Can Big Five Test Results Change? Retesting, Measurement Error, and Trait Change

7 days ago

20 min read

Author: Ukrainian Psychological Hub · Published: October 9, 2026 · Editorial Policy

Big Five results can change when you take a personality test again. A different score can reflect temporary influences on your answers, ordinary measurement imprecision, a change in the test or scoring norms, or a genuine change in your usual patterns of thinking, feeling, and behaving. The central question is therefore not whether two score reports match exactly. It is what their difference can legitimately tell you.

Research makes this distinction essential. A meta-analysis of short-interval Big Five retests found substantial, though imperfect, score dependability across instruments; its median aggregated estimate was .816, with variation by domain and measure (Gnambs, 2014). In longitudinal research, personality shows meaningful stability and genuine development over years, with different findings depending on age and the type of stability being examined (Bleidorn et al., 2022). Neither result establishes how much any one person's personality changed between two online test sessions.

This guide focuses on retaking Big Five assessments and interpreting score differences. For the broader model, see our Big Five personality traits overview; for what validated tests actually assess, see the Big Five personality test guide.

The short answer: why the same person can get different Big Five results

Imagine taking a conscientiousness questionnaire in January and scoring near the middle of its comparison group. In April, you score somewhat higher. You may have developed more consistent planning habits. You may also have answered from a different reference period, been less distracted, interpreted the wording differently, or used a changed assessment platform. A percentile may move even if the underlying raw score hardly changes, when the comparison sample or scoring model changes.

A second score is a new observation of a psychological construct, not an automatic measurement of the amount by which that construct has changed. Big Five traits are continuous dimensions. An increase in a questionnaire score does not turn someone into a new personality type, and neither a low nor a high result is a clinical diagnosis.

Four practical possibilities deserve separate consideration: the person's enduring trait level changed; their situation-dependent behavior or current feelings altered their self-report; measurement error contributed to the difference; or the test itself, its instructions, or its norms changed. More than one explanation can operate simultaneously.

What does a Big Five score actually measure?

The Big Five organizes personality variation into broad domains commonly called Openness to Experience, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. Particular instruments implement these domains differently. For example, the Big Five Inventory–2 (BFI-2) labels its fifth domain Negative Emotionality and measures 15 instrument-specific facets across the five domains (Soto & John, 2017). The domain naming and facet structure of one inventory should not be copied indiscriminately into another.

Most common inventories ask respondents to rate descriptions of typical behavior or experience. Scores summarize patterns in these responses. Their scientific usefulness depends on the specific measure, administration conditions, evidence of validity, and population for which an interpretation is intended. A self-report score estimates a trait using selected indicators; it is not direct observation of every behavior a person performs.

The distinction between broad domains, narrower facets, and situational expressions also matters for retesting. An overall conscientiousness score can stay similar while more specific tendencies concerning order or persistence change in different directions. Conversely, a narrow score can fluctuate substantially while the broader domain remains relatively stable. The precise facet names and their number depend on the assessment. See our guide to Big Five domains, aspects, and facets.

Four sources of change: trait, state, error, and scoring

1. Genuine trait change

A personality trait describes a characteristic distribution or tendency across situations and time. Genuine trait change occurs when that enduring tendency shifts. For instance, someone may gradually become more reliable in meeting obligations across many parts of life, rather than merely having one unusually organized week.

Evidence of development comes from longitudinal studies, repeated measurement, and comparisons designed to separate trait change from noise. In a large meta-analysis, Bleidorn and colleagues synthesized longitudinal evidence on rank-order stability and mean-level change, finding both substantial stability and age-related developmental patterns (Bleidorn et al., 2022). These are population findings and do not establish the trajectory of any one individual.

2. Temporary states and contexts

People can act differently from their own average in different circumstances. Someone who usually prefers quiet may be talkative during a reunion; someone generally organized may struggle during a period of acute workload. Such episodes need not imply a lasting change in extraversion or conscientiousness.

Experience-sampling research documented extensive within-person variation in Big Five-relevant states across everyday situations while also finding stable individual differences in people's typical behavioral distributions (Fleeson, 2001). A self-report completed during an unusual period can be influenced by what is most salient to the respondent even when questions request typical behavior.

