Clock Zero

A companion guide to our syndicate’s scholarly simulation
Durham — emlyon Global DBA — Innovation and Technology Management
Focal paper: Zhu et al. (2016), The impact of organizational culture on concurrent engineering, design-for-safety, and product safety performance

---

Imagine the scene in Bologna. Our syndicate exercise brief appears simple enough: read Zhu et al. (2016) and present the paper’s methodology, limitations and future research. Twenty minutes. Three headings. A respectable set of slides could do the job.

But the module leaders have already warned us not to read a journal article as if it were a textbook chapter. The syndicate brief reminds us that every paper has a "unique story and a journey" (Syndicate Exercise Requirements, 2026, p. 1). To understand that journey, we may need to read the paper several times, follow its citations and reconstruct decisions that the polished article no longer makes visible.

So instead of standing safely at the end of the journey of Zhu et al. (2016) and listing what seems strong or weak, the simulation puts us inside it.

The same paper will be encountered three times. Each encounter changes what is knowable, what can still be changed and what responsibility comes with that position.

But time is only one dimension of the investigation.

A second lens travels with the paper: Six Drivers of methodology. They ask where the study comes under pressure — through its design, sampling, measurement, unit of analysis, analytical technique and unresolved rival explanations.

That gives the game its harmonised architecture:

Three Clocks asks when judgement is being made.
Six Drivers asks where the methodology comes under pressure.

Together they ask:

What can responsibly be claimed from this evidence, at this point in the paper’s life?

This is not a test of whether we can guess what Zhu et al. did. It is a rehearsal for something closer to home: what happens when the research being designed, reviewed and eventually inherited is ours?

---

The paper at the centre of the simulation investigates a problem that matters well beyond one industry. Products are designed inside organisations, and those organisations make choices about priorities, resources, collaboration, review, safety practices and trade-offs.

Zhu et al. (2016) frame their study around five constructs: Management Commitment to Safety (MCS), Product Safety Culture (PSC), Concurrent Engineering (CE), Design-for-Safety (DFS) and Product Safety Performance (PSP).

Before inheriting that theoretical architecture, the game returns to a more phenomenon-neutral question:

What organisational conditions and product-development practices help explain product safety performance?

The distinction matters. A problem is not a theory. A theory is not a construct. A construct is not its measure. A statistical relationship among measures is not automatically the causal story suggested by their labels.

The game lives in the spaces between those things.

---

Clock	Historical position	Role	Governing question
August 2008	Before fieldwork	Supervisory team	What evidence should this study be designed to generate?
March 2016	Before publication	Editorial review panel	What claims should this evidence be permitted to support?
October 2026	After a decade of inheritance	Fellow researchers	What knowledge should later researchers be permitted to inherit?

The clocks protect the investigation from hindsight. Evidence available in 2026 cannot be used to pretend that a 2008 design decision was obviously wrong at the time. Equally, historical fairness does not require later scholarship to inherit a claim unchanged.

A hidden fourth clock — Clock 0 — belongs to us as doctoral students in the present. It is not part of the paper’s historical chronology. It is the observer and development position from which we enter the three playable clocks and to which we return afterwards.

The second lens follows the methodology rather than the calendar.

Driver	Travelling question
Design	What kind of relationship can this design establish?
Sampling	Who can the evidence legitimately speak for?
Measurement	What was actually measured?
Unit of analysis	At what level does the evidence speak?
Analytical technique	What can the analysis establish?
Rival explanations	What else could produce the same pattern?

These are not six criticisms. They are six directions from which methodological pressure can arrive.

The full Six-Driver inquiry examines 30 methodological axes. For each axis, the first task is simple: establish what the paper actually did. The answer is then compared with the strongest claim made on the same axis. Three outcomes are possible:

• Covered — the claim stays within what was done
• Not covered — the claim goes beyond what was done
• Not addressed — the paper leaves the question open

The judgement itself can then be distinguished as a reported fact, methodological inference, vulnerability of unknown magnitude or unresolved rival explanation.

A claim component survives as stated only when the necessary Drivers cover it. The point is not to make criticism harsher. It is to make criticism traceable.

The Six Drivers will change jobs as the clocks move:

design levers → methodology stress test → inheritance check

---

The same spaces exist in our own doctorates. Long before a thesis is submitted, choices are made that later look inevitable on the printed page. One theoretical perspective enters the frame and another does not. An abstract idea becomes interview questions, survey items, observations or archival variables. Someone is chosen to speak for an organisation. A source of evidence becomes an outcome measure. Months or years later, analysis can only work with what those earlier choices made observable.

Then comes the uncomfortable question:

Does the evidence really support the story?

