The process of deciding: 1) who will take the assessment (e.g., an adult, a child, or a caregiver acting as a proxy for an adult or child); 2) who will record answers to the assessment (e.g., the individual taking the assessment or a proctor); and 3) how answers will be captured (e.g., paper & pencil or computer).
Application Program Interface. Programming instructions and standards for accessing a Web-based software application or tool.
An online data collection tool that enables researchers to create study–specific Web sites that securely capture respondent data. Studies can include measures from the Assessment Center library, as well as custom instruments created or entered by the researcher.
A group of measures administered together. The NIH Toolbox®, for example, contains a battery of tests all designed to measure cognition.
Computer Adaptive Test (CAT)
A flexible, computer–driven measure that can use any items (e.g., questions) in an item bank. A CAT selects only those items that sharpen the estimate of a respondent’s score on the domain being measured. CAT length varies, but usually includes four to seven items.
Data Collection Tools
Methods for capturing answers provided during an assessment. Common methods include paper and pencil administration, online data collection software, and mobile apps.
The trait or conceptual area that is being assessed. It can be a symptom (e.g., pain) or a functional capability (e.g., physical function). A domain can also be divided into several related traits, concepts, or constructs sometimes labeled as “subdomains”.
A question or statement and its response choices in a measure.
A collection of carefully selected and tested items that provide an operational definition of a trait or construct. A good item bank covers the entire continuum of the latent trait being measured, capturing different severity levels along the continuum. (See also: Item)
Item Response Theory (IRT)
A modern psychometric theory that assumes an individual’s responses to a set of items or questions are related to an unmeasured “trait”. The trait scores for each individual predict that person’s probability of endorsing each item with which they are presented.
Widely accepted measures used in healthcare and other settings (e.g., SF-36, FACIT).
Linking (also known as cross-walking)
The process of establishing relationships between scores on two different measures (e.g., PROMIS® Depression and the PHQ-2). See Linking Measures for more information.
An assessment instrument, survey, questionnaire, or test. In HealthMeasures, these can be a fixed length short form, profile, battery, or CAT (e.g., computer adaptive test). A measure can be a self- or proxy-report survey or a performance test of function administered by a trained proctor (e.g., NIH Toolbox® Grip Strength).
Data about a particular population expected to be representative of individuals who will take a particular assessment. This data serves as a reference for interpreting individual assessment scores.
Performance Test of Function
The person being tested is asked to complete a task or activity by a trained proctor. These are sometimes called “objective tests”.
The level of variation or error in a score. Measures with greater precision have lower error.
A fixed set of short forms covering a wide range of self-reported health domains that are administered together. For example, the PROMIS-29 is a profile measure designed for adults that includes physical function, fatigue, pain interference, pain intensity, sleep disturbance, depression, anxiety, and ability to participate in social activities short forms.
A person who responds to an assessment about somebody else (e.g., an assessment about a child or about an adult who cannot respond for themselves).
The extent to which a measure yields reproducible and consistent results.
A person’s direct report on his or her feelings, function, well-being, symptoms, or life satisfaction without interpretation of that response by a clinician or other person. Sometimes this is provided by a proxy respondent (e.g., parent reporting for a child).
A subset of items selected from a larger collection of items that generates one score for a domain. Sometimes this is called a fixed length form or fixed form. (See also: Item and Item Bank).
A standardized score that has a mean of 50 and standard deviation of 10 in a reference population (e.g., U.S. General Population).
The underlying (latent) construct estimated from the responses people give to the items in a measure. These items have been previously calibrated by an IRT (e.g., item response theory) model.
The extent to which an instrument measures what it is intended to measure and can be useful for its intended purpose.