3. Measurement error and temporary response influences

No questionnaire measures its construct perfectly on each occasion. Item interpretation, response inconsistency, distraction, momentary feelings, and other occasion-specific effects can influence the observed score.

In a meta-analysis covering 682 short-interval test–retest correlations from 74 samples, encompassing 14,923 respondents, Gnambs estimated a median aggregated dependability coefficient of .816 across Big Five scales. Temporary error accounted for approximately 10% of observed score variance in that synthesis, with differences between scales and study conditions (Gnambs, 2014). These are aggregated findings, not a universal error margin for individual tests.

A separate study of several trait-related assessment domains, including the Big Five, found that transient error could account for a consequential share of apparent short-term instability. Its average estimate across the examined measures was nearly one quarter of variance, with strong variation by measure (Chmielewski & Watson, 2009). This result should not be misquoted as saying that precisely 25% of every Big Five score is error.

4. Changes in the test, scoring, and norms

Two score reports can differ because they come from different forms or providers. Tests labeled “Big Five” may contain different items, weights, response scales, normative samples, facet arrangements, and scoring rules. Even when two instruments intend to assess the same domain, a score of 70 on one does not necessarily represent the same location on the trait as a score of 70 on another.

Percentiles are especially sensitive to the reference population. A percentile expresses a person's standing relative to a particular comparison group using a particular scoring procedure. If the norms are updated or the demographic reference group changes, percentile rankings may change without a corresponding shift in the respondent's underlying tendencies. Our guide to Big Five scores and percentiles explains this distinction in detail.

Why retest reliability is important—and what it cannot tell you

Test–retest reliability, in this context, examines the consistency of scores across occasions under assumptions about the stability of the measured construct. For a short interval during which sizable average trait change is not expected, correlations across test occasions help researchers estimate temporal dependability.

A correlation such as .80 means that, within the studied sample and conditions, people with higher scores at the first measurement tend also to score higher at the second. It does not mean every individual's score is accurate to within 20%, that everyone keeps exactly 80% of their personality, or that a person's remaining 20% “changed.” Correlation primarily describes how individuals maintain their relative ordering within a sample.

Retest correlations also depend on the interval between assessments and on how stable the underlying trait is expected to be during that interval. A correlation measured over two weeks answers a different question from one measured over twenty years. Comparing the two as if they were identical measures of test quality confounds measurement error with potential personality development.

Research directly addressing temporal assessment problems emphasizes that observed instability can reflect both true psychological change and measurement limitations (Watson, 2004). Consequently, an instrument's test–retest coefficient is informative only alongside details of its sample, timing, scoring, and intended interpretation.

What the research says about short-term score variation

Short retest intervals are useful because they narrow the period in which sizable, persistent changes in habitual tendencies would ordinarily be expected. They also have limits: people may remember their earlier responses, experience different daily circumstances, or react differently after learning what the test intends to measure.

Gnambs's meta-analysis found that shorter retest intervals were associated with somewhat higher retest correlations for three Big Five domains, while domains and instruments differed in dependability (Gnambs, 2014). It would therefore be misleading to attach one universal percentage of expected change to any repeated Big Five assessment.

A slightly different result after a few days is compatible with ordinary response variability. A large difference deserves closer inspection, but size alone cannot establish its cause. We must know the scale, its reliability, its standard error of measurement, and whether the administration conditions and score metrics were comparable.

Can a real Big Five trait change over months or years?

Yes. Personality development is well supported by longitudinal research, although changes vary in magnitude, direction, and timing. Studies can detect mean-level changes across a population even when people's relative standing remains fairly stable, and an individual may change despite relatively small average change in a group.

Bleidorn and colleagues' 2022 meta-analysis included 189 estimates of rank-order stability involving 178,503 participants and 276 estimates of mean-level change involving 242,542 participants. Their results indicated that rank-order stability increases earlier in life and then tends to plateau in young adulthood, while mean-level patterns differ by domain (Bleidorn et al., 2022). These findings do not imply that a trait is fixed after age 25.

