# Corporate Value Triad Research Protocol
## Measurement development and tests of the proposed mechanisms

Proposed by Jean Nickson Pijeau
Version 0.1 | October 2026 | Research update October 19 2026
Protocol proposal | Not registered or approved | No data collected

## 1 Purpose and scope
This protocol tests whether evidence exposure, attribution, and operational dependence explain organizational recognition and development access beyond relevant contribution and role context. It also evaluates bounded interventions. CVT is a conceptual proposal. Its survey, archetypes, rubrics, and development kit have not been validated. No study in this package has recruited participants or produced results.

The program has a measurement phase, a longitudinal observational study, a randomized vignette, two optional field trials, and an attribution extension designated Study E. The extension operationalizes five proposed hypotheses from the October 19 update, with explicit identifier mapping. Investigators may conduct them separately. They must register a study-specific protocol before confirmatory data collection, obtain an independent ethics determination and required organizational permissions, and resolve the activation fields in section 15. This proposal is not evidence of approval or a completed preregistration.

## 2 Research questions and estimands
Study A asks whether earlier exposure to contribution evidence predicts later attribution and recognition, and whether dependence predicts development access among eligible employees. These are conditional associations. Study B estimates the causal effect of randomized prominence and information format on credit judgments in a fictional project. Study C estimates the effect of protected capacity plus capability transfer on development access. Study D estimates the added effect of protected practice capacity when both conditions receive the leadership curriculum.

The estimands concern specified populations and outcomes. A vignette effect does not establish workplace promotion discrimination. A field package effect does not identify the effectiveness of each component. Observational mediation remains assumption-dependent. Researchers should prefer a narrow interpretable claim over an unsupported universal mechanism.

## 3 Construct map and measurement sources
Contribution C uses role-relevant episode evidence reviewed independently of reward decisions where feasible. Reviewers record outcome quality, enabling actions, constraints, source coverage, and uncertainty. Retain indicators separately until a composite has a defensible measurement model. Do not use task counts, hours, or self-rated contribution as the sole proxy.

Visibility V records evaluator exposure to specific evidence: whether an artifact was seen, a relevant update received, or an episode discussed. It excludes favorability judgments. Attribution A is credit assigned to a named action in a named episode. Recognition N is whether formal records accurately acknowledge that action. Influence I records demonstrated ability to affect a decision or secure a resource, while documenting authority and task context. Survey reports of access are a separate source.

Operational dependence D is the baseline presence of at least one role-defined critical recurring task without demonstrated backup coverage. Define criticality before reviewing people. Record supporting indicators separately. Opportunity access O is commencement of a qualifying development assignment during follow-up among employees meeting predeclared eligibility criteria. Promotion P is a distinct authorized-record outcome over 12 months among eligible employees. Voluntary declines, no available vacancy, and missing records remain distinguishable.

The pilot diagnostic supplies 24 original items about experience in six content domains. It does not establish C or validate I. Objective or independent records require separate consent and authorization. Keep the evidence source and timing attached to every observation. The Contribution Attribution Assessment adds eight separate pilot items, an elicited accomplishment description, manager prompts, and history checks. Collective framing, signaling, exposure, credit, and recorded recognition must remain distinguishable. No composite or cutoffs for the new module are authorized.

The Collective Attribution Paradox proposes a conditional association between collective-focused framing and individual recognition. The Attribution Dilution Effect concerns loss of acknowledgment for a fixed originating action during adoption. Contribution Signaling is explicit communication of personal actions, distinct from evaluator exposure. All are proposed constructs. Related literature [R19, R20, R21, R23] motivates alternatives and measurement development, without validating the proposed effects.

