Measurement Fragmentation as a Threat to Educational and Psychological Science: The TTCT as a Case Example

Measurement is the foundation of scientific inference, yet educational and psychological research often relies on instruments that are abbreviated, selectively administered, or modified without evidence that the resulting scores remain valid. This measurement fragmentation alters what a test measures, weakens the interpretability of its scores, and undermines the comparability and reproducibility of findings. The Standards for Educational and Psychological Testing emphasize that validity is version-specific and must be demonstrated for each distinct form of an instrument; reliability alone cannot establish valid interpretation. The Torrance Tests of Creative Thinking (TTCT) offer a clear illustration of these issues. When administered in full, the TTCT has demonstrated strong construct representation and long-term predictive validity across cultures and decades. However, many researchers use only one or two TTCT activities or substitute unvalidated short forms, assuming that the validity of the complete instrument transfers to these modified versions. Such practices yield findings that reflect the limitations of the altered test rather than the properties of creative thinking, often leading to unwarranted criticisms of the TTCT or of creativity assessment more broadly. This article situates the TTCT within wider meta-scientific concerns and provides recommendations for improving measurement practice, including version-specific validation, comprehensive reporting of measurement decisions, and transparency in test materials. Addressing measurement fragmentation is essential for producing interpretable, replicable, and theoretically meaningful research.