In a longitudinal investigation of 14,718 adults in Germany, Specht, Egloff, and Schmukle examined age and major life events in relation to mean-level and rank-order patterns. Their findings showed that personality changes and stability patterns depend on the domain and stage of adulthood (Specht et al., 2011).

The article's central retesting lesson is that true trait change is plausible over a sufficiently meaningful observation window, but a changed score is not evidence enough by itself. Broad lifespan stability and development belong to dedicated analyses; the assessment question is how well our repeated measurements distinguish signal from noise.

Life events, therapy, and intentional change: what the evidence establishes

People may report different traits after new responsibilities, changes in work and relationships, psychological treatment, or sustained alterations in behavior. However, the existence of an important event in someone's life does not by itself prove that the event caused a particular score change.

A preregistered meta-analysis of life events examined 44 studies and 89 samples, totaling 121,187 participants. It found statistically detectable but generally small and inconsistent effects, with comparatively stronger patterns for work-related transitions than for relationship events (Bühler et al., 2024). Individual experiences may differ considerably from these group-level averages.

Intervention research likewise indicates that personality measures can change under structured conditions. Roberts and colleagues synthesized 207 studies of interventions and observed changes in personality measures across an average intervention duration of 24 weeks, with some changes persisting into follow-up (Roberts et al., 2017). Many included studies involved clinical interventions, so their effects should not be indiscriminately projected onto every healthy adult taking an online questionnaire.

Treatment-related symptom improvement and a changed personality questionnaire may overlap, particularly where items refer to emotional experience. Interpreting that overlap requires careful construct-level analysis. Neither a clinical diagnosis nor treatment advice follows from one Big Five score report.

Three meanings of “stability” that should not be confused

Rank-order stability

Rank-order stability asks whether people tend to preserve their relative positions compared with one another. If a group stays similarly ordered at two time points, rank-order stability can be high even if everyone's average score changes.

Mean-level stability

Mean-level stability asks whether the average level of a trait changes across a group. For example, a study might find higher average conscientiousness at a later age while many people maintain approximately the same relative standing among their peers.

Individual-level stability

Individual-level change concerns whether a particular person's latent trait level has shifted beyond plausible measurement fluctuation. This requires appropriate repeated measurements and an uncertainty model. Two reports shown on an online dashboard rarely provide everything necessary to conclude this scientifically.

These three questions are mathematically and conceptually distinct. A highly stable group average can coexist with meaningful change among individuals, while a changing group average can coexist with high rank-order stability. The longitudinal evidence reviewed by Bleidorn et al. (2022) and Specht et al. (2011) illustrates why both rank-order and mean-level measures are necessary.

How to compare two Big Five test reports responsibly

Start by establishing whether the reports refer to the same construct, instrument, version, and score metric. The following sequence helps prevent common mistakes.

Step 1: Identify the exact test

Record the measure's complete name, version, provider, item count, and administration mode. A BFI-2 domain score, a NEO inventory score, and a very brief online approximation should not automatically be placed on the same numerical scale. The BFI and BFI-2 overview offers an instrument-specific example.

Step 2: Compare like with like

Compare domain scores with the same domain scores from the same version. If one report supplies raw averages and the other supplies percentiles or standardized scores, seek the instrument's official conversion rules. If norms or scoring changed, note that your reports may not be directly equivalent. A 60th-percentile score is neither 60% of a trait nor a diagnostic category.

Step 3: Record the retest interval and circumstances

Write down how long passed between assessments, the setting, whether instructions were the same, and whether the questions asked about behavior generally or during a specific period. Record relevant changes in schedule, sleep, distress, responsibilities, or everyday environment as contextual information rather than presumed causes.

Step 4: Look at the magnitude and consistency of the difference

A modest one-off shift in one domain might reflect transient effects. Repeated findings across multiple suitably spaced assessments, especially when consistent across reliable items or independent informants, may strengthen a hypothesis of change. The necessary evidence depends on the test, the desired degree of certainty, and the construct being studied.