## 4 Falsifiable hypotheses
The directional sign and primary estimand for each hypothesis must remain fixed after registration. Use two-sided intervals and the multiplicity procedure in section 9. A failure to reject a null with wide intervals is inconclusive, not proof of no mechanism.

| ID | Prediction and primary estimand | Main design | Evidence against the proposed direction |
| --- | --- | --- | --- |
| H1 | Earlier V positively predicts later A conditional on baseline C and context; adjusted V slope | A | A precise zero or negative association within the prespecified setting |
| H2 | Structured records attenuate the prominence effect on credit; format by prominence interaction is negative | B | A precise negligible attenuation or amplification |
| H3 | Earlier A positively predicts later N conditional on C, V, and context; adjusted A slope | A | A precise negligible or negative association |
| H4 | Baseline D predicts lower O among eligible employees; adjusted probability contrast D present versus absent | A | A precise negligible or positive contrast |
| H5 | Protected capacity plus transfer increases O versus usual practice; assignment effect on probability | C | A precise negligible or negative effect |
| H6 | Protected practice plus curriculum improves decision quality versus the same curriculum with ordinary capacity; assignment effect | D | A precise negligible or negative effect |
| H7 | I strengthens the N to promotion association; average interaction contrast on probability scale | A | A precise negligible or negative interaction contrast |
| H8 | Independently assessed evaluation transparency weakens the V to N association conditional on C; cross-level interaction | A | A precise negligible or positive interaction |

H1 and H3 concern separate stages and do not establish an indirect effect. H7 is conditional on actual eligibility and sufficient promotion events. H8 requires enough organizations and variation in policy, not merely many employees in a single organization. If those conditions are absent, register them as exploratory or omit them before observing outcomes. Do not quietly substitute an employee perception item for an organizational transparency measure.

### October 19 hypothesis identifiers
The initial package already used H7 and H8. Their meanings remain unchanged. To preserve an unambiguous registry, the five additions receive H9 through H13. The update author's original labels remain visible in this crosswalk. This is an editorial identifier mapping, not a change to an existing registration. No version of this protocol has been registered.

| Update label | Stable package ID | Proposed relationship |
| --- | --- | --- |
| H7 | H9 | Collective attribution and lower individual recognition conditional on contribution |
| H8 | H10 | Managerial exposure moderates signaling and recognition |
| H9 | H11 | Originating-action acknowledgment declines during adoption |
| H10 | H12 | Explicit contribution documentation reduces attribution-style disparities |
| H11 | H13 | Team-oriented norms relate to accomplishment communication and recognition seeking |

H9 predicts a negative adjusted recognition contrast for a collective-focused account versus an account specifying the person's action. H10 predicts a nonzero exposure by signaling interaction; the update does not specify its direction. H11 predicts a negative contrast in originating-action acknowledgment at wider adoption versus creation. H12 predicts a smaller positive role-explicit versus collective-focused recognition contrast under structured documentation. H13 predicts a nonzero association between reported team-oriented norms and attribution style, with recognition seeking as a secondary outcome. These are proposed tests, not findings. If the no-documentation contrast in H12 is reversed, report that result and do not describe a more negative contrast as improved equity.

## 5 Phase 0 measurement development
Recruit a varied pilot including technical, coordination, operational, customer-facing, and managerial work. Use approximately 12 to 20 cognitive interviews as a planning range, continuing until major interpretation problems are resolved. This is a practical target, not a guarantee of content validity. Ask participants to explain each item, identify its reference period, and distinguish no opportunity from a negative experience.

Have subject-matter reviewers and employee representatives assess coverage and potential bias. Record item revisions and reasons. Pilot the evidence rubric with double-rated episodes, including prevention and joint production. Quantify agreement with appropriate uncertainty and examine disagreement qualitatively. High agreement does not establish truth.

Use separate exploratory and confirmatory samples for latent measurement work. Analyze ordinal items appropriately, examine missingness, item distributions, factor structure, reliability with intervals, test-retest behavior over a justified interval, and convergent and discriminant relations. Assess differential item functioning and measurement invariance where sample sizes permit. Cronbach and Meehl (1955 [R12]) and Flake et al. (2017 [R13]) motivate a broader validation program than reliability alone. No arbitrary reliability threshold licenses employment use.

Lock measures and the primary model before the confirmatory phase. If contribution indicators behave as formative evidence rather than interchangeable reflective items, do not force them into an internal-consistency scale. If visibility and attribution cannot be separated, redesign before claiming a test of the mechanism.

## 6 Study A longitudinal observational design
Recruit consenting adult employees through an independent research team across multiple organizations, roles, and levels. Obtain organization-level recruitment counts and eligible sampling frames where authorized. Managers must not select only favored employees or receive participation lists. A broad invitation with private opt-in reduces coercion but still permits volunteer bias, which the report must acknowledge.

