Quantifying Trust Based on the Study "The Influence of Teamwork Quality on Software Team Performance”
The study titled "The Influence of Teamwork Quality on Software Team Performance" was authored by Emily Weimar, Ariadi Nugroho, Joost Visser, Aske Plaat, Martijn Goudbeek, and Alexander P. Schouten. It was published in January 2017 and is available on arXiv. This study, along with multiple others examining the role of trust in teams, has been of significant interest to many members of the Deep Mind team developing algorithms at SelfFusion.
In this article, we briefly discuss methods for quantifying trust based on the study’s findings. More importantly, we demonstrate how SelfFusion has expanded upon this research to conduct a deeper analysis into the structure and essence of trust within teams.
Trust as a Component of Teamwork Quality
The authors of the study conceptualize trust as an essential property of Teamwork Quality (TWQ), describing it as follows:
“Friedlander (1970) found that trust is a key predictor for team performance. Following Mayer et al. (1995), we define trust as ‘the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control the other party.’ Trust is an important supporting mechanism of teamwork. It influences many team processes such as the willingness to share information, give substantial feedback, and manage time correctly (Bandow, 2001). Team members will communicate more openly and will share information more freely when they trust their team members or feel trusted by others. When they feel their contribution is not appreciated, the likelihood of withholding information increases (Boss, 1978; Zand, 1972; Jones & George, 1998; Bandow, 2001). Furthermore, trust has a positive effect on job satisfaction (Driscoll, 1978; Muchinsky, 1977), satisfaction with communication and the perceived accuracy of information shared (Roberts & O’Reilly, 1974), and satisfaction with working within the group (Ward, 1997). Given the importance of trust as a supporting mechanism for teamwork, this factor cannot be eliminated, and we decided to include trust in our TWQ model.”
The study identifies trust as one of six key factors influencing teamwork quality. The other five are:
Communication
Coordination of Expertise
Cohesion
Mutual Support
Shared Values
Study Findings and Conclusion
The authors summarize their conclusions as follows:
“Our goal in this study was to find additional factors that may influence software team performance, extending the previous empirical TWQ study by Hoegl and Gemuenden (2001). We introduced three new TWQ factors: trust, value sharing, and coordination of expertise. The relationship between TWQ and team performance was tested using data from 252 team members and stakeholders. Results showed that teamwork quality is significantly related to team performance, as rated by both team members and stakeholders: TWQ explains 66% of the variance of team performance as rated by team members and 40% as rated by stakeholders. This study shows that trust, shared values, and coordination of expertise are important factors for team leaders to consider in order to achieve high-quality software teamwork.”
These findings are intuitive and align with fundamental principles of teamwork and collaboration. In practical terms, it would be difficult — if not impossible — to argue the opposite: that teamwork quality has no impact on team performance.
Quantifying Trust with SelfFusion Models
At SelfFusion, we have taken a far more rigorous and detailed approach to analyzing trust within teams. While conventional studies recognize trust as a significant factor in teamwork quality, our models break it down further into its essential components, allowing for precise measurement and predictive analysis. Rather than treating trust as a vague or holistic concept, we examine its structure and function through measurable behavioral and cognitive markers.
Key Dimensions of Trust Measurement
Quantifying trust requires identifying and analyzing its core components. In SelfFusion’s approach, trust is deconstructed into the following measurable dimensions:
1. Consistency – The Foundation of Trust Over Time
Trust is not established solely through prolonged exposure to someone’s behavior but through repeated demonstrations of reliability in a variety of situations—including those that may seem irrelevant at first glance. One of the most counterintuitive yet consistently confirmed insights from our research is that extreme openness regarding thoughts and emotions does not necessarily foster trust. Instead, the true predictor of trust is the ability to refrain from deception altogether.
A person who is not excessively open but consistently honest across different contexts builds a stronger foundation of trust than someone who frequently expresses their thoughts but distorts facts when necessary.