A supervisor may ask it before fieldwork. A reviewer may ask it before publication. A later researcher may ask it before carrying a citation into a literature review.

Three Clocks, Six Drivers lets us practise all three conversations while also seeing which methodological choices are doing the work underneath them.

---

There is another reason for the format. The simulation is itself an Innovation & Technology Management experiment. McCarthy et al.’s 2006 complex-adaptive perspective invites us to see New Product Development not as a tidy linear sequence but as a system of interdependent decisions whose consequences can amplify, constrain or redirect what follows. Milewski et al.’s 2015 lifecycle perspective adds a second idea: as an innovation develops, technical and organisational elements adapt to one another and the space for intervention changes over time.

Now turn those lenses onto research. A doctoral study also develops through interdependent choices:

problem → theory → constructs → measures → respondents → evidence → analysis → claims → publication → later citation

The arrows are not a claim of simple causality. They show path dependence. An early definition changes what can be measured. Measurement changes what the dataset can contain. The dataset constrains what an analysis can reveal. The design constrains what a claim can responsibly say. Publication then creates a relatively durable scholarly artefact that later researchers may inherit, recombine or contest.

Seen this way, research design resembles innovation architecture. Some decisions are still reversible early on. Others become expensive or impossible to change once evidence has been collected. Review can reconfigure interpretation but cannot reconstruct data that were never gathered. Later scholarship can create a successor design, but it cannot retroactively repair the original study.

The Three Clocks make those shifting intervention spaces visible. The Six Drivers show where design choices create methodological boundaries as the study moves through its lifecycle.

That is the Innovation & Technology Management scholarly analogy behind the game:

Research itself can be examined as a path-dependent system of decisions under uncertainty, with design choices, lock-in, review gates, reconfiguration and successor innovation.

Once that analogy is visible, the simulation becomes more than a presentation device. It becomes a small laboratory for thinking about how knowledge is designed, stabilised, challenged and renewed.

---

Vitae’s 2025 Researcher Development Framework describes researcher development across personal and professional values and behaviours, research knowledge and skills, research management and wider research communities. Relevant capabilities here include being development-focused, strategic, curious, adaptable, reflexive and responsible, alongside subject knowledge, research methods, information and data literacy, digital and technological capabilities, critical thinking, creativity and innovation.

A list of capabilities can remain abstract until a situation demands them. The simulation turns them into decisions.

At Clock One, can we design and justify when the answer is not yet known? At Clock Two, can we scrutinise and restrain when an attractive result tempts us to say more than the evidence warrants? At Clock Three, can we synthesise and create when inherited claims may have travelled further than their original measures?

But those three clocks belong to the paper. There is another clock that belongs to the participants.

Clock 0 is the doctoral present. It is where we are now — learning how to read, challenge, design and eventually own scholarly judgement before the equivalent decisions in our own DBA become irreversible. Zero does not mean “chronologically before 2008”. It marks the observer position from which we enter the historical simulation and to which we return after it.

The paper is being investigated, but so are our standards of judgement. The deeper developmental question is not whether we can identify the “best” option in a game. It is whether the reasoning behind our choice becomes sufficiently visible that we can later reproduce, challenge and improve it without an answer key.

---

The first half of this companion is deliberately spoiler-light. It explains the roles and analytical lenses without supplying an answer key before play.

For each clock, carry four historical questions:

1. What is known at this moment?
2. What is not yet knowable?
3. What can this role still change?
4. What responsibility comes with that power?

Keep the Six Drivers nearby as a second set of questions:

Design · Sampling · Measurement · Unit of analysis · Analytical technique · Rival explanations

They will not all matter equally in every decision. That is deliberate. A framework becomes useful when it clarifies a judgement, not when every category is forced into every scene.

There is no score to maximise. There is a judgement record to build.

---

Before entering 2008, notice where we are actually standing. We are doctoral students in 2026, reading someone else’s completed research while still developing the habits that will shape our own.

Clock 0 is deliberately outside the paper’s historical sequence. It is the meta-clock of researcher formation. From here, we can temporarily enter the roles of supervisor, reviewer and later scholar without carrying the same consequences that those roles would have in our own projects. That makes the simulation a safe rehearsal space for decisions we will eventually have to own.

The governing question for Clock 0 is therefore:

What kind of researcher are we becoming while learning to decide what evidence can carry?

This clock also changes the meaning of the Six Drivers. They are not only a way to criticise Zhu et al. (2016). They are a prospective checklist for our own work: what design will our question require, who can our sample speak for, what will our measures really observe, at what level can we make claims, what can our analytical technique establish and which rivals will remain open?

