Every HR and L&D leader knows the feeling. You run a skills audit, employees tick boxes rating themselves as ‘proficient’ in communication or ‘expert’ in leadership, yet the performance data and on-the-ground reality tell a different story. This is not because your employees are dishonest. It is because, when it comes to accurately evaluating our own capabilities, the human brain is fundamentally flawed.
The inconvenient truth is that self-assessment measures confidence, not competence. And in the workplace, the two are often dangerously mismatched.
The Confidence Trap: Why We Are All Unreliable Narrators
The core of the problem lies in a set of well-documented cognitive biases. The most famous of these is the Dunning-Kruger effect, first described in a seminal 1999 study by psychologists David Dunning and Justin Kruger at Cornell University. [1]
Their research revealed a paradoxical dual burden for low performers: the skills required to perform a task are often the very same skills needed to recognise competence in that task. In other words, people who lack a given capability also lack the metacognitive ability to realise they lack it. As Dunning himself later summarised, the first rule of the Dunning-Kruger club is you do not know you are a member of the Dunning-Kruger club.
This leads to a predictable pattern across any workforce:
- The Unskilled and Unaware: Those with the lowest capabilities are most likely to significantly overestimate their performance.
- The Skilled and Unsure: Conversely, high performers, who are acutely aware of the nuances and complexities of their field, often underestimate their own relative competence.
This is not a niche psychological quirk. It is a widespread phenomenon, and a specific example of a broader bias known as illusory superiority, or the ‘better-than-average’ effect, where a majority of people in any given group rate themselves as above average. Statistically, this is impossible.
Recent workplace data confirms this disconnect is not merely academic. A 2024 analysis by Workera, examining over 22,000 domain assessments, found that a staggering 89% of employees could not accurately assess their own skill level. [2]
- 32% significantly overestimated their skills.
- 56% significantly underestimated their skills.
- Only 11% were accurate.
Crucially, the bias was most pronounced for behavioural and interpersonal skills rather than technical ones. While 55% of users could accurately rate their Machine Learning ability, only 27% could accurately rate their skill in communicating about AI. This is because the criteria for success in human capabilities such as communication, leadership, and adaptability are inherently ambiguous, making objective self-reflection almost impossible.
A 2024 peer-reviewed study published in Swiss Psychology Open reinforced this finding with a striking result: it found a robust negative relationship between self-reported social skills and actual performance on social intelligence tests. [3] People who claimed to be better-than-average at social skills consistently performed worse than those who did not make such a claim. The effect size was substantial, with a Cohen’s d of 0.46 to 0.58 across two independent studies. In short, the more confident someone is about their interpersonal capabilities, the more likely they are to be overestimating them.
The Measurement Mirage in L&D
For L&D professionals, this creates a critical business problem. If the baseline data from your skills gap analysis is unreliable, your entire development strategy is built on a foundation of sand. You risk investing in training for people who do not need it, while the employees with the most significant gaps remain blissfully unaware and under-developed.
This is why so many L&D initiatives struggle to prove their return on investment. The industry has long relied on the Kirkpatrick Model for evaluating training effectiveness, but most organisations rarely progress beyond the first two levels. [4]
- Level 1: Reaction. Did they enjoy the training? (The ‘happy sheet’).
- Level 2: Learning. Did they acquire new knowledge?
- Level 3: Behaviour. Are they applying the skills on the job?
- Level 4: Results. Did their behaviour change impact business outcomes?
Self-assessment is a tempting, low-cost shortcut to attempt to measure Level 3, but as the evidence shows, it is a mirage. It measures perceived confidence, not actual behaviour change. This measurement gap is a significant challenge across the industry. A 2024 LeadX survey found that while most L&D professionals measure satisfaction, only 39% attempt to measure behaviour change, and just 22% measure business impact. [5] LinkedIn’s 2024 Workplace Learning Report noted that only 4% of organisations have reached the measurement phase for their large-scale upskilling projects. [6]
Without accurate measurement of capability, L&D remains a cost centre, perpetually struggling to justify its budget and prove its strategic value to the board. The For HR & L&D Directors page sets out exactly how capability evidence changes that conversation.