Small, seemingly insignificant interactions — keeping promises, showing up on time, fulfilling minor commitments —contribute more to trust-building than dramatic acts of transparency in isolated moments.
2. Behavioral Integrity – Trust as a Function of Action, Not Declaration
Behavioral integrity measures the alignment between what individuals claim to value and how they actually behave. While many models treat behavior as a test of whether an individual adheres to their stated values, SelfFusion reverses this perspective — we assess values exclusively through observed behavior.
Our approach assumes that values are not what people say they believe but what they demonstrate through their actions.
If someone consistently acts in alignment with a particular ethical or professional standard, we infer that as their true value structure, regardless of what they verbally profess.
This is a particularly valuable tool for HR, as self-reported value structures can often be unreliable. Behavioral data, on the other hand, reveals an individual’s true framework for decision-making.
3. Individual Limitations Perception – The Cognitive Gap in Trust Formation
Trust is not only about ethical behavior — it is also about competence perception. In our analyses, a recurring pattern emerges: trust diminishes when the intelligence or experience gap between two team members is too large.
When a more experienced or cognitively superior team member perceives a significant gap in intelligence or competence with another team member, their trust is affected — not due to dishonesty, but due to concerns about predictability.
If a junior or less experienced employee consistently aligns their actions with the value structure of the organization, trust remains intact. However, if they appear unable to make sound decisions due to lack of expertise or experience, even their well-intentioned actions may be seen as unreliable.
This has strong implications for team formation and mentorship programs. It also suggests that trust-building requires not only ethical consistency but also demonstrated competence and ongoing skill development.
4. Benevolence – The Emotional Factor in Trust Perception
This dimension examines how much a team member believes that another person has their best interests at heart. Interestingly, the ability to perceive benevolence is correlated with the emotional disposition of the perceiver.
Our data shows that more emotionally-driven individuals are significantly more sensitive to perceived benevolence, whereas highly analytical team members tend to weigh consistency and integrity more heavily.
This means that within a diverse team, the perception of trustworthiness can vary significantly based on personal cognitive styles, which is crucial when designing leadership and communication strategies.
5. Transparency – The Distinction Between ‘Not Lying’ and ‘Telling the Truth’
Transparency is commonly mistaken for full disclosure, but our research suggests a more nuanced reality.
Baseline trust can be achieved simply by refraining from dishonesty. However, for higher levels of trust to emerge, particularly in high-stakes environments, selective transparency becomes necessary.
In most professional contexts, strategic discretion is tolerated — employees and leaders are not expected to reveal every thought or concern. However, in critical “binary” situations (where a decision has major consequences), simply “not lying” is insufficient. A failure to disclose crucial information in these limit cases can be perceived as deceptive omission.
For example, if an employee is aware of a major risk in a project but chooses not to mention it because they were not directly asked, their decision can still be seen as dishonest — even though they technically never lied. This insight is particularly relevant in leadership training, where trust must be maintained in both everyday operations and high-pressure decision-making moments.
Conclusion: A Data-Driven Approach to Trust
By breaking trust down into these specific dimensions, SelfFusion enables organizations to quantify and predict trust dynamics within teams more accurately than ever before. Our approach ensures that trust is not treated as an abstract or self-reported concept but as a behaviorally verifiable, structured metric.
This analytical framework allows businesses to:
🔹Identify and mitigate trust issues before they escalate.
🔹 Design teams with balanced cognitive and ethical compatibility.
🔹 Develop leadership strategies that account for trust-building at different levels.
🔹 Align workplace culture with proven trust dynamics rather than theoretical assumptions.
In future sections, we will further examine how these metrics influence overall team effectiveness, crisis resilience, and long-term organizational stability.
Methods for Quantifying Trust
Trust is typically assessed using a combination of qualitative and quantitative methods. The most commonly applied approaches include:
1. Surveys & Questionnaires
Trust Scale Surveys – Established psychological instruments such as the Organizational Trust Inventory (OTI)(Mayer & Davis, 1999) or McAllister’s Affect-Based and Cognition-Based Trust Scale (1995).
