Article
What Makes an Assessment Accurate?
“Accurate” sounds simple, but an assessment can be accurate only in relation to a defined interpretation and use.
For CheckYourDISC, the intended interpretation is that scores estimate self-reported behavioral preferences across four defined dimensions. Evidence must answer several different questions:
- Content: Do the questions cover the intended behavior without irrelevant material?
- Structure: Do response relationships support the proposed dimensions or circular pattern?
- Consistency/precision: Would scores contain too much random variation?
- Stability: Are results reasonably similar over an appropriate interval when behavior is not expected to change substantially?
- Convergence: Do scores relate to other measures they theoretically should resemble?
- Discrimination: Are they distinguishable from constructs they should not duplicate?
- Fairness: Do irrelevant language, culture, disability, device, or group factors distort interpretation?
- Usefulness/consequence: Does the report support the stated decisions without creating predictable misuse?
A high satisfaction rating is not accuracy. A report feeling personal is not accuracy. A high coefficient alpha is not accuracy. A million responses are not accuracy if the construct or questions are poor.
The honest answer can change by version. That is why the report carries assessment, scoring, and report versions and why the methodology page distinguishes evidence already collected from planned research.
Try this: When a site claims “98% accurate,” look for the definition of accuracy, comparison criterion, sample, version, analysis, and uncertainty. If none appears, treat the number as marketing.
Related terms: Validity · Reliability · Construct · Fairness · Criterion evidence
Research standard: Standards for Educational and Psychological Testing, https://www.testingstandards.net/