Understanding the Key Variables that Impact Knowledge Sharing

How ISM Reveals Direct, Indirect, and Hierarchical Relationships

Interpretive Structural Modeling (ISM) maps how selected variables may influence one another. Instead of treating every factor as an isolated predictor, it arranges the factors into levels. This makes hidden dependencies easier to see: one variable may act as a root condition, while another appears closer to the final sharing behaviour.

ISM begins with a clearly defined set of variables and a structural self-interaction matrix. For each pair, the research team asks a focused question: does variable A influence variable B, does B influence A, do both influence each other, or is there no meaningful link? The answers are then coded in a reachability matrix.

• Direct relation: A factor affects another factor without an intermediate step.
• Indirect relation: A factor reaches another through one or more intervening variables.
• Transitive relation: If A affects B and B affects C, ISM examines whether A should also be treated as capable of reaching C.
• Mutual relation: Two factors may reinforce one another, creating a feedback loop.

The resulting matrix is usually processed until it reaches closure. Researchers then compare each variable’s reachable set with its antecedent set. Variables that share the same intersection can be grouped into a level. Removing the highest-level elements and repeating the procedure produces a layered structure. The final diagram is not a statistical effect chart; it is a reasoned map of influence paths.

This distinction matters. A visible top-level variable may be the closest explanation of sharing behaviour, yet it can depend on deeper conditions. For example, a contributor’s decision may sit above perceived usefulness, social confidence, and institutional support. The model therefore guides attention toward the conditions beneath the visible action, not just the action itself.

ISM also supports the classification of variables by driving power and dependence. A high-driving, low-dependence factor can function as a foundational lever. A low-driving, high-dependence factor is more likely to reflect outcomes elsewhere in the structure. Variables with high scores on both dimensions deserve caution: they may be unstable because changes in one part of the system can quickly affect them.

To keep the model credible, researchers should document who judged each relationship, what “influence” meant, and how disagreements were resolved. A small expert panel can produce useful insight, but its conclusions remain interpretive. Agreement statistics, sensitivity checks, and a second assessment round can show whether the hierarchy remains stable when borderline links are removed.

The strongest use of ISM is diagnostic. It can reveal that a popular intervention targets a dependent variable while leaving its root causes untouched. That finding changes the order of action: address the deep driver first, monitor the intermediate mechanism next, and assess the visible sharing outcome last. In plain terms, ISM turns a flat list of factors into a navigable chain of influence.

Key Variables Shaping Effective Knowledge Sharing

Access to Knowledge Sources and Ease of Online Contribution

Access affects knowledge sharing in two separate ways: people need to find useful material, and they need to contribute with little effort. If either step fails, participation drops. A large archive can therefore remain quiet when its content is hard to locate or its contribution process feels like paperwork.

Search quality is more than a search box. Users need stable labels, filters that match their work, current metadata, and results that reveal why an item is relevant. Ranking should favour useful and recent material without hiding older records that still carry essential expertise. Clear ownership and revision dates also help users decide whether a source deserves trust.

• Findability: Use plain titles, meaningful keywords, subject tags, and filters for date, role, project, or content type.
• Access clarity: Show who can view, edit, reuse, or comment before a person starts the task.
• Content context: Attach purpose, scope, assumptions, and examples to each knowledge item.
• Result quality: Measure successful searches, not only clicks or page views.
• Maintenance: Mark stale material and route it to a named reviewer.

The contribution path should match the size of the knowledge unit. A quick correction may need only one field and a submit button. A technical lesson may require evidence, tags, version details, and review. Forcing both into the same form creates friction at one end. In practice, progressive disclosure works well: ask for the minimum first, then offer optional fields when they add value.

Small barriers accumulate. Mandatory attachments, repeated logins, unclear file limits, slow page loads, and poor mobile layouts can turn a willing contributor into a silent observer. A useful audit records the time from opening the form to successful submission, the number of failed attempts, and the abandonment rate at each step.