The Alternative: From Claimed to Demonstrated Capability
To break this cycle, the focus must shift from what people claim they can do to what they can demonstrate. The most effective way to achieve this is through simulation-based assessment.
Unlike surveys or multiple-choice tests, simulations create realistic, controlled scenarios that require individuals to apply their skills in context. They are not asked if they are a good leader; they are placed in a situation that requires them to lead. The resulting data is a record of decisions made and behaviours exhibited, not a record of self-perception.
This approach provides a fundamentally more reliable form of data because it measures behaviour, not belief. The Human Skills Index methodology is built on exactly this principle: scored decisions, not self-assessment. The contrast with traditional methods is stark:
| Feature | Self-Assessment | Simulation-Based Assessment |
| What it Measures | Confidence and belief | Decisions and demonstrated behaviour |
| Objectivity | Highly subjective, prone to bias | Objective, data-driven scoring |
| Context | Abstract, context-free | Realistic, job-relevant scenarios |
| Feedback Quality | None | Specific, behavioural feedback |
| Predictive Value | Low to negative | High correlation with on-the-job performance |
Sources: Fligby (2024), Simulation-Based Skills Assessment: why they are better [7]; Workera (2024) [2]
By placing employees in immersive scenarios that mirror real-world challenges, such as managing a difficult conversation, making a financial trade-off, or prioritising competing team demands, organisations can observe how they actually perform under pressure. The output is not a vague self-rating, but a rich dataset of decisions, actions, and consequences. Team managers can see this data in real time through the team capability dashboard, identifying strengths and gaps without relying on guesswork.
This is the difference between asking a pilot if they can land a plane and having them prove it in a flight simulator. For critical human capabilities such as leadership, communication, and decision-making, the stakes in the workplace are just as high.
Stop Measuring Confidence. Start Measuring Competence.
It is time for L&D to stop relying on broken tools. To truly understand and close the capability gaps in your organisation, you must measure what people do, not just what they say they can do. The shift from self-assessment to simulation-based assessment is a shift from measuring confidence to measuring competence, and it is the foundation of any L&D strategy that can prove its worth.
The Human Skills Index provides exactly this: a 0-100 score across 8 employer-validated capabilities, derived from demonstrated performance in realistic simulations, not from a tick-box survey. It gives HR and L&D directors the baseline data, progression tracking, and department-level analytics needed to build a genuine business case for capability development. For organisations deploying at scale, the implementation guide covers everything from pilot to full rollout.
Explore the Human Skills Index to see how simulation-based assessment can provide the capability data your organisation needs.
If you are a training provider looking to differentiate your offer with measurable capability evidence, visit the Human Skills Index for Training Providers or explore the partnership models available.
References
[1] Kruger, J., & Dunning, D. (1999). Unskilled and Unaware of It: How Difficulties in Recognizing One’s Own Incompetence Lead to Inflated Self-Assessments. Journal of Personality and Social Psychology, 77(6), 1121-1134. https://psycnet.apa.org/record/1999-15054-002
[2] Workera (2024). 7 out of 10 employees dangerously underestimate or overestimate their skill levels, new analysis finds. https://www.workera.ai/blog/7-out-of-10-employees-dangerously-underestimate-or-overestimate-their-skill-levels-new-analysis-finds
[3] Heck, P. R., Brown, M. I., & Chabris, C. F. (2024). A Robust Negative Relationship Between Self-Reports of Social Skills and Performance Measures of Social Intelligence. Swiss Psychology Open, 4(1). https://swisspsychologyopen.com/articles/10.5334/spo.78
[4] Kirkpatrick Partners (n.d.). The Kirkpatrick Model. https://www.kirkpatrickpartners.com/the-kirkpatrick-model/
[5] Insights (2024). Measuring the effectiveness of leadership effectiveness programmes. https://blog.insights.com/en-gb/blog/how-to-measure-the-impact-of-leadership-effectiveness-programmes
[6] LinkedIn Learning (2024). 2024 Workplace Learning Report. https://learning.linkedin.com/resources/workplace-learning-report-2024
[7] Fligby (2024). Simulation-Based Skills Assessment: why they are better. https://www.fligby.com/using-simulations-in-skills-assessment/
