Scientific Foundations of PRODICTA

Work sample simulations

PRODICTA replaces CV screening and theoretical assessment with realistic job simulations.

In comparison:

• Years of experience has low predictive value, around 0.18
• Education level has low predictive value, around 0.10

Traditional hiring relies on proxies:

• CVs
• Qualifications
• Self-reported answers

Work simulations measure something fundamentally different: Observed performance in job-relevant situations.

PRODICTA places candidates into Day 1 scenarios where they must:

• Prioritise tasks
• Make decisions under pressure
• Respond to real constraints
• Communicate in context

Not what they say they would do, but what they actually do.

Behaviour depends on context

Research shows behaviour is shaped by the situation.

The same individual may behave differently depending on:

• Pressure
• Competing priorities
• Stakeholder expectations
• Organisational environment

Assessing candidates outside of real context limits accuracy. To understand performance, behaviour must be observed in realistic conditions.

From traits to behaviour

Research shows that while personality traits describe general tendencies, observed behaviour in real situations provides a stronger and more reliable signal for predictive analysis.

Personality-performance research demonstrates moderate correlations between traits and job outcomes (Barrick and Mount, 1991), while behavioural science shows that performance is highly dependent on situational context (Mischel, 1968).

Meta-analyses of hiring methods further show that approaches based on real job tasks, such as work sample tests, are among the most predictive indicators of performance (Schmidt and Hunter, 1998; 2016).

Personality provides useful insight, but it does not capture how someone behaves when it matters. Observed behaviour in context provides a stronger, more reliable signal.

Assessment centre methodology

PRODICTA reflects principles used in assessment centres, a long-established approach in high-stakes hiring.

Assessment centres evaluate candidates through:

• Simulations
• Role-based exercises
• Real-world tasks

Assessment centres are effective but traditionally:

• Expensive
• Time-intensive
• Difficult to scale

PRODICTA brings the same principles into a scalable, digital format:

• On-demand simulations
• Role-specific scenarios
• Consistent evaluation

Experiential performance

Research in experiential learning shows that performance is best understood through action.

Kolb (1984)

People demonstrate capability most accurately when:

• Actively performing tasks
• Making decisions in context

PRODICTA captures behaviour as it naturally occurs while candidates reason through real workplace scenarios.

Decision-making under pressure

Real-world performance depends on how individuals operate under constraint.

Decision science shows that behaviour is influenced by:

• Time pressure
• Incomplete information
• Cognitive load

Kahneman (2011); Sweller (Cognitive Load Theory)

Traditional tests do not reflect real working conditions.

PRODICTA introduces:

• Time constraints
• Competing priorities
• Imperfect information

To reveal true decision-making behaviour.

Behaviour vs intention

Behavioural science shows that individuals do not always act in line with their self-perception.

Kahneman and Tversky

This is why personality tests, behavioural questionnaires, and self-assessment surveys have limited predictive value. People answer how they imagine themselves, not how they actually behave.

Self-reported answers:

• Can be biased
• May not reflect real behaviour

Research shows that hiring systems based on inferred traits or self-reported inputs:

• Can lack consistency
• May not reflect real-world behaviour
• Depend heavily on subjective interpretation

Understanding how someone behaves in theory is not the same as seeing how they perform in practice. This is the limitation of personality assessments, behavioural questionnaires, and trait inventories: they tell you what someone is like, not what they will do.

PRODICTA measures actual decisions, not stated intentions.

AI-driven behavioural analysis

By ‘behavioural analysis’ we do not mean personality questionnaires, behavioural surveys, or self-reported preference tests. PRODICTA analyses observed behaviour as candidates reason through real workplace scenarios, what they actually do, decide, write, and prioritise when faced with realistic job pressure. The difference is fundamental: self-reported behavioural data tells you what someone says they would do, observed behavioural data shows you what they actually did.

PRODICTA combines behavioural science with modern AI to interpret candidate performance.

Research in AI and HR analytics shows:

• Data-driven hiring improves consistency and decision quality
• Behavioural data provides stronger signals than self-reported input
• Structured evaluation reduces bias and variability

Supported by:

• IBM Smarter Workforce Institute
• McKinsey research
• Frontiers in Psychology (2022)

AI is most effective when applied to observed behavioural data captured in real work conditions, rather than relying on abstract or self-reported questionnaire responses. The distinction matters: self-reported answers can be biased, rehearsed, or AI-generated. Observed behaviour during actual work tasks cannot be faked the same way.

PRODICTA analyses what candidates actually do as they respond to real workplace scenarios:

• Decision patterns when facing trade-offs
• Prioritisation behaviour under time pressure
• Risk awareness when stakes are visible
• Consistency under pressure across multiple tasks
• Communication style in real workplace situations

This is observed performance, not self-reported preference. The data comes from how candidates actually behave during the simulation, not from questions about how they would behave.

AI, data, and modern hiring

Research shows that:

• AI improves hiring speed and efficiency
• Predictive analytics improves workforce outcomes
• Data-driven organisations make better talent decisions

AI is only as powerful as the data it is trained on.

Traditional systems rely on:

• CV data
• Keyword matching
• Static inputs

PRODICTA uses observed behavioural data from realistic job scenarios.

• Deeper pattern recognition
• Early identification of risk
• Stronger prediction of performance and retention
• More consistent and defensible decisions

The commercial reality

Hiring accuracy has direct financial impact.

CIPD reports:

• Average cost per hire around £6,000
• Manager-level hires around £19,000

REC benchmarks indicate:

• Typical bad hire cost between £30,000 and £50,000
• Mid-management failures can exceed £100,000

Even small improvements in hiring accuracy:

• Reduce cost
• Protect revenue
• Improve retention
• Strengthen client relationships

Self-perception vs performance

Research shows that self-reported ability does not reliably reflect real performance. In some studies, individuals who rated themselves most highly performed worse in practical assessments.

• Confidence does not equal competence
• CV claims do not guarantee performance
• Interviews capture perception, not behaviour

PRODICTA measures observed performance, not self-perception. The candidate who claims to be ‘great under pressure’ gets the same simulation as everyone else, and we measure how they actually respond.

Realism and ecological validity

Research in psychology shows that assessments which reflect real-world conditions provide more accurate insights. This concept, known as ecological validity, is one of the strongest principles in modern assessment science.

The closer an assessment is to real work:

• The more accurate the prediction
• The more meaningful the results
• The higher the trust from employers, HR teams, and hiring managers

Generic personality tests, situational judgement questionnaires, and abstract problem-solving exercises score low on ecological validity because they don’t resemble real work.

PRODICTA creates realistic, role-specific environments that mirror real decisions, real constraints, and real job conditions. A Marketing Manager candidate works on actual campaign briefs. A Finance Manager reviews actual variance data. A Solicitor advises on actual contract clauses. The simulation environment matches the work environment, high ecological validity, high predictive power.