Likert Scale Questions – Example: “I feel confident that my team members will support me in a crisis.” (1 = Strongly Disagree, 5 = Strongly Agree).
2. Behavioral Analytics & Observational Methods
Response Time to Requests – Measuring how quickly employees respond to each other’s work-related requests.
Error Forgiveness Rate – Evaluating how often team members tolerate and correct each other’s mistakes without placing blame.
Knowledge Sharing Behavior – Analyzing the frequency of voluntary information sharing within teams.
3. Network Analysis (Trust Propagation Models)
Social Network Analysis (SNA) – Mapping trust relationships through interactions within a team or company, such as tracking who seeks advice from whom.
Graph-Based Trust Models – Computational models analyzing trust based on weighted interactions over time.
4. Physiological and Neurological Data (Experimental Approaches)
Oxytocin and Cortisol Levels – Studies such as Kosfeld et al. (2005) have measured oxytocin levels as a biological indicator of trust.
EEG and fMRI Studies – Examining brain activity when individuals are placed in trust-dependent scenarios.
5. AI & Predictive Trust Models
Machine Learning on Communication Patterns – AI models analyzing email sentiment, message tone, and collaborative patterns to estimate trust levels within teams.
BTrust Scores – Decentralized reputation systems that quantify reliability and trustworthiness based on transaction histories, commonly used in financial or gig economy settings.
SIVH Model of SelfFusion
From these standard methodologies, we at SelfFusion have implemented and advanced several approaches to create a more precise trust analysis model. Our hybrid model consists of:
Structured tests and video interviews – These, when used together, yield the most accurate results for quantifying Structured Internal Value Hierarchies (SIVH) in relation to trust.
Behavioral data integration – Trust evaluations are cross-referenced with actual behavioral data and insights from team members who have had sufficient interaction with the analyzed individual.
While we acknowledge the conclusions of the study The Influence of Teamwork Quality on Software Team Performance, our approach diverges categorically when it comes to analyzing and conceptualizing trust.
Specifically, in team-based settings, we do not view "value sharing" and "coordination of expertise" as conceptually equivalent to trust. These factors may contribute to teamwork efficiency, but they do not operate at the same structural level as trust when analyzed through the SelfFusion model.
Trust as a By-Product of SIVH Overlapping
At SelfFusion, trust is not an independent variable — it is a function of SIVH alignment. In other words:
Trust does not exist in a vacuum or emerge randomly — it is the natural by-product of sufficient overlap in Structured Internal Value Hierarchies (SIVH) among team members.
When SIVH alignment is high — especially in the top four core values — trust formation becomes exponentially more likely.
When SIVH alignment is weak or absent, trust formation is unstable, regardless of individual competence or social strategies.
This fundamentally changes how trust should be approached in HR strategies. Instead of attempting to "generate" trust artificially or implementing short-term trust-building exercises, companies must first clarify the internal value structures of key team members.
Hierarchy of Trust: A Structural Perspective
One of the key insights from SelfFusion’s approach is that trust is not a stand-alone goal — it is subordinate to a deeper hierarchical structure.
In our framework, SIVH operates at a "higher hierarchical level" than trust or coordination of expertise.
This means that HR and leadership should not focus on trust creation as an isolated process.
Instead, they must understand that trust, or its absence, is an inevitable outcome of SIVH overlap (or misalignment).
Key Takeaway for HR Leaders
A successful HR strategy should not attempt to "manufacture" trust through surface-level interventions, such as team-building activities or temporary motivational incentives. Instead, it must begin with a structured approach to identifying and aligning the internal value hierarchies of key employees.
By ensuring that team members operate from compatible value structures, trust will naturally emerge, leading to greater team cohesion, long-term resilience, and improved performance.
