel and Shoda challenged purely trait-based models by proposing that behaviour varies systematically across situations through cognitive-affective processing. That debate was not resolved in favour of either pole. It was resolved in favour of interaction and decomposition, and commercial assessment has largely not noticed.
2.1 What follows for measurement practice
Unmeasured state influences do not disappear. In a single-occasion assessment, they may contribute to the observed score without being distinguishable from enduring person effects or transient measurement error.
The direction of that influence is not necessarily random. A respondent completing an assessment during a period of unusually high demand, low recovery or emotional strain may answer differently from the same respondent completing it under more stable conditions. A single administration cannot tell us how much of that difference belongs to disposition and how much belongs to the occasion.
Methodologically, this is not an unknown problem. LST-based frameworks for developing and evaluating personality-state measures are well specified. Items and scales can be evaluated for the proportion of reliable variance attributable to stable person effects and to occasion-specific influences. The methods are available. Commercial products rarely use them because a single-occasion, single-score product is easier to administer, explain and sell than a decomposition.
3. Why owner-operators are the acute case
The state component of behaviour is present in every population. Our narrower claim is that state-related variation may be unusually consequential in owner-operators, and may also account for a larger share of observed behavioural variation in this population. Four features of the role give us reason to test that hypothesis.
The role has less structural buffering. An employee whose capacity drops often operates inside a system that continues around them: a manager reassigns work, a process holds the decision, or a colleague covers. The owner-operator is often the final buffer. When their capacity drops, less of that drop is absorbed before it reaches the business.
Decision density is high and decision consequence is often broad. Owners make a large number of discretionary decisions across functions, frequently without the protection of a tightly bounded role. State variation that would be contained in a narrower position may therefore become materially expensive.
Recovery is often difficult to protect. Research on recovery experiences commonly distinguishes detachment, relaxation, mastery and control. The owner-operator role can interfere with all four. Detachment is difficult when the business is financially and identity-linked to the self, and control over one’s own time is frequently the first thing surrendered to the business.
The role also changes as the business grows. The evidence is not uniformly negative. Some research suggests solopreneurs may report lower burnout risk than entrepreneurs who expand and employ others. That matters because it points away from a simple personality explanation and toward an interaction between role structure, responsibility, support and load.
Put those features together and you have a population in which state-related variation may be amplified by role design, weakly buffered by organisational structure and directly connected to business outcomes. That is enough to justify a different measurement question. It is not yet enough to claim that owner-operators are universally more state-dominant than employed populations. That claim requires data.
4. What we are not claiming: the depletion caveat
It would be convenient to ground this argument in ego depletion, the strength model of self-control under which self-control is a limited resource that depletes with exertion. It would also be unwise, and we want to be explicit about why.
The depletion literature is contested to a degree that makes it unsuitable as a load-bearing foundation. The first meta-analysis reported a medium effect. Later analyses that incorporated unpublished studies and publication-bias corrections produced estimates close to zero. Large-scale registered replications also returned small or null effects, although defences and alternative interpretations remain.
Our position requires nothing from the strength model. We do not claim that self-control is a finite resource depleted through a specific physiological mechanism. We claim something narrower: observed behaviour contains both enduring and occasion-specific influences, and an assessment intended to explain current execution should not assume that the enduring component is the whole explanation.
That claim survives the total collapse of ego depletion as a theory because it never depended on it.
We state this prominently because the distinction between what a framework requires and what it merely gestures at is exactly the discipline this literature has learned the hard way.
5. Defining behavioural energy
The term energy is easy to understand and easy to misuse. In this paper it does not mean a hidden biological resource, a diagnosis, or a single general reserve that is spent in equal units across every task.
We use behavioural energy to mean a person’s current, self-reported capacity and willingness to initiate, sustain and recover from effort across specified categories of business activity.
That definition is deliberately operational. It does not assume a cause. A low score may reflect workload, fatigue, emotional strain, poor role fit, low motivation, insufficient skill, conflict, lack of support, health, or circumstances outside the business. The measure cannot determine which of those is responsible on its own.
It also leaves open an important possibility: energy patterns may contain both stable and fluctuating components. A person may show an enduring preference for some categories of work and a temporary reduction in capacity within the same category. The eventual empirical model may therefore be more complex than a clean trait-versus-state split.
6. The two-section architecture
The Behavioural Performance Index is our initial attempt to operationalise this argument. The instrument is divided into two sections, scored independently and reported separately.
The behavioural section measures tendencies across five domains. These are intended to represent relatively stable characteristics, although their actual stability must be demonstrated rather than assumed.
The energy section measures current, self-reported capacity and willingness across categories of work. These are intended to vary meaningfully across occasions and conditions.
The goal is not simply low test–retest stability. Low stability could reflect a valid state measure, but it could also reflect weak reliability or noisy items. The energy scales should instead demonstrate reliable within-person variation and change in theoretically predictable ways as relevant conditions change.
