Assessment glossary for training providers: 30 terms explained

An assessment glossary is a reference list that defines the words used to build, deliver, score and report tests, so you can read a vendor proposal, a standard or a psychometric report without guessing. This one explains 30 terms in plain English for training providers, trainers and L&D staff who are new to assessment vocabulary.

The glossaries that rank for this phrase are mostly written by colleges for faculty, with entries such as curriculum map and program review. This one covers running exams, quizzes and certificates: items, pass marks, item statistics, interoperability standards and credentials.

Each entry gives a direct definition and then one line on what the term means in practice for a provider. The Standards for Educational and Psychological Testing cited here are the 2014 edition, and AERA announced the committee to revise them in June 2024, so check for a newer edition. Facts that can change are dated as of October 2026.

Key takeaways

• Validity asks whether evidence supports using scores for a stated purpose, while reliability asks whether scores are consistent, so a test can have one without the other.
• A cut score is a judgment, not a discovered fact, so record who set it, with which method and on what date.
• QTI moves questions and results, LTI launches a tool inside a learning platform, SCORM packages courses and xAPI records activity, so each solves a different problem.
• A certificate of completion, a CEU and certification are three different claims, and each needs different evidence behind it.

What are the main types of assessment training providers use?

Training providers mainly use five kinds of assessment: formative checks that guide learning, summative tests that decide an outcome, diagnostic checks that find gaps, placement tests that sort learners into levels, and performance assessments that judge real work. Purpose, not question format, defines each type.

What is formative assessment?

Formative assessment is a planned, ongoing process in which teachers and learners gather evidence of learning during instruction and use it to improve understanding. Carnegie Mellon’s Eberly Center calls it generally low stakes. The wording follows the 2017 CCSSO FAST SCASS definition.

In practice: use a short mid-session check to decide what to reteach, and keep it out of the final pass.

What is summative assessment?

Summative assessment evaluates learning at the end of an instructional unit against an established yardstick, such as a standard or benchmark. Carnegie Mellon’s Eberly Center describes it as often high stakes.

In practice: in a multi-stage program, let one or two stages decide the pass and keep the others formative.

What is diagnostic assessment?

Diagnostic assessment is a check given before instruction that shows what learners already know and where their gaps are. A pre-test is one common form; Carnegie Mellon’s Eberly Center lists pre-tests as direct measures of capability on entry.

In practice: build any pre-test from the same blueprint as the post-test, so the score difference can show gain.

What is a placement test?

A placement test is a brief check that assigns each learner to the course level or group that fits their current skills. It differs from a diagnostic, which maps specific skill gaps, and from a pre-test, which sets a baseline for measuring gain.

In practice: set a separate cut score for each level and review learners who turn out to be misplaced.

What is performance assessment?

Performance assessment is an assessment in which learners complete an authentic task, such as running a procedure or analyzing a case, and a trained observer scores the work against criteria. UL/SCALE’s quality criteria link such tasks to mastery of standards-aligned content and skills.

In practice: hands-on skills need an observation checklist or rubric and a rater practice session before the first live cohort.

Which terms describe how a test is designed and built?

Test design starts with a learning objective, becomes a blueprint that allocates items across topics, and ends with individual items stored in an item bank. Distractors, the wrong options in multiple-choice items, often decide whether an item works, so they get their own entry here.

What is a learning objective?

A learning objective is a statement of what a learner should be able to do after instruction, written so that it can be observed and assessed. The 2001 revision of Bloom’s taxonomy offers six levels for the verb: remember, understand, apply, analyze, evaluate and create.

In practice: tag every item to one objective so coverage can be counted.

What is a test blueprint?

A test blueprint, also called test specifications, is the formal design of an exam. It lists content areas with the percentage of items for each and the total item count, and may add cognitive levels and rules such as time limits, following Assessment Systems’ description.

In practice: build the pre-test, post-test and final exam from one blueprint and record which version each learner took.

What is a test item?

A test item is a single question or task on a test, together with its response options and scoring key. In a multiple-choice item, the lead-in is the stem and the correct option is the key.

In practice: trainers say question and psychometricians say item, but both mean one scored unit, including a practical task.

What is an item bank?

An item bank is a managed collection of test items, each stored with tags, a status and a revision history, from which exams are assembled.

In practice: keep version history so you can reproduce any past exam, as covered in building an item bank for a training company.

What is a distractor?

A distractor is an incorrect answer option in a multiple-choice item, written to attract a learner who holds a specific misconception. Penn State’s Guide to Item Analysis defines distractors as the response options that are not the correct answer.

In practice: write each distractor from a real learner error, because an option nobody picks does no work.

What do scoring and standard-setting terms mean?

Scoring terms describe how a performance becomes a result and which result counts as passing. A rubric scores judged work, a cut score marks the pass point, standard setting is the method for choosing it, and criterion-referenced reporting compares learners with that standard instead of with each other.

What is a rubric?

A rubric is a scoring tool that describes performance expectations for a piece of work. It has three parts: the criteria being assessed, descriptors of what each criterion looks like, and a rating scale of performance levels, as Carnegie Mellon’s Eberly Center describes it.