At month 0 obtain baseline episode evidence, role and level, relevant tenure, prior recognition and opportunity, D, workload, assignment resources, preferences, eligibility, and evaluator context. At month 1 record evaluator evidence exposure. At month 2 collect attribution judgments. At month 3 collect subsequent formal recognition records, preserving actual dates and checking that recognition followed attribution measurement. At month 6 observe development access and workload. At month 12 record promotion outcomes for separately consenting participants. If an organization uses a different cycle, register the actual timing and preserve temporal order before recruitment.

Sample episodes using a predeclared rule such as the most recent completed qualifying assignment, not the employee's most successful example. Define a qualifying episode by duration, outcome relevance, and available evidence before selection. Record excluded and inaccessible episodes. Avoid selecting only people who survived to promotion review. Track departures as outcomes or competing events when appropriate and consent permits.

Use mixed models or fixed effects suited to the estimand and number of clusters. Episodes may be crossed by employee and evaluator and nested within teams and organizations. For H1 model later A from prior V and prespecified C indicators with baseline attribution where available. For H3 model later accurate acknowledgment from prior A. For H4 estimate standardized probability contrasts for O. For H7 use N, I, and their interaction, reporting the prespecified probability-scale difference in the N contrast between higher and lower I values within common support. State those I values before outcome analysis. For H8 use independently double-coded policy evidence and a cross-level V interaction.

Baseline controls include role family, level, tenure, prior outcome, assignment resources, task interdependence, and evaluator relationship where justified. Compensation models would additionally need relevant compensation context and remain optional. Draw a study-specific causal diagram first. Do not add post-exposure controls merely because they improve fit. Policy transparency uses document availability, stated criteria, correction access, and consistency records as separate indicators until justified otherwise.

## 7 Study B randomized vignette
Recruit adult evaluators with defined supervisory or project-review experience, or explicitly identify a different target population. Use a 2 by 2 between-participant design with equal allocation: prominent focal presenter versus no distinctive presenter, crossed with narrative versus structured contribution evidence. Hold contribution facts, project outcome, and role responsibility constant. Use neutral names and counterbalance the focal identity independently.

Primary outcome: credit assigned to the focal contributor on a 0 to 100 scale. Fit a linear model with prominence, format, their interaction, and counterbalance factors. Code format 1 for structured and prominence 1 for prominent. H2 predicts a negative interaction. Report all four cell means and the difference in prominence effects with intervals. A robust or randomization-based sensitivity analysis checks dependence on distributional assumptions.

Pretest whether prominence implies extra contribution or responsibility and whether the formats differ in comprehension. Repair the stimuli before confirmation if the manipulation fails. Measure comprehension after the credit outcome to avoid priming. Exclude ineligible or duplicate respondents by rules established before assignment. Include attention failures in the primary randomized analysis and report a registered sensitivity analysis. Document randomization method, allocation concealment, and attrition by cell.

## 8 Field trials and the attribution extension
Study C assigns eligible teams to protected development time with actual work transfer and backup practice, or usual approved practice with delayed access where ethical and feasible. Primary endpoint: commencement of a qualifying development assignment within 90 days. Study D gives both conditions the same curriculum and equivalent scheduled contact. Only the intervention condition receives a formal agreement that removes or transfers specified obligations and protects practice time. Primary endpoint: independently rated decision quality on a role-relevant task at day 90, adjusted for baseline task quality.

Use team allocation to limit contamination, stratifying on organization and baseline task context where feasible. Recruit and measure baseline before randomization. Analyze assignment rather than adherence as primary, accounting for clustering and stratification. For small cluster counts, use small-sample corrections or a prespecified randomization analysis. Reviewers should be unaware of allocation where feasible. Report violations of masking. Keep O, task quality, delivery reliability, participant workload, colleague workload, and adverse experiences distinct.

Do not remove required training, ordinary rights, or needed support from a comparison group. The investigator and independent reviewer must approve workload and service-quality guardrails before launch. Pause new assignments for serious retaliation, privacy breach, or safety concerns. Record what happened and obtain independent clearance before restarting. An intervention that improves access but harms service or colleagues requires a qualified interpretation.