Step 5: Examine documented measurement uncertainty

Look for instrument-specific retest estimates, standard errors of measurement, and relevant validation evidence for the population in which the test was used. Without those details, any statement that a difference of “five points” is meaningful is arbitrary. The interpretation should follow the scale's measurement properties, not a generic internet rule.

Step 6: Interpret the result in everyday context

Ask whether the reported change corresponds with a sustained and varied pattern of behavior, including contexts outside the test. Independent observations can contribute useful evidence, while remembering that informant ratings contain their own sources of disagreement and error. One person's self-perception and another's observations are complementary information, not an automatic truth-versus-error test.

Can you calculate whether a score difference is “real”?

Under a classical measurement framework, the standard error of measurement (SEM) is often expressed as the standard deviation of observed scores multiplied by the square root of one minus an appropriate reliability coefficient. Some applications compare a change score with the expected error of the difference between two measurements.

For two occasions with independent errors, the standard error of a difference can be expressed as the square root of the sum of the two error variances. This gives researchers a way to assess whether an observed difference is larger than what would be expected from measurement imprecision alone, under the model's assumptions.

An individual reader usually lacks the necessary instrument-specific information, appropriate reliability estimate, population norms, and evidence that errors at the two assessments are sufficiently independent. Long-term comparisons add a further complication: genuine developmental change may occur between tests. These formulas therefore explain why uncertainty matters; they do not supply a universal Big Five change threshold.

A statistical threshold, even when validated, would address whether a difference exceeds a specified uncertainty boundary. It would not by itself identify why the underlying trait changed, whether an intervention caused the shift, or what the result predicts for someone's future.

How often should you retake a Big Five test?

There is no scientifically established universal schedule that suits every Big Five inventory and every purpose. A retest interval should match the question you are asking and the validation evidence for the chosen instrument.

Taking the same questionnaire repeatedly over a few hours is unlikely to yield a clean measure of stable personality development. Responses may become familiar, and apparent differences can reflect attention or momentary interpretation. A repeat after a short interval may be relevant when examining score consistency using a protocol designed for that purpose.

A repeat after months or years can be relevant to longitudinal self-observation, especially when the same validated measure and comparable conditions are maintained. Such repeated observations are more informative when the goal is clearly defined in advance and when the results are viewed alongside uncertainty rather than as absolute personal labels.

For many casual users, retesting only when there is a meaningful reason to revisit their assessment is more useful than collecting frequent scores. A single brief questionnaire should not become a running numerical verdict on daily emotional experience.

Why different Big Five tests can give different scores on the same day

Some Big Five measures are longer and assess multiple facets; others are designed for brevity. Different test forms have different tradeoffs in coverage, precision, and practicality.

The 60-item BFI-2, for example, was developed to assess five domains and 15 facets (Soto & John, 2017). Shorter BFI-2 versions were developed for circumstances where assessment length matters, with measurable tradeoffs between efficiency and psychometric performance (Soto & John, 2017, short forms).

The Ten-Item Personality Inventory was designed as an exceptionally brief measure. Its original evaluation examined convergence with longer instruments, observer agreement, and retest stability while acknowledging limitations relative to longer questionnaires (Gosling et al., 200300046-1)). A difference between a 10-item estimate and a comprehensive inventory is not automatically evidence of personality change.

The NEO family uses a different instrument-specific organization from the BFI-2. Research on the NEO-PI-3, for example, describes item revisions designed to improve readability and measurement properties while retaining its intended structure (McCrae et al., 2005). Comparing facet labels across these families as though they were identical would create false precision.

How mood, stress, fatigue, and recent experiences can influence retesting

Questionnaire responses involve memory, attention, self-judgment, and interpretation. A respondent may temporarily rely on unusually vivid recent experiences when estimating what they “generally” do. Someone under intense pressure might focus on current difficulties rather than behavior across ordinary weeks.

This is a plausible mechanism of occasion-specific variability, consistent with research on transient measurement error and behavioral state fluctuations (Chmielewski & Watson, 2009; Fleeson, 2001). The magnitude and direction of a mood-related effect cannot be inferred for an individual without an appropriate study of that instrument and context.

