Can We Really Predict Job Performance? What Science Tells Us

An impressive CV, fifteen years of experience, a convincing interview. On paper, the candidate ticks every box. Three months later, the collaboration falls apart. Not because they lack expertise, but because something simply isn't working in the way they interact with others.

It's a scenario many recruiters and managers know all too well. And it raises an important question: can we really anticipate whether a professional relationship will succeed before it even begins?

Organizational psychology has been exploring this question for decades. Its findings challenge some deeply held assumptions: technical expertise alone does not explain performance, recruiters' intuition is far from infallible, and behavioral alignment plays a critical role in successful professional relationships.

Can We Really Predict Job Performance?

Yes, but first we need to understand what we're actually trying to predict.

For years, organizations have evaluated professional performance primarily through a technical lens: knowledge, experience, qualifications, and the ability to meet objectives. Yet being highly skilled at a job doesn't necessarily mean performing equally well in every situation.

Research by Campbell et al. (1993), further developed by Campbell and Wiernik (2015), demonstrates that job performance is multidimensional. It also depends on an individual's ability to collaborate, adapt, take initiative, and operate effectively within a team.

This distinction matters. An outstanding expert can deliver flawless work while disrupting an entire team. Conversely, a less experienced professional may contribute more to a project's overall success through their ability to bring people together, communicate effectively, and adapt to changing circumstances.

Research by Koopmans et al. (2014), involving 1,181 workers, identifies several dimensions of job performance, including task performance, contextual performance, adaptive performance, and counterproductive work behavior.

In the performance framework presented in the Migbirds research study, the latter three dimensions collectively account for approximately 63% of the dimensions considered. This does not mean that soft skills statistically explain 63% of professional success. Rather, it highlights how central behavioral factors are to the way performance is understood and assessed.

The real question is no longer simply whether someone can do the job, but under what conditions they will be able to perform at their best.

What Does a Job Interview Actually Predict?

An interview can help assess a candidate. But can it really tell us how that person will behave once they're on the job?

The answer largely depends on how the interview is conducted.

Research by Schmidt and Hunter (1998), subsequently reassessed by Sackett et al. (2022), shows that recruitment methods vary considerably in their predictive validity. Structured interviews rank among the strongest methods, with an estimated correlation of 0.42 with subsequent job performance.

By contrast, unstructured interviews, those relatively informal conversations in which recruiters rely heavily on experience and instinct, demonstrate substantially lower predictive validity, estimated at around 0.19 in the revised findings.

Why such a difference?

Because a conventional interview often measures a candidate's ability to present themselves just as much as their ability to succeed in a professional environment. Some people are exceptionally good at describing achievements, highlighting their experience, and giving the answers interviewers want to hear. But that tells us surprisingly little about how they will respond to disagreement, uncertainty, or a team under pressure.

Structured interviews help reduce some of these biases by asking every candidate the same questions, using predefined evaluation criteria, exploring specific professional situations, and applying consistent scoring methods.

Yet even structured interviews have an important limitation. However well designed, an interview remains a snapshot of an individual in a situation that cannot fully replicate real working conditions.

This is one of the central challenges identified in the first part of the Migbirds scientific study: moving beyond self-reported qualities to better understand the behavioral dynamics that emerge within a specific professional context.

Which Methods Best Predict Job Performance?

Organizational psychology has extensively compared the main methods used to select candidates. The updated meta-analysis published by Sackett et al. (2022) notably revised several historical estimates of predictive validity downward.

Sources: Schmidt & Hunter (1998); Sackett et al. (2022, 2023). Estimated correlations with job performance, depending on the methods and statistical corrections applied.

These findings point to an initial conclusion: no single method can perfectly predict professional success.

But they also reveal a deeper limitation. Most selection tools attempt to evaluate candidates' characteristics independently of the environments in which they will be working.

Yet research by Kristof-Brown, Zimmerman, and Johnson (2005), covering more than 200 studies, demonstrates that the fit between individuals and their professional environments is associated with important outcomes, including job satisfaction, commitment, and intentions to leave an organization.

In other words, identifying a professional's strengths is not enough. We also need to understand how those strengths will translate into performance when interacting with a particular team, manager, organizational culture, or type of mission.

This is where context-based behavioral assessment offers a particularly promising direction. Not as a replacement for existing selection tools, but as a way to complement them with relational and organizational insights.

Combining several sources of information can improve decision-making, provided those inputs are relevant, properly measured, and appropriately weighted. Research by Kuncel et al. (2013) notably demonstrates the advantages of mechanical data combination over relying solely on overall human judgment.