Four things may follow from the separation if the architecture performs as intended.
Interpretation may become clearer. A behavioural score and an energy score together may help distinguish between an enduring tendency and a current constraint. The same behavioural profile at high and low energy should produce materially different guidance.
State-related influences may become more visible. Measuring current-state variables alongside behavioural tendencies may reduce the risk of treating every observed pattern as disposition. Whether it materially reduces state contamination in behavioural scores is an empirical question.
Falsifiable predictions become available. Behavioural scales should show stronger stable-person components. Energy scales should show reliable occasion-specific variation and theoretically predictable change. The two sections should also remain empirically distinguishable rather than collapsing into a single general factor. If those predictions fail, the architecture is wrong.
Reassessment acquires a rationale. Frequent reassessment of a pure trait instrument adds little if the construct is genuinely stable. Reassessment of the energy section is central to its purpose because it is intended to track what changes.
7. Implications for commercial assessment generally
Three propositions follow that extend beyond our own instrument.
A single-occasion trait score, reported without reference to current conditions, may overstate what has been measured. This is not a criticism of any particular publisher. It is a structural limitation of single-occasion assessment that LST theory identified decades ago and commercial practice has only partly absorbed.
Instruments marketed to owner-operators may carry a heavier interpretive burden than instruments used in more bounded roles because state-related constraints can travel more directly into operational outcomes.
State and trait scales should not be judged by identical criteria. High test–retest stability may support the interpretation of a trait measure. It is not, by itself, the correct quality standard for a state measure. State measures need reliability within occasions, meaningful within-person variation and evidence that change occurs for reasons the theory predicts.
8. Limitations and what comes next
This is a theoretical paper. It advances an architectural argument grounded in established measurement theory and in a growing but still incomplete literature on owner-operator strain, recovery and burnout. It does not report data.
Several claims require empirical support. The first is that the BPI behavioural scales show substantial stable-person variance across occasions. The second is that the energy scales show reliable within-person variation and sensitivity to relevant change. The third is that the two sections demonstrate discriminant validity rather than collapsing into a single factor. The fourth is that joint interpretation predicts operational outcomes more usefully than trait scores alone.
The owner-operator population claim also needs direct testing. We have given structural reasons to expect state-related variation to be unusually consequential in this group. We have not shown that it is larger than in employed populations.
We note the obvious conflict of interest. This paper argues for an architecture we have built and sell. The appropriate response to that conflict is not a disclaimer but falsifiability, and we have tried to state the argument in a form that could be shown to be wrong.
References
Bader, M., Columbus, S., Zettler, I., & Mayer, A. (2025). Developing, evaluating, and interpreting personality state measures: A framework based on the revised latent state–trait theory. European Journal of Personality.
Carter, E. C., Kofler, L. M., Forster, D. E., & McCullough, M. E. (2015). A series of meta-analytic tests of the depletion effect. Journal of Experimental Psychology: General.
Dang, J., et al. (2021). A multilab replication of the ego depletion effect. Social Psychological and Personality Science.
Freudenberger, H. J. (1974). Staff burn-out. Journal of Social Issues, 30(1), 159–165.
Geiser, C., & Lockhart, G. (2012). A comparison of four approaches to account for method effects in latent state–trait analysis. Psychological Methods, 17, 255–283.
Hagger, M. S., & Chatzisarantis, N. L. D., et al. (2016). A multilab preregistered replication of the ego-depletion effect. Perspectives on Psychological Science, 11(4), 546–573.
Hagger, M. S., Wood, C., Stiff, C., & Chatzisarantis, N. L. D. (2010). Ego depletion and the strength model of self-control: A meta-analysis. Psychological Bulletin, 136(4), 495–525.
Inzlicht, M., & Friese, M. (2019). The past, present, and future of ego depletion. Social Psychology, 50(5–6).
Maslach, C., & Leiter, M. P. (2016). Understanding the burnout experience. World Psychiatry, 15(2), 103–111.
Mischel, W., & Shoda, Y. (1995). A cognitive-affective system theory of personality. Psychological Review, 102(2), 246–268.
Pines, A., & Aronson, E. (1988). Career Burnout: Causes and Cures. Free Press.
Steyer, R., Mayer, A., Geiser, C., & Cole, D. A. (2015). A theory of states and traits—revised. Annual Review of Clinical Psychology, 11, 71–98.
Steyer, R., Schmitt, M., & Eid, M. (1999). Latent state–trait theory and research in personality and individual differences. European Journal of Personality, 13(5), 389–408.
Vohs, K. D., et al. (2021). A multisite preregistered paradigmatic test of the ego-depletion effect. Psychological Science, 32(10), 1566–1581.
Wright, J. C., & Mischel, W. (1987). A conditional approach to dispositional constructs. Journal of Personality and Social Psychology, 53(6), 1159–1177.
Institute of Behavioural Performance working papers are pre-publication documents circulated for comment. Correspondence: behaviouralperformance.org