### Study E measurement and observational cohort
E1 follows consenting employee-innovation episodes prospectively from creation through initial adoption and wider institutional use. Include an eligible comparison cohort that has not yet reached each stage. Specify stage criteria independently of recognition: first use by another authorized work group and later integration into an approved routine process are possible definitions, to be adapted before recruitment. Record actual elapsed time, adoption decisions, new contributors, assignment resources, evaluator turnover, and record accessibility. Do not treat adoption as randomized.

Before evaluative outcomes, ask each participant for the same neutral accomplishment description with the same space allowance. Two trained coders, unaware of later recognition, code whether the account specifies a personal action, acknowledges others, or describes only a team result. These are separate flags. For the primary H9 contrast, collective-focused means team-result language without a specified personal action, while role-explicit includes a specified personal action with or without team acknowledgment. Ambiguous and unclassifiable accounts remain separate. Pronouns alone cannot determine the code. Pilot a detailed codebook, double code, report agreement with intervals, and retain disagreement before adjudication. A description under this prompt does not establish a stable communication trait.

H9's primary estimand is the adjusted probability difference in subsequent accurate acknowledgment of the focal corroborated action, collective-focused minus role-explicit. Define the acknowledgment window and qualifying record before launch. Independently review contribution indicators before measuring recognition. Adjust using a prespecified causal diagram for baseline contribution evidence, task, role, level, prior recognition, evaluator knowledge, and opportunity to communicate. Report the total style association as primary. Exposure can be on the proposed pathway, so add it only in a separately labeled conditional model. Residual confounding and contribution measurement error preclude a causal claim.

For H11, preserve a fixed set of corroborated originating actions as the denominator across stages. Code whether the relevant later account accurately names each originating action, and separately code later improvements. The primary contrast is standardized probability of originating-action acknowledgment at wider adoption minus creation, conditional on elapsed time, contribution context, later work, and evaluator continuity. Use repeated-measure models with employee, episode, evaluator, and organization dependence as appropriate. Model unavailable accounts separately from observed accounts without acknowledgment. Compare with episodes at the same elapsed time that have not yet advanced, examine selection and departures, and report limitations of the adoption comparison. A declining credit share alone is not evidence for H11.

For H13, obtain baseline TN1 about encouragement of collective framing before eliciting the description. The primary estimand is the adjusted association of the ordinal TN1 response with the probability of a collective-focused description. Specify ordinal coding or category contrasts prospectively, without treating the eight-item module as one scale. The prediction is two-sided. TN2 and interview questions about asking for accurate credit address recognition seeking as secondary outcomes. Do not infer norms from military service, profession, nationality, gender, or any other identity. Prior communication style and selection into teams remain rival accounts.

### Study E factorial communication experiment
E2 uses a fictional work episode in a 2 by 2 by 2 between-evaluator design. Randomize attribution framing in a truthful update, evaluator receipt of that update, and contribution-record format. Framing compares a team-focused outcome account with a matched account that specifies the focal person's action while acknowledging others. Receipt means the evaluator sees the update; the comparison receives neutral matched-length material. Record format compares narrative and structured histories containing the same contribution facts. All evaluators initially receive the same baseline contribution facts. Counterbalance identities. No real employee messages or company materials enter the stimuli.

Keep outcome value, originating work, later improvements, formal responsibility, and factual contribution constant. Match length and readability as far as feasible. Pretest whether a role-explicit account implies extra work, whether receipt is understood, and whether structured records change factual content. Redesign failed manipulations before confirmation. The style manipulation necessarily changes communication wording; interpretation is limited to that specified wording. An E2 effect is not evidence of a stable employee characteristic.

The primary endpoint is accurate acknowledgment of the focal originating action in an evaluator's written review, coded yes or no by assessors unaware of assignment using a preregistered rubric. Preserve partial, incorrect, and unavailable responses for secondary and missingness analyses. Elicited credit on a 0 to 100 scale is secondary and distinct from written recognition. Analyze all eight randomized cell probabilities using a prespecified saturated model and marginal contrasts, with randomization-based sensitivity analysis.

