Pre-Employment Assessment

The main types of pre-employment assessment

Each type measures something different, and the right mix depends on what the job actually needs.

• Skills tests. Typing speed, data entry accuracy, Excel, coding, language proficiency. Direct and easy to defend. See skills assessment.
• Cognitive ability tests. Problem solving, numerical and verbal reasoning, learning speed. Among the strongest single predictors of job performance, but also the most likely to show score gaps between demographic groups.
• Personality and behavioral questionnaires. Conscientiousness, service orientation, reliability. Moderate predictors; easy to fake if candidates guess what you want. Part of the broader family of psychometric tests.
• Situational judgment tests. Realistic scenarios with ranked responses. Good for customer-facing and supervisory roles.
• Work samples and simulations. A mock customer call, a chat queue, a pick-and-pack exercise. High predictive power and candidates see them as fair.
• Integrity tests. Attitudes toward theft, rules and safety, mostly used in retail, cash handling and security roles.

Where assessments fit in the hiring funnel

Placement matters as much as the test itself. Put a 45-minute battery right after a two-minute mobile application and you will lose a large share of hourly applicants before anyone has spoken to them.

A common high-volume design: short knockout questions at application, then a 10 to 20 minute assessment the candidate can take immediately on their phone, then a structured interview for those who pass. For professional roles, assessments usually come after a recruiter screen, when the candidate has more reason to invest the time.

Rule of thumb: the earlier the test, the shorter it must be. Under 20 minutes for anything before a human conversation, and tell candidates up front how long it takes.

A worked example: building a contact center assessment

A healthcare billing company hires 25 patient account representatives a month. Its old process was a phone screen and one interview, and 35% of new hires left during training.

The TA lead, Priya Shah, adds a 15-minute assessment between application and interview: a typing test (target 35 words per minute at 95% accuracy), a short data-entry accuracy exercise and a six-item situational judgment section about frustrated callers. Cut scores are set using the results of current agents who were rated good performers.

After three cohorts, about 55% of applicants pass. Interview volume drops by nearly half, which frees supervisors, and training attrition falls from 35% to roughly 20%. She also runs an adverse impact check each quarter and finds the typing component passing groups at similar rates, while the original typing cut score of 45 wpm had been screening out more older applicants without predicting anything extra, so she lowered it.

Validity: what makes an assessment worth using

Validity is whether scores actually relate to job performance. Ask any vendor for evidence of three things:

• Content validity. The test content clearly mirrors tasks of the job, which is easiest to show for skills tests and simulations.
• Criterion validity. Scores correlate with real outcomes such as quality scores, sales, attendance or retention, ideally in roles like yours.
• Reliability. The same person would score similarly if they took it again.

Then validate locally. After a few hundred hires, check whether people who scored higher actually performed better on your own metrics. If they did not, the test is costing you candidates for no return.

Legal and compliance rules in the US

Under Title VII of the Civil Rights Act, an employment test that screens out a protected group at a disproportionate rate must be job related and consistent with business necessity. The Supreme Court set this standard in Griggs v. Duke Power Co. (1971). The Uniform Guidelines on Employee Selection Procedures (1978) describe how employers should monitor adverse impact, including the four-fifths rule of thumb: a selection rate for any group below 80% of the highest group's rate is generally treated as evidence of adverse impact.

Other rules to know:

• ADA. Offer reasonable accommodations, such as extra time or screen reader compatibility, and do not use tests that amount to a medical exam before a conditional offer.
• No score adjustments by group. The Civil Rights Act of 1991 prohibits adjusting scores or using different cut scores based on race, color, religion, sex or national origin.
• Automated tools. NYC Local Law 144 requires a bias audit within the past year and candidate notice before using an automated employment decision tool for candidates in New York City. Other states and the EU AI Act add their own rules for AI-scored assessments.

Rules vary by state and country, so review your assessment program with employment counsel.

Common mistakes

• Buying a generic test for a specific job. A general aptitude battery for a forklift role tells you little. Start from the job's actual tasks.
• Too long, too early. Every extra minute before human contact costs completions.
• Cut scores set by guesswork. Base them on incumbent data or vendor norms for similar roles, then review.
• Never checking adverse impact. Pull pass rates by group at least quarterly.
• Treating scores as the whole decision. Use them as one input alongside a structured interview, not a lone gate unless the skill is truly mandatory.
• Ignoring cheating. Unproctored tests invite proxy test-takers. Use assessment tools with time limits, question randomization and identity checks for higher-stakes roles.