Access also includes inclusive design. Keyboard navigation, readable contrast, captions for recorded material, descriptive link text, and support for assistive technologies widen participation. The Web Content Accessibility Guidelines 2.2 provide a practical benchmark. Accessibility is not decorative; it changes who can contribute and whose expertise enters the shared record.

Managers should track the full path from discovery to reuse. Useful measures include search success rate, median time to locate an applicable item, contribution completion rate, review delay, and the share of contributions that receive a later view or reuse. One metric alone can mislead: a high posting count may hide low-quality or duplicate material.

The strongest design treats access as a workflow rather than a doorway. Someone finds a source, judges its relevance, adapts it, submits an addition, and later sees how that addition is used. Improving each handoff makes online knowledge sharing feel less like an administrative chore and more like a sensible part of the workday.

Trust, Reciprocity, and Willingness to Share Knowledge

Trust lowers the perceived risk of sharing. People are more willing to post an answer, disclose a work method, or admit uncertainty when they believe others will treat that contribution fairly. Without this confidence, valuable knowledge may remain private, especially when sharing could affect status, job security, or professional credibility.

Trust has several distinct targets. A contributor may trust the other members, the organisation behind the community, the moderation process, or the rules for handling stored content. These forms should not be merged into one vague score. Someone may trust colleagues but distrust managers, or trust the platform while doubting the motives of its users.

• Competence trust: Members believe that others can understand and use the contribution correctly.
• Integrity trust: Members expect honest use, accurate attribution, and no deliberate distortion.
• Procedural trust: Members expect complaints, disputes, and misuse reports to receive a fair response.
• Identity trust: Members can judge who is speaking and whether the person is accountable for the claim.

Reciprocity adds a practical calculation. People often ask, consciously or not, “Will my effort help me or someone else later?” This does not always mean direct exchange. A contributor may share because past help created a sense of mutual obligation, because future assistance seems likely, or because helping strengthens the group’s shared norm.

Reciprocity becomes fragile when benefits are uneven. A small group may carry most of the answering burden while many users only consume. Contribution records, rotating expert duties, peer acknowledgement, and visible follow-up can make exchange more balanced. Rewards should support useful behaviour, not encourage a flood of shallow posts.

Recognition works best when it is specific. A short note explaining how a contribution solved a problem can be more credible than a generic badge. Credit should also follow the knowledge itself. If an idea is copied without attribution, trust declines quickly; once that happens, symbolic incentives rarely repair the damage.

Leaders can strengthen willingness by setting clear social rules:

• Permit questions and corrections without ridicule.
• Separate good-faith mistakes from careless or harmful conduct.
• Explain how authorship and reuse are recorded.
• Respond to misuse reports within a stated time.
• Show concrete examples of contributions that produced useful results.

Measurement should distinguish willingness from actual behaviour. A survey may capture intentions, while contribution logs show what people do under real conditions. Compare both, and examine whether trust predicts first-time participation, repeated contribution, response quality, or help offered to newcomers. These outcomes are related, but they are not interchangeable.

Trust also changes over time. A single breach can have a sharp effect, while consistent fair treatment builds confidence slowly. For that reason, managers should monitor complaint patterns, ignored questions, uncredited reuse, and the return rate of contributors. The quiet signals often matter most: when experienced members stop posting, the community may look active while its deeper knowledge is draining away.

Communication Quality and Shared Meaning Across Domains

Shared meaning is a practical condition for useful knowledge exchange. People may use the same word while assigning it different limits, risks, or outcomes. In cross-functional work, terms such as “approved,” “complete,” or “urgent” can therefore create disagreement even when everyone communicates in good faith.

Communication quality depends on more than speed or message volume. A strong contribution states its purpose, audience, evidence, assumptions, and required action. It also separates facts from interpretation. This structure gives readers a better chance of applying the knowledge correctly in a new setting.