For H10, estimate the role-explicit minus collective-focused probability contrast under receipt minus the corresponding contrast without receipt, averaged equally over record formats. Test the exposure by framing contrast two-sided. For H12, within the receipt condition estimate that style contrast under structured documentation minus the contrast under narrative documentation. A negative difference alone does not establish reduced disparity. The predicted attenuation requires an explicit-attribution advantage under narrative documentation and a smaller absolute style difference under structured documentation. Prespecify uncertainty-based interpretation of these component contrasts and a bound for a meaningful reverse disparity. A large reversal is an adverse or contradictory result. Report all component probabilities and contrasts even if the predicted disadvantage never appears. The experiment does not test actual compensation or promotion. Do not merge H12 with legacy H2, which manipulates presenter prominence rather than attribution framing.

Allocate equally across the eight cells, conceal assignment, and register counterbalancing, eligibility, duplicate removal, outcome coding, and missingness rules. Include attention failures in the primary assigned analysis and report prespecified sensitivity analyses. Assessor disagreement is not a missing value to discard. Fix an adjudication procedure before observing condition results. Adequate power for interaction contrasts requires its own simulation, not reuse of a main-effect sample calculation.

### Study E safeguards and disconfirmation
The reported episode that prompted the extension is not E1 pilot data, an interview dataset, or evidence for any hypothesis. Do not recruit colleagues through the reporting employee, inspect internal messages, or contact an employer on the strength of this package. Any future qualitative study requires separate consent, independently collected perspectives, company confidentiality clearance where required, and a disclosure plan. Published materials omit real employee names, screenshots, company and product identifiers, and distinctive service histories. External clearance and permission are not established by anonymizing a narrative.

H9 is weakened by a precise negligible or positive collective-focused contrast. H10 and H13 are weakened by precisely negligible moderation or association under justified equivalence bounds. H11 is weakened by stable or increased originating-action acknowledgment with adequate precision. H12 is weakened by precisely negligible attenuation or an amplification of the documented style disadvantage. An imprecise null remains inconclusive. Successful acknowledgment of the originator in both styles would challenge a universal disadvantage claim. Test productive communication, new contribution, recall decay, evaluator knowledge, and selection as alternatives.

## 9 Precision sample size and multiplicity
No sample in this proposal has demonstrated adequate power. Before recruitment, define the smallest effect of interest with stakeholders and justify it independently of expected statistical significance. Illustrative planning values are a 5-point interaction on the 0 to 100 credit scale, a 5 percentage-point development-access change, and a 0.15 baseline-SD decision-quality difference. These are proposed research design choices, not CVT thresholds, established benefits, or validated norms. Register final values and reasons before data collection.

Simulate the exact models over plausible variance, intraclass correlation, event rates, cluster counts, unequal cluster sizes, measurement error, and attrition. Target at least 90 percent power for the selected smallest effect and inspect expected interval width. Report assumptions and sensitivity rather than a single optimistic N. Study E adds separate requirements for the style, exposure interaction, documentation interaction, adoption contrast, and norms association. Define their smallest effects and equivalence bounds before data collection; the illustrative Study B credit-scale value cannot be transferred to a binary acknowledgment outcome. Include coder error, adoption selection, repeated measures, and cluster counts in E1 simulations and all eight cell sizes in E2 simulations. For rare promotion events, more clusters and longer follow-up may matter more than extra respondents. Lakens (2022 [R17]) provides methodological guidance on explicit sample-size justification.

For a single integrated program, apply Holm adjustment across the thirteen primary hypothesis tests at familywise alpha .05, with two-sided tests and a sign check. If independent registered studies form separate families, state those families prospectively and report all tests without selective regrouping. Report unadjusted estimates, intervals, raw p-values, and adjusted p-values. The five extension tests may form a distinct family only if that choice is registered before outcomes are inspected and all thirteen tests remain visible. H13 recognition-seeking outcomes, secondary mediation, subgroup comparisons, and alternative models remain explicitly exploratory unless separately registered with their own multiplicity plan.