Innovation & Technology Management lens

Clock 0 treats doctoral development as capability-building for an uncertain innovation process. We do not yet know the final form of our DBA contribution, but we are already accumulating design commitments, conceptual preferences, methods knowledge, information practices and ways of working with technology.

McCarthy et al.’s 2006 complex-adaptive perspective is useful because the doctoral trajectory is not a linear pipeline in which one perfect choice follows another. Feedback from supervisors, peers, literature and evidence can change the next decision. Choices interact. A theoretically elegant question may prove difficult to operationalise. A feasible dataset may force the research problem to narrow. A new source may expose a construct boundary that had previously seemed settled.

Milewski et al.’s 2015 lifecycle perspective adds the importance of timing and intervention space. In the present, many of our choices remain relatively malleable. Once fieldwork begins, some become costly to change. Once a thesis or paper is submitted, the main room for manoeuvre shifts from redesign towards interpretation, qualification and future research. Clock 0 therefore asks us to build the capacity to recognise which decisions are still open and which are becoming locked in.

The simulation is itself a prototype: a low-risk environment in which we can experience design lock-in, methodological pressure and later reconfiguration before those dynamics become consequential in our own research.

Researcher-Development lens

Vitae’s 2025 Researcher Development Framework makes Clock 0 particularly important. At this stage, development is not an incidental benefit of completing the exercise; it is part of the exercise’s purpose.

Several capabilities are visible at once. Development-focused practice means noticing which scholarly skills need strengthening rather than hiding uncertainty behind polished language. Curiosity means keeping credible alternatives alive long enough to test them. Strategic thinking means understanding how today’s design choices affect the claims and contribution that may become possible later. Reflexivity means examining how our assumptions, preferred theories and confidence in particular methods shape what we notice. Responsibility means recognising that claims have consequences for participants, organisations, practice and later scholarship.

On the research side, Clock 0 develops subject knowledge by tracing theory and construct lineage, research methods by asking what a design can actually establish, information and data literacy by maintaining provenance and claim boundaries, digital and technological capabilities by using AI and other tools without surrendering judgement, critical thinking and analysis by testing competing explanations and creativity and innovation by designing better successor questions rather than merely cataloguing weaknesses.

Clock 0 is therefore the hidden developmental purpose of the whole game:

The aim is not to become better at criticising someone else’s paper. It is to become more capable of designing, judging and defending our own.

Before entering Clock One

Carry one question with you:

Which habits of judgement do we want to have in place before our own evidence architecture becomes difficult to change?

Then enter the Supervision Room.

The conference fieldwork that will later feed the published study has not yet happened. There are no significant paths because there is no dataset. There is no mediation result to debate. There is no published conclusion available to tell us what theory apparently "won".

What exists instead is design freedom.

Theory can still change. Constructs can still be redrawn. Measures can still be challenged. Respondents can still be reconsidered. The evidence architecture can still be strengthened.

The question on the agenda

What evidence should this study be designed to generate?

The five decisions tighten progressively. The simulation begins with competing theoretical disciplines, moves into rival perspectives on organisational culture, then commits to a conceptual architecture. Only after that does the questionnaire appear, followed by the evidence design.

The sequence matters because each decision leaves fingerprints on the next. If something is called "culture", what evidence would show that it is genuinely shared? If a bundle of practices is called "Concurrent Engineering", which parts of that construct need to appear in the measure? If the outcome is "product safety performance", whose judgement counts and what would independent evidence look like?

Published papers can make methodology look as though it begins with a sampling paragraph. Clock One makes a different point:

Methodological limitations are often born before the method section is written.

At this clock the Six Drivers are design levers, not retrospective criticisms.

Early theory choices do not activate every Driver. Methodological pressure becomes visible when constructs are operationalised and the evidence architecture is approved.

By the end of Clock One, the key design questions have largely been fixed:

• what kind of relationship a cross-sectional design can establish
• who enters the evidence
• what the questionnaire actually observes
• whether managerial reports can establish shared organisational properties
• what same-source measurement leaves unresolved

The Six Drivers have not judged the study. Their answers are becoming fixed.

Innovation & Technology Management lens

Clock One is the closest analogue to the front end of an innovation process, when multiple architectures are still possible and early choices have disproportionate downstream effects.

McCarthy et al.’s 2006 complex-adaptive perspective is useful here because the research design is not a row of independent boxes. Theory, construct definition, measurement, respondents, evidence and ambition interact. A change in one element can alter what another element means. Choosing an organisational-culture frame changes the constructs that seem salient. Defining those constructs changes what must be measured. The available observation channel may then expose a mismatch between what the theory names and what the study can actually see.

