Hiring Assessment Glossary: 35 Terms Every Recruiter Should Know
Umid Rahimov9 min read
Key takeaways
- Short, plain-language definitions of the terms used in pre-employment testing.
- Grouped into: test types, scoring, fairness and quality, anti-cheating, and process.
- Use it to brief hiring managers, write job ads or compare vendors.
Test types
Pre-employment test: A standardised task or set of questions candidates complete before being hired, used to measure job-relevant skills.
Skills test: A test of a specific, learnable skill, such as Excel, SQL, accounting or a language.
Cognitive ability test: A test of reasoning and problem solving (numerical, verbal, logical). It predicts how quickly someone learns.
Critical thinking test: A test of evaluating arguments, assumptions and data to reach sound conclusions.
Situational judgment test (SJT): Realistic work scenarios where the candidate picks the most (and least) effective response.
Personality test: A questionnaire measuring stable traits, e.g. the Big Five (openness, conscientiousness, extraversion, agreeableness, emotional stability).
Work sample: A small version of the real job task, e.g. reconciling a spreadsheet or answering a customer email.
Coding assessment: A programming task run against test cases to check that the code works.
Screening questions: Must-have questions (work permit, notice period, salary expectations), often used as knock-outs.
One-way (asynchronous) video interview: The candidate records answers to preset questions, and the team reviews them later.
Essay question: A written answer, usually reviewed and scored by a person.
Scoring
Raw score: The number of points earned.
Percentage score: Raw score divided by the maximum, times 100.
Weighted score: An overall score where some tests count more than others.
Percentile: The share of candidates who scored lower. The 80th percentile means a higher score than 80% of candidates.
Cut score (pass mark): The minimum score needed to move to the next stage.
Partial credit: Points for a partly correct answer, e.g. 2 of 3 correct options.
Auto-scored vs manual review: Choice and coding questions are scored automatically. Essays and videos are scored by a person.
Item: A single question in a test.
Item difficulty (p-value): The share of candidates who answer an item correctly. Very high or very low values suggest an item to review.
Item discrimination: How well an item separates strong from weak candidates.
Fairness and quality
Validity: How well test results predict what you care about (e.g. job performance).
Reliability: How consistent results are, i.e. whether the same person would score similarly again.
Adverse impact: When a selection step disadvantages a protected group at a noticeably higher rate.
Bias: Systematic error that favours or disadvantages candidates for reasons unrelated to the job.
Structured interview: An interview with the same job-related questions and a scoring guide for every candidate.
Reasonable adjustment (accommodation): A change to the test conditions, such as extra time, for candidates with a disability.
Anti-cheating
Proctoring: Supervising an online test to prevent and detect cheating, by a person or by software.
Session recording: A screen and/or webcam recording of the test session for later review.
Behaviour monitoring: Detecting suspicious events such as tab switches, copy-paste, screenshots or leaving full screen.
Question randomisation: Showing each candidate a different set or order of questions from a pool.
Risk grade: A summary of how suspicious a candidate's session looked, e.g. Clean to Severe risk.
Process
Assessment: The full package a candidate takes for a role (several tests, questions and settings).
Question bank (pool): The set of questions a test draws from.
Candidate experience: How candidates perceive your process: length, clarity, fairness and speed of feedback.
Completion rate: The share of invited candidates who finish the assessment.
Put the terms into practice
TheyTest uses these concepts throughout: weighted scores, percentiles, randomised question banks, risk grades and session recordings, in 9 languages.
FAQ
What is a good cut score?
There's no universal number. Set it after looking at the first group of candidates, and check that people above the cut perform well on the job.
What is the difference between validity and reliability?
Reliability means the test gives consistent results. Validity means those results predict job performance. A test can be reliable without being valid.
What does proctoring mean in online tests?
Supervising the test to prevent and detect cheating, for example with recording, tab-switch detection and time limits.