Use equivalence testing only with justified bounds fixed in advance and adequate precision. A nonsignificant test does not show equivalence. Do not compare raw slopes to claim visibility matters more than contribution. A separate predictive analysis may compare cross-validated performance using common outcomes, appropriate scaling, and uncertainty, with no causal interpretation.

## 10 Missingness selection and robustness
Maintain distinct codes for skipped, not applicable, no opportunity, unavailable record, withdrawal, and lost follow-up. Do not code these as zero or the scale midpoint. Report missingness and retention by condition, role, and permitted group summaries. Prespecify handling for each source. For observational estimates, use multiple imputation under a justified missing-at-random model including cluster and outcome information, with complete-case and departure-from-MAR sensitivity analyses. Imputation does not create unavailable ground truth.

For trials, retain assignment in the primary analysis and describe outcome availability. Use a registered missing-outcome strategy and bounds or sensitivity analyses for differential attrition. Report a separate treatment uptake analysis. Check common support, influential clusters, alternative plausible contribution indicators, evaluator effects, and models that include coordination as productive work. Do not select the version that best supports CVT.

Mediation is secondary. State the exposure-mediator and mediator-outcome identification assumptions, address post-treatment confounding, and conduct sensitivity analysis (Imai et al., 2010 [R15]). A temporal association or bootstrap interval alone does not establish a causal mechanism. Shared-source concerns require design controls, not only a statistical diagnostic (Podsakoff et al., 2003 [R09]).

## 11 Ethics and participant protections
Use respect for persons, beneficence, and justice as guiding principles (National Commission, 1979 [R18]). Obtain an independent ethics determination appropriate to the jurisdiction and institution. The Belmont Report is an ethical foundation, not a claim that one regulatory regime applies everywhere. Confirm applicable employment, privacy, and research obligations locally before launch.

Recruitment must explain that participation is voluntary, refusal has no employment consequence, and questions may be skipped. Supervisors must not collect consent or watch completion. Offer accessible formats, translation developed through testing, and reasonable paid participation time where approved. Avoid incentives contingent on favorable responses. Separate consent for surveys, records linkage, interviews, and recontact.

The three new constructs and all Contribution Attribution Assessment items remain unvalidated proposals. Managers may not use a communication-style code to reward assertiveness, penalize collective credit, or identify allegedly manipulative colleagues. No individual CVT score or archetype enters hiring, compensation, promotion, discipline, redundancy, or a personnel file. Employers receive only disclosure-reviewed aggregates and process findings. The research team must have authority to refuse identifiable data requests and publish null or adverse findings. State sponsor relationships and conflicts. Participants can withdraw identifiable data until the stated de-identification cutoff, after which removal from pooled data may be impossible.

## 12 Data governance and disclosure
Assign random study IDs. Keep the identity key with an independent custodian in a separately encrypted location with access logging. Researchers analyzing data use pseudonymous records. Pseudonymity is not anonymity. Do not promise anonymity for linked longitudinal work. Collect no private messages or communication surveillance. Network nominations, if later justified, require a separate review because they reveal information about nonparticipants.

Use a written access matrix. Custodian: identity key and consent. Analysts: minimized pseudonymous research data. Independent safety reviewer: necessary incident information. Employer: approved aggregate outputs only. A proposed default retains the key through linkage plus 90 days, then deletes it, and retains minimized data for five years after final publication, subject to the approved protocol and applicable requirements. Register actual dates and document deletion or justified extension.

Suppress cells with fewer than 10 respondents as a conservative operational starting rule, apply complementary suppression, and withhold larger cells when distinctive roles or combinations remain identifiable. Ten is not a proven anonymity threshold. Review free text, quotations, longitudinal combinations, and differences between tables. Seek explicit quotation permission and remove identifying details. Do not release row-level employee data publicly. Share synthetic demonstrations, code, dictionaries, and disclosure-reviewed aggregate results where feasible.

## 13 Participant information and consent template
Study title: Corporate Value Triad study of work evidence and recognition. This study asks how organizations acknowledge work and provide opportunities. The framework and questionnaire are experimental. Participation may involve a survey, discussion of a work episode, and follow-up. The local information sheet must state the actual activities, time commitment, compensation, and any randomized assignment.