Domain boundaries make the problem sharper. An engineer, teacher, manager, and software specialist may describe one issue through different concepts. Their vocabulary reflects different goals. Translation alone may not solve the gap because a direct equivalent may not exist. The missing element is often a shared frame: what the term covers, what it excludes, and when it should be used.

• Define critical terms: Add a short meaning, scope, example, and excluded case.
• State the context: Identify the project stage, audience, location, version, or business rule.
• Use controlled alternatives: Provide preferred terms and list common synonyms or abbreviations.
• Show relationships: Explain whether a concept is a cause, result, part, exception, or measure.
• Invite semantic checks: Ask a subject expert from another domain to restate the message in their own words.

A glossary is useful, but a static list is not enough. Terms change meaning as processes, regulations, and roles change. A living vocabulary should record an owner, approval date, version, and link to the relevant rule or example. This turns terminology into an operational asset rather than a dusty reference page.

Structured templates can improve message quality without making every contribution stiff. A short incident template might ask: “What happened?”, “What was affected?”, “What evidence exists?”, and “What should happen next?” A lesson template could add conditions, limitations, and transfer advice. These prompts reduce ambiguity while leaving room for professional judgment.

Meaning should be tested at the point of use. Ask recipients to classify a case, complete a small task, or explain the decision that follows from the shared information. If two competent readers reach different conclusions, the communication needs refinement. Read receipts and message counts cannot reveal this semantic failure.

Language access matters too. Plain wording, translated key terms, captions, and examples from more than one domain can widen understanding. Automated translation may accelerate drafts, but specialised terms, legal wording, and safety instructions still need qualified review. A fluent sentence can remain conceptually wrong—smooth words are not proof of shared meaning.

The useful outcome is not identical language. It is aligned interpretation: different specialists can retain their own vocabulary while reaching the same decision, action, or data meaning. That is the communication variable worth tracking in knowledge-sharing research.

Technology Usability, Search, and Social Interaction

Technology usability shapes whether people can turn an intention to share into a completed action. A system may offer advanced functions yet still discourage participation when controls are unclear, responses are slow, or users cannot recover from mistakes. The key variable is not the number of features. It is the effort required to complete a meaningful knowledge task.

Usability has several measurable dimensions:

• Learnability: New users can understand the main actions without lengthy training.
• Efficiency: Experienced users can complete routine tasks with few steps.
• Error recovery: Users can undo, restore, or correct an action without losing work.
• Consistency: Similar actions behave in similar ways across pages and devices.
• Responsiveness: The system gives clear feedback when a request is processing, saved, or rejected.

Search technology also shapes what users notice. A useful system should handle spelling variants, abbreviations, natural-language questions, and related concepts. Ranking can combine text relevance with signals such as freshness, expert review, usage, and context. Yet ranking signals need care: popularity may favour familiar material, while specialist content remains invisible. Search logs should therefore be examined for reformulated queries, empty-result sessions, rapid returns, and repeated failed attempts.

Personalised results can reduce cognitive load, but excessive personalisation creates a narrow view of the knowledge base. Users may repeatedly see material similar to what they already know and miss a useful alternative. A balanced design offers both relevant recommendations and a clear route to broader discovery. Explainable filters help users understand why a result appeared and how to change the selection.

Social functions add a different layer. Profiles, replies, mentions, reactions, version discussions, and shared annotations help contributors connect knowledge to people and situations. These features should support the work, not become a popularity contest. Reaction counts alone say little about accuracy or reuse, and public rankings may discourage cautious experts from contributing.

Moderation design is especially important in social knowledge spaces. Clear reporting paths, edit histories, rate limits, and role-based permissions can reduce spam, harassment, and deliberate manipulation. A lightweight reputation signal may help readers assess contributions, but it should not replace evidence, review, or context.

Researchers can assess the technology variable through controlled usability tasks and real usage data. Useful indicators include task completion time, error frequency, successful searches, return visits after a failed search, contribution edits, and use of discussion features. Compare results across novice and experienced users; averages can hide a steep learning curve.