Recent events can also alter someone's perception of their own baseline. Becoming responsible for a new job may cause a respondent to compare themselves with more organized colleagues, for instance. Research on reference-group effects shows that changing the group implied by instructions can influence personality self-reports (Credé et al., 2010). A change in self-evaluation is meaningful information, yet it is not identical to proof of an enduring trait shift.

If emotional or medical concerns are prominent, interpreting them through a five-domain personality chart is especially limited. The Big Five is a model of normal-range personality variation; it is not a substitute for a clinical evaluation.

Scores, percentiles, and labels: why a “high” result can become “average”

Some reports display a raw score, some show an average item rating, and some report standardized scores or percentiles. Percentiles depend on a comparison distribution; they are not direct quantities of the characteristic being measured.

Suppose the raw responses are unchanged while the provider introduces age-adjusted norms or a revised normative sample. The same response profile might then correspond to a different percentile. A high-versus-average label may also change if the report's categories are computed with different conventions.

No universal score cutoffs separate high and low Big Five traits across all instruments. Use only thresholds defined for a particular validated assessment and appropriate comparison sample, and interpret them as descriptive score classifications rather than as clinical boundaries. For conceptual orientation, see why Big Five traits are dimensions rather than personality types.

Can you change one facet without changing the whole trait?

Yes, narrower components may show different patterns from broad domain totals. A person can become more systematic in organizing work without experiencing a parallel increase in every aspect of conscientiousness. Whether an inventory detects this depends on the facets it includes and how those facets are measured.

Facet-level change should not be asserted when an instrument reports only domain-level scores. Nor is there a universal facet inventory shared by every Big Five test. BFI-2's 15 facets, for instance, are part of that instrument's design rather than a fixed property of all Big Five research (Soto & John, 2017).

Broad scores provide a useful summary, but they may conceal meaningful and opposing changes in narrower behaviors. Interpretation requires looking at the correct level of measurement and ensuring that the available evidence supports that level.

Can repeating a personality test itself change your results?

Retesting may change how familiar the questions feel and which examples come to mind. Seeing an earlier report may lead someone to reflect on a trait, notice a behavior, or consciously respond in a way that matches or resists the prior feedback.

These mechanisms are plausible influences on repeated self-report. However, a change after seeing feedback does not automatically prove that taking the test caused trait development. Demonstrating that causal claim requires a design that compares comparable groups or otherwise separates testing effects from changes due to time and context.

Using repeated assessments for self-reflection can be worthwhile, provided that the feedback is not treated as an obligation to behave according to a label or as a definitive map of identity.

When repeated scores are especially hard to interpret

Several circumstances make direct comparison weaker: using different instruments or test versions; switching languages without suitable validation; having different instructions or reference periods; changing the normative comparison group; completing one test hurriedly and the other carefully; comparing facet scores with domain scores; or relying on very brief quizzes with poorly documented psychometrics.

Research on test reliability and instrument brevity shows why these design details are consequential (Gnambs, 2014; Soto & John, 2017, short forms). Language and culture also affect the meaning and functioning of psychological measures. Equivalent translations and appropriately tested scoring cannot be assumed merely because two websites use the same English trait names.

The most defensible conclusion from incomparable reports is often that they provide two related but methodologically different estimates. Treat their discrepancy as a reason to inspect the measures rather than as proof that one report exposed a hidden, permanent personality.

What should you do when two results contradict each other?

First, check for clerical or scoring differences: reversed directions, mislabeled axes, norm updates, or differences between the report's raw and relative scores. Second, verify that both instruments actually assess the same version of the intended Big Five construct.

Third, examine whether the questions were interpreted in reference to the same period and everyday context. Fourth, check the quality of the measure and any reported uncertainty. Fifth, consider a planned later assessment using the same scientifically documented instrument if the question is genuinely important.

If the discrepancy concerns an important employment, educational, or clinical decision, a casual online retest is insufficient. A qualified professional can evaluate what the specific instrument is valid for, whether its norms fit the individual, and how much interpretive weight the result deserves.

The best response is not to average incompatible percentiles, choose whichever profile feels more flattering, or declare that your “real” personality lies halfway between them.