Why Are So Many Recruitment Failures Behavioral?

It would be misleading to claim that every failed hire comes down to behavioral factors. However, research clearly shows that human dynamics are a major source of organizational difficulties.

A study by Azim et al. (2010), involving 47 project management professionals in the aerospace industry, offers a particularly revealing insight: 74.5% of respondents identified human and behavioral issues among the main drivers of project complexity, far ahead of challenges related to processes or products.

This reflects something managers experience regularly: the hardest problems to solve are not always technical.

A consultant may have outstanding expertise yet struggle to work with a demanding executive. A freelancer accustomed to considerable independence may feel constrained in a highly hierarchical organization. A particularly direct manager may deliver exceptional results in one environment while creating tension in another.

In these situations, the problem is not necessarily a lack of competence, or even an individual's personality. Sometimes, it is simply a mismatch between how someone operates and what their environment requires.

Our guide to freelance IT missions describes three of these mismatches as they show up in practice: the pace of decisions, the level of autonomy expected and the relationship to quality. Our other sector guides do the same for remote work, communications, finance and admin.

This is what researchers refer to as Person-Environment Fit.

The meta-analysis by Kristof-Brown et al. (2005) found a correlation of 0.56 between person-job fit and job satisfaction, alongside a correlation of -0.46 with intentions to leave. The relationship with rated job performance was more moderate, at approximately 0.20.

These findings matter because fit does not simply influence what someone is capable of delivering. It also helps explain whether a professional relationship will be satisfying and sustainable.

For an organization engaging an external expert for three, six, or twelve months, that distinction becomes strategically important.

It is also one of the central hypotheses of the Migbirds research program: understanding the extent to which behavioral mismatches contribute to failed collaborations, integration difficulties, and reduced performance in expert outsourcing.

What Is the Real Cost of a Poor Prediction?

A poor hiring decision costs far more than the recruitment process itself.

When a professional relationship breaks down, an organization may need to restart its search, onboard a replacement, reorganize teams, and absorb accumulated delays.

With external experts, the consequences can emerge particularly quickly. An mission terminated after just a few weeks may already represent thousands of euros in fees, on top of replacement costs, management time, and potential project disruption.

Consider a simple example: a consultant charging €800 per day whose mission ends after fifteen working days. The fees already amount to €12,000. If finding a replacement requires another fifteen working days, the opportunity cost can become substantial, even if not every day translates into a direct financial loss.

And that calculation doesn't account for less visible consequences: damaged trust, internal tensions, organizational disruption, or a deteriorating client relationship.

This is precisely the economics of behavioral mismatch that Migbirds aims to investigate in the second and third parts of its scientific study.

The research begins with qualitative interviews involving operational decision-makers, HR leaders, and procurement professionals. The goal is to document real incidents, identify the behaviors contributing to difficulties, and estimate their economic consequences.

A subsequent field research protocol will compare missions staffed through conventional selection processes with those benefiting from predictive behavioral matching.

Five performance indicators will be examined: job performance, time to full productivity, team effectiveness, collaboration stability, and mutual satisfaction.

The ambition is clear: to move beyond the widely shared belief that people matter and toward a measurable understanding of their economic impact.

What Are the Limits of Behavioral Prediction?

Predicting human behavior is not the same as predicting how a machine will operate. People evolve, organizations change, and professional relationships are shaped by countless interacting factors.

Several precautions are therefore essential.

Correlation is not certainty. Even a method with strong predictive validity can produce an incorrect individual prediction. It improves the probability of making a successful decision without guaranteeing the outcome.

Context matters as much as individual characteristics. Someone may perform exceptionally well in one organization and struggle in another. Assessing a person without understanding their environment means overlooking a significant part of the equation.

Self-reported qualities are not the same as observed behaviors. Personality questionnaires can provide valuable information, but what people say about themselves does not always reflect how they act in real situations. Social desirability bias and the limitations of self-assessment must be considered.

Behavior is not fixed. The way people work evolves with experience, confidence, relationships, and working conditions. Predictive tools must account for this complexity rather than reduce individuals to permanent categories.

Scientific validation is essential. An assessment method does not become reliable simply because it draws on neuroscience, psychology, or artificial intelligence. Its effectiveness must be demonstrated through real-world data, observable criteria, and reproducible research protocols.

This is why the Migbirds research program is structured around three stages: consolidating existing scientific knowledge, documenting decision-makers' real-world experiences, and empirically testing the ability of behavioral matching to anticipate mission success.