Technology improves sharing when it reduces needless effort while preserving discovery, discussion, and accountability. The winning formula is rarely “more features.” It is a quiet interface, strong feedback, flexible search, and social functions that make expertise easier to connect with the right problem.

Communities of Practice and the Exchange of Tacit Knowledge

Communities of practice turn experience into a social learning process. Their value lies not in storing answers alone, but in helping members explain how they reached a decision, which signals they noticed, and where a familiar method may fail. This is the kind of knowledge that is difficult to capture in a checklist.

Tacit knowledge often appears in stories, demonstrations, comparisons, and small corrections. An experienced worker may sense that a plan is unrealistic before a formal metric shows why. In a community, others can question that judgment, test its limits, and gradually learn the pattern behind it.

Effective communities usually develop three forms of participation:

• Peripheral participation: New members observe discussions and learn local practices before contributing.
• Active participation: Members ask questions, share cases, and compare approaches.
• Stewardship: Experienced members guide discussions, connect people, and protect the group’s learning purpose.

Online settings need deliberate substitutes for the cues found in face-to-face work. A photo sequence, annotated screen recording, case timeline, or short audio explanation can reveal details that a polished final document hides. Follow-up questions are equally important. They expose the reasoning behind an action, not just the action itself.

Capture should happen close to the event. A useful workflow invites the contributor to record four elements:

• What situation occurred?
• What clues shaped the decision?
• What action was taken, and why?
• When would this advice not apply?

This format preserves nuance without demanding a long report. A curator can later extract a concise lesson, link it to a case, and retain the original conversation for context. The aim is not to flatten experience into rigid rules. It is to make expert judgment easier to inspect and adapt.

Community health depends on its boundary conditions. If membership is too open, discussions may lose focus. If it is too closed, newcomers cannot refresh old assumptions. A clear purpose, rotating facilitation, and periodic review of unresolved questions help maintain a useful balance.

Leaders should assess more than activity. Strong indicators include the number of new members who become contributors, the diversity of roles in a discussion, the reuse of lessons in later cases, and the survival of knowledge when an expert leaves. A busy conversation can still produce little learning; a quiet thread may contain a crucial insight.

Communities of practice work best when recognition follows learning value rather than posting volume. Thank contributors for revealing a difficult judgment, testing an assumption, or explaining a failure without blame. That small signal tells members that thoughtful experience—not merely polished certainty—is worth sharing.

BIM, Visual Context, and Project-Based Knowledge Sharing

Building Information Modeling (BIM) changes the setting in which project knowledge is shared. Instead of discussing an isolated file, participants can attach a note, decision, or lesson to a specific object, location, system, or construction phase. This spatial link gives the contribution a precise point of reference.

The key variable is context density. A useful BIM-based contribution can connect several layers at once:

• Object: The wall, beam, door, pipe, or system concerned.
• Location: The floor, zone, room, or site position.
• Time: The design stage, installation sequence, inspection date, or revision.
• Responsibility: The party that created, checked, approved, or must act on the information.
• Evidence: A drawing, photograph, specification, inspection record, or field observation.

This structure reduces the ambiguity of ordinary messages. “The opening is wrong” is weak project knowledge. “The east stairwell opening on level three conflicts with the fire-door clearance shown in revision 12” is far more actionable. The second statement can be checked, assigned, and resolved because its meaning is anchored to model context.

Visual context also helps users understand relationships that are hard to express in words. A clipped 3D view can show interference between services. A sequence view can reveal why an installation step must happen earlier. A linked photograph can show the difference between the planned condition and the condition found on site. These views support faster diagnosis, but they must retain scale, orientation, date, and revision details.

Interoperability remains a decisive constraint. Project participants may use different authoring tools, file formats, classification systems, or naming rules. The ISO 19650 series offers a widely used framework for managing information across the asset life cycle, but a standard does not remove every exchange problem. Model versions, object identifiers, permissions, and responsibility for updates still need explicit control.

