1 of 6

Tarek Azzam

Director

tarekazza@ucsb.edu

Artificial Intelligence and Validity

Center for Evaluation & Assessment

The Gevirtz School

2 of 6

High quality evaluation

is not a technocratic exercise!!!

3 of 6

AI & Validity

Quantitative Validity

Stages

Types of Validity

Evaluator/ Researcher

Artificial Intelligence

Methods

Internal validity

Yes

Not Yet

External validity

Yes

Yes

Measurement

Construct validity

Yes

Maybe

Concurrent validity

Yes

Yes

Predictive validity

Yes

Yes

Analysis

Statistical conclusion validity

Yes

Yes

Qualitative Trustworthiness

Types of Validity

Evaluator/ Researcher

Artificial Intelligence

Credibility

Yes

Not Yet

Transferability

Yes

Maybe

Dependability or trustworthiness

Yes

Not Yet

4 of 6

Concepts and examples within educational assessment

From: Dorsey, D. W., & Michaels, H. R. (2022). Validity Arguments Meet Artificial Intelligence in Innovative Educational Assessment. Journal of Educational Measurement59(3), 267-271.

“AI in the form of Natural Language Processing (NLP) methods can “score interviews, essays, or other forms of text-based assessments, while overlooking the fact that language/text exists at multiple levels of abstraction: syntactic, lexical, semantic, and pragmatic, involving a host of construct-based inferences—ones that involve complex interactions between culture and language” pg. 268

5 of 6

Hype Cycle

Source: https://www.gartner.com/en/research/methodologies/gartner-hype-cycle

6 of 6

Thank you