The final stage is particularly important. Until the experimental results are available, any expected performance improvement remains a hypothesis to be tested, not a scientifically demonstrated benefit.

How Does Migbirds Apply These Findings?

At Migbirds, we start from a simple conviction: the best expert is not necessarily the one with the most impressive CV. It's the person who combines the right expertise with the behavioral conditions needed to succeed in a specific mission.

Our approach enriches traditional skills assessment with an analysis of behavioral dynamics, motivations, and professional context.

The objective is not to divide people into good and bad profiles, nor to claim that we can decode their personalities. It is to better understand the conditions under which a professional collaboration is most likely to succeed.

This thinking lies at the heart of the Human Alignment Score (HAS), an alignment indicator being developed as part of our research program.

The experimental model combines three dimensions: behavioral alignment, fit with the mission context, and technical skills alignment. The proposed initial weightings are 50%, 30%, and 20%, respectively. These are starting hypotheses that will need to be refined based on real-world observations.

To evaluate the model, we have designed a six-to-twelve-month quasi-experimental protocol involving at least two partner consulting firms. It will assess whether missions supported by Migbirds matching achieve stronger outcomes, whether experts reach full productivity faster, and whether professional collaborations prove more sustainable.

The protocol will also examine the correlation between the Human Alignment Score and observed job performance.

We are not claiming to have eliminated uncertainty from recruitment. Our goal is to understand it better, identify the behavioral factors that influence professional relationships, and progressively measure their predictive value.

Because perhaps the real question is no longer simply: "Can we predict job performance?"

It is this:

"Why do we continue to select people primarily based on what they know how to do, when their success also depends on how they will work with others?"

That is the question we aim to answer through a scientific, practical, and measurable approach.

Key references: Campbell et al. (1993); Borman & Motowidlo (1997); Schmidt & Hunter (1998); Kristof-Brown et al. (2005); Azim et al. (2010); Koopmans et al. (2012, 2014); Kuncel et al. (2013); Campbell & Wiernik (2015); Sackett et al. (2022, 2023). Migbirds Scientific Research Program, Parts I, II and III, 2026.

References

• Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range. Journal of Applied Psychology, 107(11), 2040–2068. doi.org/10.1037/apl0000994
• Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2023). Revisiting the design of selection systems in light of new findings regarding the validity of widely used predictors. Industrial and Organizational Psychology, 16(3), 283–300. doi.org/10.1017/iop.2023.24
• Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology. Psychological Bulletin, 124(2), 262–274. doi.org/10.1037/0033-2909.124.2.262
• Kuncel, N. R., Klieger, D. M., Connelly, B. S., & Ones, D. S. (2013). Mechanical versus clinical data combination in selection and admissions decisions: A meta-analysis. Journal of Applied Psychology, 98(6), 1060–1072. doi.org/10.1037/a0034156
• Kristof-Brown, A. L., Zimmerman, R. D., & Johnson, E. C. (2005). Consequences of individuals' fit at work. Personnel Psychology, 58(2), 281–342. doi.org/10.1111/j.1744-6570.2005.00672.x
• Highhouse, S. (2008). Stubborn reliance on intuition and subjectivity in employee selection. Industrial and Organizational Psychology, 1(3), 333–342. doi.org/10.1111/j.1754-9434.2008.00058.x
• Campbell, J. P., McCloy, R. A., Oppler, S. H., & Sager, C. E. (1993). A theory of performance. In N. Schmitt & W. C. Borman (Eds.), Personnel selection in organizations. Jossey-Bass.
• Campbell, J. P., & Wiernik, B. M. (2015). The modeling and assessment of work performance. Annual Review of Organizational Psychology and Organizational Behavior, 2, 47–74.
• Borman, W. C., & Motowidlo, S. J. (1997). Task performance and contextual performance: The meaning for personnel selection research. Human Performance, 10(2), 99–109.
• Koopmans, L., Bernaards, C. M., Hildebrandt, V. H., de Vet, H. C. W., & van der Beek, A. J. (2014). Construct validity of the Individual Work Performance Questionnaire. Journal of Occupational and Environmental Medicine, 56(3), 331–337.
• Azim, S., Gale, A., Lawlor-Wright, T., Kirkham, R., Khan, A., & Alam, M. (2010). The importance of soft skills in complex projects. International Journal of Managing Projects in Business, 3(3), 387–401.
• Migbirds Scientific Research Program, Parts I, II and III (2026). Internal study.