Milewski et al.’s 2015 lifecycle perspective adds the idea of mutual adaptation. In a coherent research design, questions, theory, constructs, measures and evidence need to remain aligned as the study develops. Misalignment at this stage can become path dependence later: once the instrument is fielded, the missing evidence cannot simply be wished back into the dataset.

The Six Drivers therefore function as design levers at Clock One. They make the architecture explicit before it hardens. This is analogous to asking, during innovation development, whether requirements, interfaces and verification plans are coherent before costly commitment.

The Innovation & Technology Management lesson is not “plan everything perfectly”. It is more demanding:

Know which early decisions create later lock-in, and make those decisions with the eventual evidence claim in view.

Researcher-Development lens

Clock One puts a broad cluster of capabilities under pressure because pre-fieldwork design requires integration rather than isolated technical competence.

Subject knowledge matters because theory cannot be selected merely for familiarity; it must illuminate the research problem and make credible mechanisms visible. Research methods matters because the researcher must explain why the chosen design, source and analytical strategy fit the research question rather than simply naming techniques. Information and data literacy appears before data collection in the ability to trace construct origins, distinguish theoretical meaning from operational indicators and anticipate what evidence will later be needed.

Critical thinking and analysis is exercised by comparing plausible frameworks rather than searching for a single obviously correct one. Creativity and innovation appears in the ability to combine or adapt theoretical resources without producing an incoherent hybrid. Reflexivity matters because researchers can become attached to a theory, instrument or method before they have tested whether it fits the phenomenon. Responsibility enters because once participants contribute data, weak design choices are no longer merely private intellectual experiments.

Notice the order in which we instinctively solve the problem. Do we begin with theory? Measurement? feasibility? the eventual claim? No single starting instinct is automatically wrong, but some combinations become incoherent.

Clock One therefore develops a doctoral capability that is easy to underestimate:

the ability to justify not only what we intend to do, but why the evidence architecture can answer the question we say we are asking.

Before entering

If this were our study and not a single response had been collected, what would need to be defensible before asking participants for their data?

---

Eight years later, the intervention space has changed.

The questionnaire has been fielded. Managers have answered it. Items have been retained or removed. A five-factor measurement model has been fitted. Structural relationships have been estimated. The manuscript has a story it wants to tell.

The reviewer cannot redesign the study. The reviewer can still decide what the evidence deserves to become in the scholarly record.

The question on the desk

What claims should this evidence be permitted to support?

This is where the Six Drivers become a methodology stress test.

The question is no longer "Could the study have been designed differently?" It becomes:

Does what was actually done support what the manuscript now wants to say?

The full inquiry uses 30 methodological axes. The simulation surfaces only the load-bearing stress tests. Covered / Not covered / Not addressed now becomes a way of locating the manuscript’s claim ceiling.

The evidence arrives in stages. First comes a respectable-looking measurement model. Then the relationships among the factors begin to look uncomfortably close. Then a mediation pattern appears. Finally, the paper as a whole reaches the editorial desk.

A model can fit. A coefficient can be large. A path can be significant. None of those facts is trivial. None automatically settles the theoretical question printed beside it.

The Six Drivers keep returning the investigation to the same discipline:

What was actually observed, at what level, through which design and what remains another possible explanation?

Innovation & Technology Management lens

Clock Two introduces path dependence and design lock-in. By 2016, the earlier choices have hardened into an empirical configuration. The reviewer inherits the theoretical frame that was chosen, the items that were fielded, the respondents who answered, the observations those items produced and the analytical opportunities available within that record.

In innovation terms, the design has crossed a major commitment point. The product can still be reviewed and its claims reconfigured, but its core architecture cannot be rebuilt without returning to development. The equivalent scholarly distinction is crucial: reviewers can demand narrower interpretation, clearer construct boundaries, more defensible causal language or acknowledgement of unresolved alternatives, but they cannot manufacture longitudinal evidence, independent outcomes or missing construct content after the fact.

The Six Drivers now operate like a methodological design review or stress test. Design asks what relationship the evidence can establish. Sampling asks how far the result can travel. Measurement asks what the numbers actually represent. Unit of analysis asks whether the entity in the claim matches the entity in the data. Analytical technique asks what the statistical machinery adds — and what it cannot add. Rival explanations ask what remains capable of producing the same pattern.

This creates a second Innovation & Technology Management lesson:

When an architecture is fixed, responsible innovation often means reconfiguring the claim rather than pretending the underlying design can still be changed.

Researcher-Development lens

The developmental pressure now shifts from designing evidence to judging evidence proportionately.