A BIM knowledge item should therefore carry a small evidence trail:

• the model version and authoring date;
• the affected object or coordinate;
• the issue, decision, or lesson;
• the supporting source;
• the responsible reviewer;
• the status and closing date.

Visual richness can mislead. A detailed model may look authoritative while hiding missing attributes, outdated geometry, or uncertain assumptions. Users should be able to distinguish design intent, verified site data, temporary information, and unresolved interpretation. A simple status label often prevents a polished image from becoming false certainty.

Project teams can measure BIM-supported sharing through issue resolution time, repeated coordination errors, reuse of approved details, linked contributions per project phase, and the percentage of decisions connected to a model object. These measures show whether visual context improves work, rather than merely increasing the number of uploaded files.

The strongest approach treats the model as a navigable knowledge map, not as a decorative 3D picture. When decisions, evidence, responsibilities, and lessons remain linked to the right project location and revision, knowledge can travel from design to construction and later maintenance with much less loss.

From Tacit Experience to Structured and Reusable Knowledge

Turning experience into reusable knowledge requires more than recording what happened. The essential step is to preserve the conditions, reasoning, limits, and evidence that make an action understandable outside its original moment. Without those elements, a polished document may be easy to read yet unsafe to apply.

A practical conversion process separates observation from interpretation. First, record the event in concrete terms. Next, explain the decision and identify the signal that shaped it. Then test whether the lesson holds in another case. This sequence helps prevent one person’s habit from being mistaken for a universal rule.

• Capture: Record the situation, action, result, and relevant evidence.
• Explain: Add the reasoning, assumptions, and trade-offs behind the decision.
• Qualify: State exceptions, risks, and conditions where the lesson may fail.
• Abstract: Extract a principle that can guide similar cases.
• Validate: Ask another competent person to apply the lesson to a new example.
• Package: Create a short reusable unit with a title, purpose, scope, owner, and review date.

The right unit is often smaller than a manual and richer than a slogan. A reusable knowledge card might contain a problem pattern, recommended response, warning signs, and one counterexample. Short formats reduce reading time, while linked source material allows users to inspect the deeper reasoning when needed.

Classification affects reuse. A lesson should be tagged by task, process stage, risk, asset, role, and level of confidence where relevant. Tags must describe how people will look for the lesson later, not merely how the original author thought about it. Otherwise, valuable material becomes technically stored but practically invisible.

Quality control should test transfer, not just grammar. Give a new user a realistic case and ask whether the recorded lesson supports a sound decision. Reviewers can score completeness, clarity, evidence, transferability, and known limitations. A simple four-level scale—draft, reviewed, tested, retired—helps prevent untested advice from appearing authoritative.

Reuse also depends on controlled change. When a rule, process, or design changes, linked lessons need a status review. Keep the original version for traceability, but mark whether it remains valid, needs revision, or is withdrawn. This protects users from applying yesterday’s solution to today’s conditions.

Knowledge owners should not become permanent bottlenecks. Assign stewardship for accuracy, while allowing qualified users to suggest edits and add contrasting cases. Two lessons may appear to conflict because they apply under different conditions; linking those conditions is often more valuable than deleting one of them.

Success can be measured through repeat application, reduced time to solve comparable problems, fewer avoidable errors, and the number of lessons updated after use. Storage volume is a weak signal. A smaller collection that changes decisions is far more valuable than a vast archive no one can trust.

Practical Priorities for Online Learning and Knowledge Systems

Practical priorities should follow the influence structure of the knowledge-sharing system. Start with the conditions that shape participation, then improve the learning process, and only afterward judge the final volume of shared material. This order prevents teams from treating visible activity as the main goal.

• Set a measurable learning target: Define the decision, skill, or task that online sharing should improve.
• Assign intervention owners: Give each priority to a named role with authority, time, and a clear success measure.
• Use a staged rollout: Test the approach with one course, team, or project before expanding it.
• Protect learning time: Reserve time for contribution, review, reflection, and application.
• Measure outcomes: Track improved decisions, completed tasks, error reduction, and knowledge reuse rather than post counts alone.