A practical example: what two reports can and cannot establish

Consider an imaginary respondent who completes the same validated conscientiousness scale twice, six months apart. The first report shows a moderate score and the second a higher one. During this period, the respondent started a new role, used a more structured calendar, and describes being more consistent in meeting deadlines.

This pattern is consistent with possible trait change. It is also compatible with shifting self-perception or partly temporary variation. Without sufficient evidence on the measure's reliability, score metric, and expected error—and without carefully assessed behavior across contexts—the amount of enduring change remains uncertain.

Now imagine the second result instead comes from a short quiz on another website using different norms. The comparison is even weaker. A percentage or percentile difference between these two tools cannot be interpreted as a quantitative estimate of personality development.

In a better longitudinal approach, the same validated measure would be administered under comparable conditions on several occasions, with documented scoring and an appropriate model of change. Even then, interpreting the causes of any change would require evidence beyond the repeated scores alone.

Big Five retesting and mental health: essential boundaries

A Big Five inventory may describe tendencies relevant to emotional experience, interpersonal behavior, and self-regulation. A changed score does not establish the onset or remission of a psychiatric disorder, the success or failure of treatment, or someone's need for medication.

A very high or low domain score is also not, on its own, evidence of a personality disorder. The constructs, assessment purposes, diagnostic criteria, and clinical implications differ. Mental-health assessment requires clinical context and tools appropriate to the question being asked.

If emotional distress is the reason you are retesting, a qualified mental-health professional can help evaluate the experience itself rather than attempting to infer a diagnosis from a score change.

Frequently asked questions

Is it normal to get different Big Five results every time?

Some difference is expected because personality questionnaires contain measurement uncertainty and people respond in different circumstances. How much is expected depends on the specific test and interval. Large or repeated discrepancies merit inspection of the instrument and conditions rather than a universal interpretation.

Does a different score mean my personality changed?

It means your observed test result changed. Genuine personality change is one possible explanation; temporary states, response variability, and scoring differences are others. One score difference rarely separates these explanations conclusively.

Can my Big Five score change in a week?

Yes, an observed score can change within a week. A short-term score change may largely reflect situational or measurement influences. Concluding that an enduring trait changed during that week requires stronger evidence than two questionnaire totals.

Can personality really change after age 30 or 50?

Yes. Longitudinal evidence shows personality stability and development throughout adulthood, with different trajectories across domains and individuals (Bleidorn et al., 2022; Specht et al., 2011). The retest interpretation still depends on measurement quality.

If I retake the same test immediately, should my result be identical?

Not necessarily. Remembering questions, changes in attention, and ordinary response fluctuation can influence answers. Immediate retesting is not a validated method for detecting enduring trait change.

What if only one of my five trait scores changed?

A change confined to one domain may reflect a real domain-specific shift or differences in measurement and context. The other four scores remaining similar does not establish which explanation is correct.

Can my percentile change if my raw score stays the same?

Yes. A percentile is calculated relative to a normative distribution. Different norms, updates, or reference groups can produce a different percentile from the same raw responses.

Are free Big Five tests reliable enough for retesting?

Price is not itself a measure of validity. Some publicly available instruments have published psychometric research; other online quizzes give little information about item development, scoring, norms, or retest evidence. Evaluate the exact instrument rather than the site's “free” or “premium” label.

Which result should I trust if two tests disagree?

Give greater weight to the instrument with appropriate validation for the purpose and population, clear scoring, and known measurement limitations. Where measures differ, neither report automatically invalidates the other. Their outputs may not be directly comparable.

Can therapy or important life events alter Big Five scores?

Yes, changes in personality measures have been reported in intervention and longitudinal life-event research. The size and cause of change vary, and effects found in groups do not predict what will happen for one person (Roberts et al., 2017; Bühler et al., 2024).

Is there a number of points that proves personality changed?

There is no universal cutoff across all Big Five tests. Interpreting a numerical difference requires that instrument's metric, reliability evidence, relevant uncertainty, and comparison conditions.

Can a Big Five retest diagnose a mental-health condition?