Critical thinking and analysis requires separating what the model reports from the theoretical sentence somebody wants to write about it. A significant path, a large coefficient or acceptable fit can be meaningful without settling construct identity or causal mechanism. Research methods becomes the ability to recognise design ceilings — particularly where temporal order, sharedness, level, source or rival explanations are not established.

Information and data literacy appears in numerical and documentary discipline: checking what was actually reported, distinguishing β from (R^2), tracking deleted items and preserving the difference between a methodological vulnerability and a demonstrated bias. Responsibility and research integrity require keeping the public claim proportionate to the evidence even when stronger wording would make the paper more impressive. Communication becomes a research skill because qualification must be expressed clearly enough that readers understand what survives rather than hearing only a catalogue of defects.

Reflexivity becomes especially important because reviewers and authors face opposite temptations. Authors may defend an attractive story because they have invested years in it. Reviewers may over-correct and treat every limitation as fatal. Mature judgement resists both.

Clock Two therefore asks whether we can hold two thoughts at once:

A study can make a real contribution, and the strongest sentence attached to that contribution can still need narrowing.

Before entering

What sentence would be tempting to write — and what sentence would the evidence actually allow us to defend?

---

Ten years after publication, Zhu et al. (2016) has acquired a different kind of power. It is now literature.

A study becomes a finding. A finding becomes a sentence. The sentence acquires a citation. Eventually the citation can circulate without its questionnaire, sampling route, deleted items, alternative explanations or inferential boundaries travelling with it.

The original study cannot be altered. The 2016 publication cannot be renegotiated. The responsibility now is to decide what the paper is entitled to become inside contemporary knowledge.

The question at the frontier

What knowledge should later researchers be permitted to inherit?

The stress test has already located the claim ceiling. Now the Six Drivers become an inheritance check.

A claim that does not survive as stated is not automatically empty. The underlying evidence may need to be narrowed, reframed, retained as a research proposition or rejected.

This distinction is central:

The Six Drivers ask what survives as stated. Clock Three asks what responsible scholarship should do with what remains.

The final three decisions therefore move from citation inheritance, to contemporary interpretation, to successor-study design.

Innovation & Technology Management lens

The last playable clock treats a published paper as one configuration in an evolving knowledge system. Publication stabilises a particular combination of theory, constructs, measures, analytical choices and claims, but it does not freeze the surrounding field.

Later researchers can reproduce that configuration, recombine it with neighbouring theory, challenge its construct boundaries, identify path dependence in the literature or design evidence that makes a previously unresolved distinction observable. Citation itself can create path dependence: once a convenient interpretation is repeated often enough, later researchers may inherit the compressed claim rather than the original evidence architecture.

This makes Clock Three an innovation problem of reconfiguration and successor design. The task is not merely to repair an old paper. It is to decide which components still create value, which assumptions should be retired and what next-generation evidence would discriminate among the serious explanations that remain.

The Six Drivers become an inheritance check. Their limits do not automatically become a shopping list of future methods. A longitudinal design, multi-source data, larger sample or objective outcome is useful only when it resolves a consequential uncertainty. This is the same discipline used in effective innovation portfolios: invest in the next capability because it changes what can be learned or done, not because it is fashionable.

Knowledge, in other words, does not simply accumulate. It is selected, recombined and reconfigured.

Researcher-Development lens

Clock Three brings together capabilities that are central to doctoral contribution.

Information and data literacy means reconstructing what a cited study actually observed before allowing its claims into a literature review. Subject knowledge means understanding enough construct and theory history to notice when the same noun has changed meaning across studies. Critical thinking and analysis requires synthesis across competing explanations rather than counting how many papers support each side.

Creativity and innovation becomes important because future research should do more than repeat a standard limitations paragraph. A strong successor study identifies a live theoretical or methodological distinction and designs evidence capable of discriminating it. Reflexivity asks whether our preferred contemporary explanation is being treated as established merely because it fits our own conceptual commitments. Responsibility concerns inheritance: our literature reviews become part of the chain through which other researchers will later understand the field.

Clock Three therefore moves researcher development from critique to knowledge stewardship:

It is not enough to find more literature. We have to decide what the literature collectively warrants, what uncertainty still matters and what evidence would genuinely move the field.

Before entering

If another researcher opened the paper behind our citation, would they find the claim we said was there?

---

Clock Three is the last playable clock. The simulation now returns us to the clock that was running before play began: Clock 0, the doctoral present.

The three historical clocks ask different forms of scholarly responsibility:

• The supervisor asks, “Is this researchable?”
• The reviewer asks, “Is this claimable?”
• The fellow researcher asks, “Is this inheritable?”