In practice: write descriptors an observer can see or count, and have two raters score sample work first.

What is a cut score?

A cut score is a selected point on a test’s score scale that separates one result category from another, such as pass from fail. ETS authors Zieky and Perie state that there are no “true” cut scores; the numbers reflect the combined judgments of the people involved.

In practice: a flat 70 percent is a cut score too, so record who set it and how, as shown in setting pass marks for professional training courses.

What is standard setting?

Standard setting is the structured process of deciding where a passing score falls, which Cizek describes as following a prescribed, rational system of rules or procedures. In the Angoff method, judges estimate the probability that a minimally acceptable person answers each item correctly.

In practice: a small provider can run a modified Angoff session, which Cizek notes often has two or more rounds of ratings.

What is criterion-referenced assessment?

Criterion-referenced assessment compares a learner’s score with a fixed criterion, such as a cut score or a body of knowledge, instead of with other learners’ scores. Norm-referenced assessment does the reverse. Both describe how scores are interpreted, not types of test.

In practice: a course that requires 80 percent should pass every learner who reaches 80, however many that is.

Which assessment glossary terms cover test quality and fairness?

Quality terms test whether results can be trusted and used. Validity asks whether evidence supports an interpretation of scores, reliability asks whether scores are consistent, the standard error of measurement sizes the uncertainty, difficulty and discrimination screen items, and accommodations keep the exam fair for people with disabilities.

What is validity?

Validity is the degree to which evidence and theory support the interpretations of test scores for their proposed uses. The Standards for Educational and Psychological Testing call “the validity of the test” an incorrect unqualified phrase (page 11).

In practice: state the decision a score supports, such as “may work unsupervised on this task”, and keep the evidence for it.

What is reliability?

Reliability, which the Standards call reliability/precision, is the consistency of test scores. A reliable exam gives similar results for the same learner under equivalent conditions. Consistent scores can still measure the wrong thing, so reliability alone does not show validity.

In practice: ask for a reliability estimate per exam form. Penn State’s guide calls alpha above 0.80 ideal, its own rule of thumb, not a Standards requirement.

What is the standard error of measurement?

The standard error of measurement (SEM) estimates how much a learner’s observed scores would typically vary around their true score. In Harvill’s NCME module example, with an SEM of 2, about two chances in three put one observed score within 2 points of the true score.

In practice: a learner one point under the pass mark sits inside that noise, so decide in advance how near-misses are handled.

What is item difficulty?

Item difficulty is the percentage of learners who answered an item correctly, often written as a p-value. A higher value means an easier item. Penn State’s guide defines it this way, and suggested target ranges vary by source and by the item’s purpose.

In practice: judge each value against the item’s job; item analysis for trainers compares ranges by source.

What is the discrimination index?

The discrimination index shows how well an item separates learners who scored high on the whole test from those who scored low. Positive values mean stronger learners get it right more often. Penn State’s guide uses the point-biserial correlation, which ranges from -1.0 to 1.0.

In practice: a negative value means weaker learners did better, so check the answer key before blaming the learners.

What is an accommodation?

An accommodation is a change to how a test is presented, taken or answered, such as extra time, screen-reading software or a scribe, that lets a person with a disability show the skill the test measures. These examples come from the ADA’s testing guidance.

In practice: record each request and decision. The ADA page (updated February 2020) covers licensing and credentialing exams, so ask a lawyer if it applies to you.

What do delivery, security, accessibility and interoperability terms mean?

Delivery terms cover how an exam reaches the learner and how its data moves between systems. Proctoring supervises the test taker, QTI moves questions and results, LTI launches a tool inside a learning platform, SCORM packages content for a learning management system, xAPI records learning activity, and WCAG sets accessibility targets.

What is proctoring?

Proctoring is the supervision of an exam to confirm who is taking it and to deter or detect rule-breaking. It can be live, recorded for later review, or automated with software flags. The 2022 ITC and ATP guidelines list remote online proctoring among common applications.

In practice: match supervision to the stakes, so a low-stakes quiz does not need certification-exam controls.

What is QTI?

QTI (Question and Test Interoperability) is, in 1EdTech’s words, a way of packaging and moving tests and questions from one application to another. QTI 3.0 is the current version, and its core assessment, section and item features reached final release in May 2022.

In practice: QTI is your exit route, so check that your items can leave a platform before you build in it. See the 1EdTech QTI page.

What is LTI?

LTI (Learning Tools Interoperability) is a 1EdTech technical standard that connects learning tools to a learning environment, such as an LMS, without logging into each tool. LTI 1.3 is current, and its LTI Advantage services include grade return and user roles.

In practice: an LTI launch lets learners open an outside exam from your LMS, and grade return sends scores back.

What is SCORM?

SCORM (Sharable Content Object Reference Model) is a set of ADL e-learning standards that lets browser-based courses work across different learning management systems. The most recent release is SCORM 2004 4th Edition, from 2009.

In practice: check which SCORM version your LMS imports, because SCORM is still widely used.

What is xAPI?

xAPI (the Experience API) is, per ADL, a data and interface standard that lets software applications capture and share data on human performance. It became IEEE 9274.1.1-2023, also called xAPI 2.0.

