Why Organizations Keep Using Invalid Hiring Methods (and How to Change That)

Organizations keep using hiring methods that don't work. Discover the empirical validity of hiring methods and why organizations ignore the evidence.
Why Organizations Keep Using Invalid Hiring Methods (and How to Change That)
"The illusion that we understand the past fosters overconfidence in our ability to predict the future. In hiring, interviewers confidently believe they can predict performance from a brief conversation—the evidence says otherwise." — Daniel Kahneman, Thinking, Fast and Slow (2011)
Executive Summary
If the evidence is clear that resumes and unstructured interviews barely predict job performance, why do 82% of organizations still use them? The evidence is unambiguous and decades old: certain hiring methods reliably predict job performance while others are essentially useless. Yet most organizations continue using the invalid methods despite—or perhaps because of—their familiarity, social acceptability, and organizational inertia.
Understanding this paradox reveals critical gaps between what we know works and what we actually do. This disconnect represents one of the costliest inefficiencies in organizational management.
The path forward requires both understanding the psychology of why organizations resist change and implementing structural interventions that make valid methods the default.
The Validity Evidence: What Actually Predicts Job Performance
Industrial-organizational psychology has spent decades systematically measuring the predictive validity of hiring methods—whether a hiring method's results correlate with actual job performance. The research is remarkably consistent.
Highest Validity Methods
Structured Interviews (r = .43 to .70): Structured interviews use standardized questions applied consistently to all candidates, with predetermined rating criteria. Research by Harvard's Google People Analytics team found that structured interviews are more predictive of job performance than unstructured interviews (r = .43 vs. r = .24)—nearly twice as effective. A single structured interview by one interviewer yields the same validity as three to four unstructured interviews.
Work-Sample Tests (r = .39 to .56): Work-sample tests present candidates with actual job tasks or simulations. Their predictive validity increases with job complexity (r = .39 for low-complexity jobs to r = .56 for high-complexity jobs), making them exceptionally valuable for technical and complex roles.
Cognitive Ability Tests (r = .51): These assess reasoning, problem-solving, and learning ability. They predict job performance across diverse roles and are particularly strong for complex, knowledge-intensive work.
Lowest Validity Methods (Yet Most Commonly Used)
Unstructured Interviews (r = .14 to .24): Open-ended conversations without standardized questions or criteria predict job performance only slightly better than random selection. Research shows unstructured interviews explain just 14-20% of job performance variance, meaning 80-86% of variation in performance is not captured. One study noted this is "not much better than a coin toss."
Resume Screening (r = .13): Despite being the universal first filter, resumes show minimal predictive validity. Research found that 43% of employers struggle to determine an applicant's actual skills from their resume, and 51% struggle to determine whether resumes are accurate. Yet 82% of employers used resumes to hire in the past 12 months.
Unlock the Full Research
Create a free account or sign in to access the complete methodology, deep-dive analysis, and key implementation frameworks.
Don't have an account? It's free to join.
Organization Learning Labs
Research & Insights Division



Comments