Clock 0 asks the developmental question underneath all three:

What kind of researcher are we becoming when nobody else is in the room to ask those questions for us?

This is why the companion guide should not treat researcher development as a short reflection appended to the game. Clock 0 is the frame around the entire simulation. The playable decisions are rehearsals for capabilities we need before our own research reaches equivalent commitment points.

Vitae’s 2025 Researcher Development Framework is useful here because it gives names to those capabilities without turning them into a scorecard. The developmental aim is independence with accountability: enough subject knowledge to recognise conceptual boundaries, enough methods understanding to know what evidence can support, enough information literacy to preserve provenance, enough criticality to test a compelling story, enough creativity to design a better alternative, enough reflexivity to notice our own preferences and enough responsibility to represent evidence honestly.

The Innovation & Technology Management lens adds another layer. Our DBA itself is a developing knowledge artefact. It will move through periods of exploration, commitment, lock-in, review, reconfiguration and eventual inheritance. The skill is not to eliminate uncertainty before moving. It is to manage uncertainty without disguising it, to preserve optionality where it matters and to recognise when a decision has become consequential.

Clock 0 therefore turns the simulation back onto us:

What would we change in our own current research practice if we knew that a future reviewer — or a future doctoral student — would reconstruct our evidence architecture as carefully as we have reconstructed Zhu et al. (2016)?

The aim is not dependence on an answer key. It is increasing visibility of the reasoning behind our own answers until those habits become portable into our own doctorate.

---

The simulation can be experienced as several lines of music moving at once. They are easier to distinguish after you know what to listen for.

Layer	Listen for...
1. The syndicate brief	Methodology, limitations and future research
2. The paper’s journey	How Zhu et al. move from problem to theory, measurement, evidence and publication
3. Three Clocks	How historical position changes knowledge, agency and responsibility
4. Six Drivers	Where design, sampling, measurement, level, analysis and rivals constrain the claim
5. Innovation & Technology	NPD, systems, lifecycle, path dependence, reconfiguration and decisions under uncertainty
6. Clock 0 — the doctoral journey	Pre-fieldwork approval, peer review and literature inheritance
7. Researcher development and reflexivity	How methodological judgement becomes a habit rather than an answer
8. AI-gamified investigation	Whether AI can support an auditable environment for scholarly judgement rather than merely summarising content

The layers are not eight separate frameworks to memorise. They are different things to notice while one paper moves through the simulation.

---

There is an obvious irony in using AI to build a game about inferential discipline. Fluent prose can make an unsupported claim feel deceptively settled.

For that reason, the investigation keeps an evidential chain of custody:

• DIRECT — focal paper: Zhu et al. (2016) explicitly report it
• DIRECT — external source: another cited source explicitly reports or argues it within its own domain
• INFERRED — scholarly synthesis: the investigation reaches the conclusion by comparing sources, reconstructing constructs or applying methodological reasoning
• RESEARCH PROPOSITION: the explanation is plausible but remains unresolved and empirically testable
• DESIGN — simulation choice: the clocks, personas, adjudication language or other pedagogical machinery created to make the reasoning playable

The Six-Driver labels do something different. Reported fact / methodological inference / vulnerability of unknown magnitude / unresolved rival explanation describe the status of a methodological judgement.

The two systems should not be collapsed:

Chain of custody asks where a statement comes from.
Six-Driver status asks what kind of methodological judgement it is.

The game is allowed to be dramatic. The evidence is not allowed to become more certain because the prose is compelling.

---

Remember the two dimensions.

Three Clocks

2008 — What evidence should this study be designed to generate?

2016 — What claims should this evidence be permitted to support?

2026 — What knowledge should later researchers be permitted to inherit?

Six Drivers

Design · Sampling · Measurement · Unit of analysis · Analytical technique · Rival explanations

There is no score to maximise. There is a judgement record to build.

Pay attention to the option that first feels obvious. Notice when later evidence makes an earlier choice uncomfortable. Notice when a satisfying theory becomes difficult to measure. Notice when a methodological criticism begins to sound relevant to our own doctorate.

That is where the game begins to do its real work.

Enter Clock One.

---

Spoiler boundary: The sections below discuss the investigation’s principal evidential conclusions. Play all three clocks first if the aim is to preserve an uncontaminated initial judgement.

The simulation does not end by declaring Zhu et al. (2016) "right" or "wrong". It asks what survives the combined pressure of time and methodology.

The terminal judgement is deliberately balanced:

The paper survives. Its claims change.

