A Dictionary of Assessment Terminology

Affective Domain - This refers to the nonintellectual aspects of behavior.

Aptitude - This traditionally refers to a relatively homogenous and clearly defined segment of ability.

Behavioral Styles - These refer to any of several categorizations of personality into a construct or matrix of characteristic types. These types are essentially models by which observed behavior and interaction may be discussed and understood. (Also called Social Styles)

Benchmark Pattern - This refers to a composite picture of the characteristics of top performers produced by assessing and analyzing a sample group. (Also called a Success Pattern or Success Profile)

Cognitive Abilities - These include various elements of intelligence, characterized as Reasoning, Numerical Reasoning, Verbal Reasoning, Spatial Reasoning, Mental Alertness, etc.

Construct - A psychological characteristic that is considered to vary across individuals. A construct is not directly observable, but it is a theoretical concept derived from research and experience that has been constructed to explain observable behavior patterns.

Core Personality - This refers to an individual's fundamental traits of personality that are established during childhood, and tend to remain the same over time in the absence of some life trauma.

Conative - This refers to the individual differences in motivational content or to the differences in the things for which people strive (Miller, 1991).

Equivocation - The frequent selection of the middle response (sometimes, uncertain, etc.) of several extremes or the response that is the least committal. This has the effect of diluting the information provided.

Factor Analysis - Any of several methods of analyzing the inter-correlations or co-variances among variables by constructing hypothetical factors, which are fewer in number than the original variables. It indicates how much of the variation in each original measure can be accounted for by each of the hypothetical factors.

Faking - This refers to attempts by the test participant to misrepresent their true behavior through exaggeration, distortion, equivocation, avoidance, or some other means.

Generations of Assessment Instruments - This refers to a construct which sorts assessment instruments into categories based on various factors reflecting the relative level of psychometric methodology used to produce that instrument's report. (I.e. First Generation tools use older and much less advanced types of items, while Fifth or Sixth Generation tools use newer and more sophisticated forms of items, consequently resulting in more accurate and more detailed information.)

Intelligence

Crystallized Intelligence - This refers to intelligence that is dependent upon culture, education, or experience.

Fluid Intelligence - This refers to raw intelligence or reasoning ability that is not dependent upon culture, education, or experience. Measurements of fluid intelligence are much more effective in predicting performance in diverse situations. Measures of fluid intelligence are less likely to create adverse impact problems.

Item - This refers to a question or a problem on an instrument.

Ipsative - This is a type of scoring generated by forced choice items (e.g. Select the word that MOST describes you and the word that LEAST describes you from the following: moody, thoughtful, enthusiastic, or intense.) For more information on ipsatively scored tests, see Q & A Choosing Assessments.

Item Analysis - The process of assessing certain characteristics of test items, usually the difficulty value, the discriminating power, and sometimes the correlation with an external criterion.

Normative - This is a type of scoring produced by testing a large population and generating a normal bell curve distribution of the results. The distribution is then divided into standard tenths (or ninths in older instruments), creating a quantified, normal scale with which to measure and compare individuals.

Personality Types - These are categories of people who exhibit particular combinations of psychological characteristics, the assumption being that this combination is unique and distinguishes this type from another (Miller, 1991).

Psychometrics - The science of measuring the characteristics of human behavior, personality, cognitive abilities, interests, or aptitudes.

Reliability

Test-Retest Reliability - This refers to a test's stability over time. Lower test-retest reliability indicates that the instrument is not measuring core behavior traits, but is assessing states, which are subject to change with mood or circumstances.

Internal Reliability - This refers to the ability of a test to measure discrete variables. The degree to which variable measurement is cross-related lowers internal reliability.

Social Desirability - This refers to the set of responses in which participants tend to answer in such a way as to portray themselves in the most favorable light.

Social Styles - These refer to any of several categorizations of personality into a construct or matrix of characteristic types. These types are essentially models by which observed behavior and interaction may be discussed and understood. (Also called Behavioral Styles)

Standard Deviation - This is a measure of the variability of a sample of scores from the average or mean of that same sample.

Success Pattern or Success Profile - This refers to a composite picture of the characteristics of top performers produced by assessing and analyzing a sample group. (Also called a Benchmark Pattern)

Technical Manual - This is a step-by-step description of how the instrument was constructed. It outlines the various constructs used by the assessment, and the basis of their formulation. The numerous validity studies are detailed with the description of the various populations used in the studies.

Types - This refers to the concept of sorting people into various categories or sets of behavior for the purpose of discussing interaction. Early assessments were based on this concept.

Validity

Validity - A test is said to be valid if it measures what it claims to measure. There is no one validity coefficient for a test. A test is always valid for some purpose, and therefore is more valid in some circumstances than in others (Kline, 1993).

Deductive Validation starts with a theory in order that the content of the test is defined and that hypotheses are generated concerning what should correlate with the test scores. Inductive Validation starts with the test measure and then tries to infer what it must be a measure of by examining its relationship with other things.

Construct Validity - This refers to whether a test is measuring what it claims to measure as judged by accumulated evidence.

Concurrent Validity - A test is said to have concurrent validity if it correlates highly with a "benchmark" test of the same variables.

Content Validity - This refers to tests such as skills, ability or attainment tests where the domain of items is very defined.

Criterion Validity - This refers to evidence that shows the extent to which scores on a test are related to a criterion measure.

Concurrent Criterion-Related Validity - This refers to evidence of criterion validity in which predictor and criterion information are obtained at approximately the same time.

Predictive Criterion-Related Validity - This refers to evidence of criterion validity in which criterion scores are observed at a later date (e.g. after job performance).

Face Validity - An instrument is said to be face valid if it appears to be measuring what it claims to measure.

Predictive Validity - A test is said to have predictive validity if it will predict some variable.

Synthetic Validity - This refers to the practice of using validity generalization to "synthesize" the criteria for a new job through extrapolation from known predictive criteria in other jobs.

Validity Generalization - This refers to applying validity evidence obtained in one or more situations to other similar situations on the basis of simultaneous estimation, meta-analysis, or synthetic validation arguments.

Validity Scales - This refers to any of a variety of scales designed to indicate exaggeration, faking, equivocation, or deception by test participants.