Online learning systems should connect contribution with a real learning cycle. A participant shares a case, receives a response, applies the lesson, and records what happened next. This feedback loop reveals whether knowledge changed behaviour. It also gives educators and managers a stronger basis for improving course content, assignments, and support.

Evaluation should combine several levels. Immediate reactions show whether users understood the activity. Short tests can examine recall and interpretation. Later observations reveal whether learners apply the knowledge in practice. A useful assessment plan may review progress after one week, one month, and one work cycle. Timing matters because knowledge can appear successful at first and then fade.

Different groups need different success measures. In education, useful indicators may include completion of peer tasks, quality of explanations, and transfer to a new problem. In project work, teams may examine fewer repeated mistakes, faster handovers, and better decision records. A single universal score would hide these differences.

Governance should be light but visible. Define who may create learning content, who can approve it, how conflicts are handled, and when material must be reviewed. Keep a decision record for major changes. This creates accountability without turning every contribution into a formal publication.

Accessibility and inclusion should be tested through actual participation. Compare completion and contribution patterns across language groups, devices, roles, seniority levels, and working arrangements. If one group contributes far less, investigate the learning design, workload, permissions, or social climate before assuming a lack of interest.

Responsible data use is another priority. Collect only information needed for learning or system improvement, explain retention periods, and separate developmental feedback from disciplinary assessment where possible. Under the EU General Data Protection Regulation, identifiable learning records require a lawful basis and transparent handling. Automated profiling may also raise obligations under the EU Artificial Intelligence Act, depending on its purpose and risk level.

The most reliable improvement cycle is simple: define the learning result, select a small intervention, observe use, review evidence, and adjust. Online knowledge sharing succeeds when it improves what people can understand, decide, and do. Activity is only the signal; changed practice is the result.

Fazit: Focus on Root Variables Before Improving Online Sharing

The central lesson is simple: improve the conditions beneath sharing before chasing higher activity. A larger number of posts does not prove that a knowledge system works. The stronger test is whether people can make better decisions, solve comparable problems, and learn from one another with less delay.

The integrated logit–ISM perspective supports a disciplined conclusion. Statistical analysis can identify variables linked with online sharing, while structural analysis can show which factors may sit beneath others. Together, these views help decision-makers avoid a common mistake: treating the most visible behaviour as the deepest cause.

Use the findings as a sequence of questions:

• Which variables show a reliable relationship with the target behaviour?
• Which of those variables appear to drive several others?
• Which links are supported by data, and which remain expert interpretations?
• What change can be tested without confusing correlation with causation?
• How will improved knowledge use be measured after implementation?

A careful conclusion should also state its limits. Results from one population, platform, sector, or time period may not transfer directly to another. Self-reported willingness can differ from observed contribution. Expert judgments in a structural model can expose useful patterns, yet they do not replace longitudinal evidence or controlled evaluation.

The most valuable next step is a focused intervention on a root variable, followed by measurement of the intermediate changes and the final sharing outcome. If the result is weak, revise the assumed pathway rather than simply adding more features or incentives. This approach turns the hierarchy into a practical learning cycle.

For researchers, the priority is transparent reporting: define each construct, explain the sampling process, publish effect estimates and uncertainty, document structural judgments, and test whether the hierarchy remains stable under alternative assumptions. For practitioners, the priority is narrower but equally important: fund the condition that unlocks several downstream improvements.

Online knowledge sharing is not a single action. It is a system of motives, relationships, meanings, processes, and conditions. Once the root variables are understood, improvement becomes more precise—and far less dependent on guesswork.

Useful links on the topic

• the role of daily knowledge sharing expectations, learning goal ...
• Knowledge Sharing: Two-Dimensional Motivation Perspective and ...
• Can any one provide me with articles that have knowledge sharing ...