Zhu et al. (2016) produced a substantive empirical study connecting management commitment, product-safety culture, integrated product-development practices, Design-for-Safety practices and product-safety performance. The final quantitative sample comprised 255 managers from 126 firms. The paper reports a five-factor measurement model and a structural model testing eight hypotheses.

The inquiry does not erase that contribution. It changes how confidently some theoretical labels and causal claims can travel.

The central problem is not that every construct is invalid in the same way. The displacement is uneven.

Culture: useful architecture, not a demonstrated causal ladder

Schein’s 1990, 1992 and 2004 three-level model remains a defensible organising architecture. The stronger claim — that the empirical model demonstrates a clean sequence from underlying assumptions through espoused values to artefacts — travels beyond what the measures directly observe.

MCS and PSC: keep the observation channel attached

MCS most securely reflects reported managerial safety governance and priority. PSC most securely reflects a managerial key informant’s view of NPD product-safety-management conditions.

The study therefore does not independently establish genuinely shared organisational culture simply because a factor carries a cultural label.

CE and DFS: practices matter, but labels have boundaries

The CE measure contains genuine integrated-NPD content but under-covers some classical Concurrent Engineering territory, particularly temporal overlap and early involvement (Brown & Eisenhardt, 1995; Gerwin & Barrowman, 2002).

DFS more directly captures reported safety-assurance practice than realised hazard elimination.

PSP: performance through managerial eyes

PSP is based on two managerial ratings — outgoing product audits and customers’ assessment in the market — rather than independently observed realised harm.

The most secure label is therefore managerial dual-referent evaluative product-safety performance.

---

The published measurement model retains 26 of 30 questionnaire items. Reported reliability and fit evidence support statistically coherent factors, but statistical coherence does not by itself establish that a factor exhausts the theoretical meaning of its label (MacKenzie et al., 2011).

Several factors are also highly related. The investigation treats the four legacy AVE/shared-variance flags — MCS — DFS .77, PSC — CE .84, PSC — DFS .86 and CE — DFS .77 — as diagnostic proximity flags, not proof that the constructs are one thing.

Same-source reporting creates a vulnerability, but does not demonstrate that common-method covariance explains the observed relationships (Conway & Lance, 2010; Podsakoff et al., 2003).

The mediation result requires similar restraint. CE→PSP is about .57 before DFS enters. In the full model CE→PSP falls to about .06, CE→DFS is about .21 and DFS→PSP is about .61. The reported indirect effect is about .13. This is a meaningful statistical intermediary pattern, but the cross-sectional design does not independently identify temporal causal mediation (Antonakis et al., 2010; Maxwell & Cole, 2007; Preacher & Hayes, 2008).

A second numerical distinction matters:

DFS→PSP β≈.61 is not 61% explained variance. PSP structural (R^2=.43).

Driver	What the stress test permits
Design	Cross-sectional associations survive; causal direction does not
Sampling	Evidence from reached firms survives; industry representation does not
Measurement	Same-source ratings survive; realised safety and some theoretical labels narrow
Unit of analysis	Manager/unit reports survive; whole-firm sharedness is not established
Analytical technique	The relational pattern survives; 61% variance and strong full-mediation wording do not
Rival explanations	Reverse causality, selection and omitted causes remain live; no structural controls close them

These are six directions of pressure, not six reasons to reject the paper.

In the Six-Driver audit’s eight-part reconstruction of the paper’s largest claim, the two associational components survive most securely as stated. Stronger components concerning causal direction, full mediation, realised safety, whole-firm reach, industry reach and 61% explained variance require narrowing, reframing or withdrawal.

That is where Clock Three becomes essential.

A failed published claim does not automatically erase the evidence underneath it.

---

Later evidence makes the original construct boundaries more interesting rather than simply resolving them.

Work on product-safety culture and safety climate reinforces the importance of management, shared perceptions and system/process content while showing how heterogeneous these constructs can become (Maier et al., 2024; Shea et al., 2021; Suhanyiova et al., 2021). Later Concurrent Engineering work makes mechanisms such as overlap, integration and early involvement more explicit (Rihar et al., 2021).

Placed beside Zhu et al. (2016), the inquiry’s preferred contemporary reading becomes implementation-primary, with product-safety management-system coherence as the strongest serious rival.

This is a scholarly synthesis, not a finding directly reported by the cited studies. Neither interpretation is established as the "true" latent ontology.

That is why the investigation ends with a research design rather than a verdict.

The successor study should define rival explanations clearly enough that evidence can discriminate between them. Construct specification comes first. Source, level, temporal, equivalence and independent-criterion modules should then be added where the consequential claim requires them.

The best next study is not necessarily the largest or most technically elaborate.

It is the smallest coherent design capable of discriminating the claim that matters.