No. Big Five results measure personality dimensions and are not a stand-alone clinical diagnostic method. A change in a score does not determine whether any mental-health condition is present.

What to remember when retaking the Big Five

Big Five test scores are measurements made under particular conditions, not permanent inscriptions of personality. Repeated results can vary because of error, temporary context, changed scoring, and authentic changes in enduring characteristics. Scientific evidence supports both substantial trait stability and personality development; distinguishing the two at the individual level requires careful measurement.

The most useful comparison is between equivalent instruments, equivalent metrics, documented intervals, and comparable circumstances, with instrument-specific uncertainty taken seriously. A different result is information worth examining. Its interpretation should follow the evidence and the purpose of the assessment.

References

Bleidorn, W., Schwaba, T., Zheng, A., Hopwood, C. J., Sosa, S. S., Roberts, B. W., & Briley, D. A. (2022). Personality stability and change: A meta-analysis of longitudinal studies. Psychological Bulletin, 148(7–8), 588–619. https://doi.org/10.1037/bul0000365.

Bühler, J. L., Orth, U., Bleidorn, W., Weber, E., Kretzschmar, A., Scheling, L., & Hopwood, C. J. (2024). Life events and personality change: A systematic review and meta-analysis. European Journal of Personality, 38(3), 544–568. https://doi.org/10.1177/08902070231190219.

Chmielewski, M., & Watson, D. (2009). What is being assessed and why it matters: The impact of transient error on trait research. Journal of Personality and Social Psychology, 97(1), 186–202. https://doi.org/10.1037/a0015618.

Credé, M., Bashshur, M., & Niehorster, S. (2010). Reference group effects in the measurement of personality and attitudes. Journal of Personality Assessment, 92(5), 390–399. https://doi.org/10.1080/00223891.2010.497393.

Fleeson, W. (2001). Toward a structure- and process-integrated view of personality: Traits as density distributions of states. Journal of Personality and Social Psychology, 80(6), 1011–1027. https://pubmed.ncbi.nlm.nih.gov/11414368/.

Gnambs, T. (2014). A meta-analysis of dependability coefficients (test–retest reliabilities) for measures of the Big Five. Journal of Research in Personality, 52, 20–28. https://doi.org/10.1016/j.jrp.2014.06.003.

Gosling, S. D., Rentfrow, P. J., & Swann, W. B., Jr. (2003). A very brief measure of the Big-Five personality domains. Journal of Research in Personality, 37(6), 504–528. https://doi.org/10.1016/S0092-6566(03)00046-100046-1).

McCrae, R. R., Costa, P. T., Jr., & Martin, T. A. (2005). The NEO–PI–3: A more readable Revised NEO Personality Inventory. Journal of Personality Assessment, 84(3), 261–270. https://doi.org/10.1207/s15327752jpa8403_05.

Roberts, B. W., Luo, J., Briley, D. A., Chow, P. I., Su, R., & Hill, P. L. (2017). A systematic review of personality trait change through intervention. Psychological Bulletin, 143(2), 117–141. https://doi.org/10.1037/bul0000088.

Soto, C. J., & John, O. P. (2017a). The next Big Five Inventory (BFI-2): Developing and assessing a hierarchical model with 15 facets to enhance bandwidth, fidelity, and predictive power. Journal of Personality and Social Psychology, 113(1), 117–143. https://doi.org/10.1037/pspp0000096.

Soto, C. J., & John, O. P. (2017b). Short and extra-short forms of the Big Five Inventory–2: The BFI-2-S and BFI-2-XS. Journal of Research in Personality, 68, 69–81. https://doi.org/10.1016/j.jrp.2017.02.004.

Specht, J., Egloff, B., & Schmukle, S. C. (2011). Stability and change of personality across the life course: The impact of age and major life events on mean-level and rank-order stability of the Big Five. Journal of Personality and Social Psychology, 101(4), 862–882. https://doi.org/10.1037/a0024950.

Watson, D. (2004). Stability versus change, dependability versus error: Issues in the assessment of personality over time. Journal of Research in Personality, 38(4), 319–350. https://doi.org/10.1016/j.jrp.2004.03.001.

Psychological Encyclopedia