In practice: item-level results help when a client asks how a score was reached, but you need a learning record store.

What is WCAG?

WCAG (Web Content Accessibility Guidelines) is the W3C’s set of recommendations for making web content more accessible, organized under four principles and three conformance levels: A, AA and AAA. WCAG 2.2 became a W3C Recommendation on 12 December 2024 and does not replace 2.1.

In practice: when a vendor says an exam is accessible, ask for the version, level and pages tested. See the W3C WCAG 2.2 text.

The table below compares the four interoperability standards by what each moves, who maintains it and what its status was as of October 2026.

Pro tip: Test the Export, Not the Checkbox. The 1EdTech QTI page lists separate certification types for authoring, import, export, import and export, and delivery, so “QTI support” can mean any one of them. Ask which type a vendor holds, then export five of your own items and import them into a second tool.

Which assessment glossary terms cover credentials and compliance?

Credential terms separate what you can honestly claim. A certificate of completion records that a learner finished a course, certification confirms competence against a published scheme through an independent process, a CEU quantifies contact hours, and accreditation is an independent review of the provider, not of the learner.

What is certification?

Certification is a credential that gives assurance an individual meets defined competence requirements within a specific certification scheme. ISO/IEC 17024:2026, published in March 2026, sets requirements for bodies that certify persons; the 2012 edition is withdrawn.

In practice: unless an independent scheme assessed the person, call your credential a certificate of completion; see certificate of completion vs certification.

What is a certificate of completion?

A certificate of completion is a record that a learner finished a course or program. On its own it documents the learning event, not verified competence. The Institute for Credentialing Excellence separates attendance certificates, where mastery need not be shown, from assessment-based certificates, where it must.

In practice: word it as “completed” and state the criteria met, such as attendance or a passed quiz.

What is a CEU?

A continuing education unit (CEU) is a unit of non-credit learning: IACET states that one CEU equals ten contact hours of participation in an organized learning experience. A non-accredited CEU may or may not reflect external review, while an accredited one reflects third-party verification.

In practice: 14 contact hours equal 1.4 CEUs, but ask your accreditor whether assessment time counts. See IACET’s CEU page.

What is accreditation?

Accreditation, in IACET’s usage, is independent third-party verification that a provider’s programs meet a published standard, the ANSI/IACET 1-2018 Standard for Continuing Education and Training. The word has no single meaning across regulators, so check which body’s rules apply to you.

In practice: IACET evaluates the organization’s instructional design and delivery management system, so keep assessment and learner records.

Pro tip: Fix Your Own Definitions First. Before you configure a platform or answer a client’s data request, write one line each for “attempt”, “completed”, “passed” and “certified” as your records will use them. In a hypothetical program with 120 enrolled, 96 completers and 78 credentials, the rate is 65 percent of enrollees or 81 percent of completers.

What should you do next?

Pick the five or six terms your next decision needs. For a vendor call that is item bank, QTI and LTI; for a pass-mark review it is cut score, standard setting and the standard error of measurement; for a client report it is reliability, validity and accommodation.

Then write one-line definitions of the words your own policies use, so every trainer and platform setting means the same thing by “passed” or “completed”. When comparing platforms, ask each vendor to show a QTI 3.0 export and an LTI 1.3 launch instead of accepting a checkbox answer.

As of October 2026, SimpliAssess’s features page lists item banking, tag management, an AI feature for building quizzes, tests and exams, and an API-first design. It does not name QTI, LTI, SCORM, xAPI or accessibility conformance, so ask any vendor, including SimpliAssess, which standards it supports.

Frequently asked questions

What is the difference between validity and reliability?

Validity asks whether evidence supports using a score for a stated purpose, while reliability asks whether scores are consistent. A scale that is always five pounds off is consistent but wrong. Reliability alone never shows that an exam is valid for a decision.

What is the difference between formative and summative assessment?

Formative assessment monitors learning during instruction and is generally low stakes, while summative assessment evaluates learning at the end of a unit and is often high stakes, according to Carnegie Mellon’s Eberly Center. The difference is purpose, not question format.

What is the difference between an assessment and a test?

A test is a device or procedure that collects a sample of a person’s behavior in a specified domain, while assessment is the broader process that integrates test information with other evidence. The Standards for Educational and Psychological Testing describe assessment as a broader term than test.

What is a good passing score for a training course?

No universal passing score exists. ETS authors Zieky and Perie state that there are no “true” cut scores, so a flat 70 percent is a cut score too. Defend yours by documenting the method, the people who set it and the date.

Is a CEU the same as a certificate of completion?

No. A CEU is a unit of measure, and IACET defines one CEU as ten contact hours. A certificate of completion is a document recording that a learner finished a course. A certificate can state CEUs awarded, but only an accredited CEU reflects third-party verification.

Do training providers need QTI, LTI, SCORM and xAPI?

Not necessarily all four; it depends on your LMS and where results must go. QTI matters for portable items, LTI for launching exams from an LMS, SCORM for packaged courses and xAPI for detailed activity records. Ask your LMS vendor which of them it supports.