You may decline, skip questions, or stop without an employment penalty. Your manager will not receive your individual answers or a CVT label. The main risks are discomfort about workplace experiences and accidental identification. The research team will limit access and review reported summaries, but cannot promise absolute confidentiality. Confidentiality limits, including any required disclosures, must be explained before consent.

Choose separately: I agree to the survey [ ]; I agree to authorized organizational-record linkage [ ]; I agree to an interview [ ]; I agree to recontact [ ]. Each choice is voluntary. Consent to one activity does not authorize another. The local sheet must identify the withdrawal cutoff, researcher contact, independent concerns contact, privacy contact, and approved data-retention dates before distribution. These roles cannot be invented in a generic package.

## 14 Reporting and open research
Register hypotheses, outcomes, exclusions, randomization, sample plan, analysis, and decision rules before inspecting confirmatory outcomes. Keep a dated deviation log with reasons. Preregistration helps distinguish prospective tests from later exploration (Nosek et al., 2018 [R16]). It does not make a weak measure valid or remove ethical responsibilities.

Report recruitment flow, eligible population, missingness, intervention delivery, effect estimates, uncertainty, multiplicity, adverse effects, and all registered outcomes. Separate theory, prior evidence, new results, and practical recommendations. Invite employee review of disclosure risk without giving sponsors veto over unfavorable findings. Publish an explicit statement if the data contradict CVT. Update the framework version rather than rewriting original predictions after results.

## 15 Activation record and launch gate
Complete before recruitment: named principal investigator and institution; independent ethics decision and identifier; research and privacy contacts; sites and target population; actual sampling frame; approved consent sheet; exact measures and stimuli; eligibility and episode rules; simulation-based sample justification; randomization and analysis scripts; disclosure rules; retention dates; sponsor agreement; safety and workload guardrails; withdrawal cutoff; and timestamped registration link.

Decision: proceed [ ] revise [ ] stop [ ]. Decision owner: __________. Date: __________. Unresolved conditions and resolution owner: __________. This record makes the proposal operational without misrepresenting a generic document as an approved study.


## References

[R09] Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879-903. https://doi.org/10.1037/0021-9010.88.5.879

[R12] Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52(4), 281-302. https://doi.org/10.1037/h0040957

[R13] Flake, J. K., Pek, J., & Hehman, E. (2017). Construct validation in social and personality research: Current practice and recommendations. Social Psychological and Personality Science, 8(4), 370-378. https://doi.org/10.1177/1948550617693063

[R15] Imai, K., Keele, L., & Tingley, D. (2010). A general approach to causal mediation analysis. Psychological Methods, 15(4), 309-334. https://doi.org/10.1037/a0020761

[R16] Nosek, B. A., Ebersole, C. R., DeHaven, A. C., & Mellor, D. T. (2018). The preregistration revolution. Proceedings of the National Academy of Sciences, 115(11), 2600-2606. https://doi.org/10.1073/pnas.1708274114

[R17] Lakens, D. (2022). Sample size justification. Collabra: Psychology, 8(1), Article 33267. https://doi.org/10.1525/collabra.33267

[R18] National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research (1979). The Belmont Report: Ethical principles and guidelines for the protection of human subjects of research. U.S. Department of Health and Human Services. https://www.hhs.gov/ohrp/regulations-and-policy/belmont-report/read-the-belmont-report/index.html

[R19] Exley, C. L., & Kessler, J. B. (2022). The gender gap in self-promotion. The Quarterly Journal of Economics, 137(3), 1345-1381. https://doi.org/10.1093/qje/qjac003

[R20] VanEpps, E. M., Hart, E., & Schweitzer, M. E. (2024). Dual-promotion: Bragging better by promoting peers. Journal of Personality and Social Psychology, 126(4), 603-623. https://doi.org/10.1037/pspi0000431

[R21] Ross, M., & Sicoly, F. (1979). Egocentric biases in availability and attribution. Journal of Personality and Social Psychology, 37(3), 322-336. https://doi.org/10.1037/0022-3514.37.3.322

[R23] Connelly, B. L., Certo, S. T., Ireland, R. D., & Reutzel, C. R. (2011). Signaling theory: A review and assessment. Journal of Management, 37(1), 39-67. https://doi.org/10.1177/0149206310388419