---

The easiest way to leave the simulation is to remember what was "wrong" with Zhu et al. (2016). That would miss the point.

A more useful exit carries both lenses into our own research.

Before fieldwork

Three Clocks: What evidence must this study generate?
Six Drivers: Which design, sample, measures, levels, analyses and rival explanations will determine what can later be claimed?

Before submission

Three Clocks: What is the strongest sentence this evidence can support?
Six Drivers: Which Driver places the tightest ceiling on that sentence?

Before citing another paper

Three Clocks: What knowledge are we choosing to inherit?
Six Drivers: Does the inherited claim still fit the methodology that produced it?

If those questions become reflexive habits, the simulation has travelled beyond the classroom.

---

Antonakis, J., Bendahan, S., Jacquart, P., & Lalive, R. (2010). On making causal claims: A review and recommendations. The Leadership Quarterly, 21(6), 1086 — 1120. https://doi.org/10.1016/j.leaqua.2010.10.010

Brown, S. L., & Eisenhardt, K. M. (1995). Product development: Past research, present findings, and future directions. Academy of Management Review, 20(2), 343 — 378. https://doi.org/10.5465/amr.1995.9507312922

Conway, J. M., & Lance, C. E. (2010). What reviewers should expect from authors regarding common method bias in organizational research. Journal of Business and Psychology, 25(3), 325 — 334. https://doi.org/10.1007/s10869-010-9181-6

Gerwin, D., & Barrowman, N. J. (2002). An evaluation of research on integrated product development. Management Science, 48(7), 938 — 953. https://doi.org/10.1287/mnsc.48.7.938.2818

MacKenzie, S. B., Podsakoff, P. M., & Podsakoff, N. P. (2011). Construct measurement and validation procedures in MIS and behavioral research: Integrating new and existing techniques. MIS Quarterly, 35(2), 293 — 334. https://doi.org/10.2307/23044045

Maier, T. R., Aymans, S. C., & Mühlfelder, M. (2024). Product safety culture as a concept for the development of safety-critical products in regulated industries. Safety Science, 169, 106336. https://doi.org/10.1016/j.ssci.2023.106336

Maxwell, S. E., & Cole, D. A. (2007). Bias in cross-sectional analyses of longitudinal mediation. Psychological Methods, 12(1), 23 — 44. https://doi.org/10.1037/1082-989X.12.1.23

McCarthy, I. P., Tsinopoulos, C., Allen, P., & Rose-Anderssen, C. (2006). New product development as a complex adaptive system of decisions. Journal of Product Innovation Management, 23(5), 437 — 456. https://doi.org/10.1111/j.1540-5885.2006.00215.x

Milewski, S. K., Fernandes, K. J., & Mount, M. P. (2015). Exploring technological process innovation from a lifecycle perspective. International Journal of Operations & Production Management, 35(9), 1312 — 1331. https://doi.org/10.1108/IJOPM-02-2015-0105

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

Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879 — 891. https://doi.org/10.3758/BRM.40.3.879

Rihar, L., Žužek, T., & Kušar, J. (2021). How to successfully introduce concurrent engineering into new product development? Concurrent Engineering, 29(2), 87 — 101. https://doi.org/10.1177/1063293X20967929

Schein, E. H. (1990). Organizational culture. American Psychologist, 45(2), 109 — 119. https://doi.org/10.1037/0003-066X.45.2.109

Schein, E. H. (1992). Organizational culture and leadership (2nd ed.). Jossey-Bass.

Schein, E. H. (2004). Organizational culture and leadership (3rd ed.). Jossey-Bass.

Shea, T., De Cieri, H., Vu, T., & Pettit, T. (2021). How is safety climate measured? A review and evaluation. Safety Science, 143, 105413. https://doi.org/10.1016/j.ssci.2021.105413

Suhanyiova, L., Irwin, A., & Flin, R. (2021). Product safety culture: A preliminary study in the UK manufacturing industry. Journal of Risk Research, 24(8), 1030 — 1048. https://doi.org/10.1080/13669877.2020.1805637

Syndicate exercise requirements. (2026). [Course handout]. Durham University & emlyon business school.

Vitae. (2025). Researcher development framework. Careers Research and Advisory Centre (CRAC) Limited. https://vitae.ac.uk/vitae-researcher-development-framework/

Zhu, A. Y., von Zedtwitz, M., Assimakopoulos, D. G., & Fernandes, K. (2016). The impact of organizational culture on concurrent engineering, design-for-safety, and product safety performance. International Journal of Production Economics, 176, 69 — 81. https://doi.org/10.1016/j.ijpe.2016.03.007