Key points:
- Assessment discussions are examining how AI like ChatGPT might play a role
- Individualized assessment offers a fair and equitable option to meet varying students’ needs
- See related article: Strategies for using alternative assessments in the classroom
The National Council on Measurement in Education’s (NCME) annual meeting has always offered an opportunity to learn about innovative research and new trends in student assessment. It is a chance to get hints of where the field is moving and what will be available to school districts, teachers, and students.
This year did not disappoint. There were three notable topics at the conference that signal new directions in assessment: through-year assessment, artificial intelligence and machine learning, and individualized assessment.
1. Through-Year Assessment
Through-year assessment is a type of assessment that has several different versions. In general, through-year assessment refers to testing that takes place throughout the school year to provide feedback on student progress and ultimately make a proficiency determination. At least three different sessions at NCME’s annual meeting dealt with through-year, and each one focused on a different version of it.
In one version, through-year assessment blends interim and summative assessment together. The interim assessment is given in the fall and winter, while the summative test is given in the spring, to make a proficiency determination. A second version of through-year makes proficiency determinations for specific curriculum standards as they are taught. Another version of through-year assessment is somewhat a blend of the prior two. It uses an interim assessment during the school year, but test content is cumulative, and the proficiency determination is made in the spring.
Each through-year design has a strength that distinguishes it from another type. The first version is a strong way to measure growth because test content is similar from one test to the next. The second approach addresses opportunity to learn because test content focuses on content recently taught. The third version of through-year is like the first, but the test blueprint shifts over time
2. Artificial Intelligence and Machine Learning
Artificial Intelligence and machine learning (AI/ML) were hot topics–but not new topics–at the conference. ChatGPT caught everyone’s attention this year, including people in the measurement community. There was a lot of water cooler discussion about ChatGPT, what it could do for assessment, and what it should not do for assessment. Automatic item generation and reading passage creation were among potential benefits discussed. The main concerns centered around how ChatGPT might perpetuate bias and stereotypes, and that it may create problems with intellectual property or lead to copyright violations.
AI/ML has been a part of measurement research for years. Initial uses were focused on automated essay scoring. There were several sessions on this topic and advances in natural language processing. Newer machine learning topics focused on using it for cheating detection and for managing item banks.
3. Individualized Assessment
NCME’s annual meeting had numerous sessions that related to individualized assessment that will ultimately lead the field beyond standardized testing. I am using individualized assessment as a broad term that encompasses many lines of measurement work including fairness and equity, culturally responsive assessment, and test-taking behavior and engagement. Individualization is the opposite of standardization, where conditions of measurement are made the same for everyone (i.e. a one-size-fits-all approach).
The idea behind individualized assessment is that a test experience is optimized for every examinee to maximize their motivation and engagement. Tests are tailored to individuals and their personal experiences and background. For example, an AI/ML generated reading passage could be written with a theme that uses a student’s personal interest in music. Individualized assessment is fair and equitable for everyone because every student is presented with conditions that give them the best chance to succeed.
Related:
What the F? Grading strategies for early career teachers
4 simple strategies to help students ace standardized tests
- High school students say AI will change the workforce - April 18, 2024
- Motivating students using the Self-Determination Theory - April 17, 2024
- Michigan Virtual’s statewide workgroup releasing AI guidance for K-12 educators - April 17, 2024