Scoring divergent thinking tests by computer with a semantics-based algorithm

Kenes Beketayev, Mark A. Runco

Research output: Contribution to journalArticlepeer-review

25 Citations (Scopus)


Divergent thinking (DT) tests are useful for the assessment of creative potentials. This article reports the semantics-based algorithmic (SBA) method for assessing DT. This algorithm is fully automated: Examinees receive DT questions on a computer or mobile device and their ideas are immediately compared with norms and semantic networks. This investigation compared the scores generated by the SBA method with the traditional methods of scoring DT (i.e., fluency, originality, and flexibility). Data were collected from 250 examinees using the “Many Uses Test” of DT. The most important finding involved the flexibility scores from both scoring methods. This was critical because semantic networks are based on conceptual structures, and thus a high SBA score should be highly correlated with the traditional flexibility score from DT tests. Results confirmed this correlation (r =.74). This supports the use of algorithmic scoring of DT. The nearly-immediate computation time required by SBA method may make it the method of choice, especially when it comes to moderate- and large-scale DT assessment investigations. Correlations between SBA scores and GPA were insignificant, providing evidence of the discriminant and construct validity of SBA scores. Limitations of the present study and directions for future research are offered.

Original languageEnglish
Pages (from-to)210-220
Number of pages11
JournalEurope's Journal of Psychology
Issue number2
Publication statusPublished - May 2016


  • Assessing creativity
  • Associative networks
  • Computer creativity
  • Creativity test
  • Divergent thinking
  • Flexibility
  • Ideas
  • Ideational fluency
  • Originality
  • Semantic networks

ASJC Scopus subject areas

  • Psychology(all)

Fingerprint Dive into the research topics of 'Scoring divergent thinking tests by computer with a semantics-based algorithm'. Together they form a unique fingerprint.

Cite this