The Promise of Interim Data
High-stakes summative testing has been the dominant model for accountability for many years. However, interim assessments administered throughout the year could provide a better picture of student learning with less pressure. Schools and districts often administer both interim and summative assessments, but face challenges comparing the results.
The question is, how can educators connect these different score scales and evaluations of proficiency?
The Assessment Process
The process starts with gathering student score reports for both the interim and summative assessments. Summative tests, like state accountability exams, typically use standardized scaled scoring, while interim assessments have their own score metrics based on the test design.
The next step involves identifying any existing linking or equating tables that can convert the interim score scales into the summative score scheme. Many test publishers have conducted alignment studies that provide tables for translating interim results into the summative scale. If no tables currently exist, the test publishers need to be engaged to conduct an equating study that will derive the needed scoring crosswalk.
Once conversion tables are obtained, all student interim test scores can be translated into the summative scale scores. Students can then be classified into broad proficiency groups, such as “proficient” or “not proficient,” based on where their converted interim scores fall relative to the summative proficiency cut points.
This categorization can then be compared to the actual proficiency levels determined by the summative assessments, to quantify agreement and discrepancy rates. Statistical methods, such as correlation, can also gauge the degree of association between the aligned test scales.
Let’s look at one such study, which was conducted in collaboration with the Success Ready Students Network (SRSN), a group of Missouri school districts focused on developing next-generation accountability models.
The Study
The goal behind the study was to understand how interim or benchmark assessments administered in schools align with annual summative state assessments used for accountability purposes.
Data from over 15,000 students in 11 SRSN districts were utilized. The study focused on grades 3-8 in math and English language arts. Popular interim assessments like the Northwest Evaluation Association (NWEA) assessment, i-Ready, and Star Renaissance were compared to the Missouri summative assessments (MAP).
Findings
- The strongest relationships were between MAP and the NWEA and i-Ready interim tests in both math and English Language Arts (ELA). Correlations exceeded 0.75 in most cases.
- MAP-NWEA correlations were consistently high, ranging from 0.77 to 0.88 across grades and subjects.
- i-Ready results also showed robust correlations with MAP in the 0.76 to 0.87 range.
- Star had more variability, but still correlated reasonably well with MAP results.
- Categorizing students into “proficient” and “not proficient” groups based on interim results aligned reasonably well with summative benchmarks. Agreement rates reached 85% in some cases. This demonstrates interim assessments can differentiate proficiency similarly to summative tests.
- The link between interim and MAP tests strengthened later in the year. Spring interim scores correlated more closely with MAP compared to earlier administrations.
- The largest proficiency categorization alignment was between MAP ELA and the spring NWEA ELA test (83-84% agreement across grades).
- False positives, where students were rated proficient on interim tests but not MAP, were rare across all test types and grades.
- False negatives, where students were deemed not proficient on interim tests but proficient on MAP, represented the bulk of discrepancies.
- Proficiency thresholds vary across assessments. Interim tests do not mirror summative proficiency levels.
- Using interim results for summative accountability judgments could underestimate or overestimate proficiency for 15-25% of students.
- Layering interim and summative results provides a more accurate picture than relying heavily on one or the other.
- Interim assessments are useful instructional tools but may not be ready to fully replace summative exams for accountability purposes.
With further refinement, states could transition toward through-year proficiency models by using high-quality interim tests developed specifically for accountability purposes. This approach offers flexibility, allows for demonstrating growth, and reduces stress compared to single high-stakes exams.
Of course, interim assessments should not stand alone. Combining adaptive interim testing, classroom work, and focused performance tasks would provide a more holistic picture, superior to any single test.
The Future of Assessment
With thoughtful design, interim assessments could become the foundation of next-generation accountability models, which focus on demonstrating competency rather than point-in-time test performance. This approach holds promise for personalizing learning while maintaining high expectations.
This partnership with SRSN enabled the use of a robust statewide dataset to delve into a policy question with significant instructional and accountability implications. The insights from this analysis help advance the dialogue on competency-based models and the appropriate use of interim assessments. The collaborative effort demonstrates how researchers and practitioners can work together to promote evidence-based improvements for schools.
The time has come to rethink summative exams, which often fall short. Interim assessments open the door to focus on meaningful growth and continuous improvement, rather than instilling fear through once-yearly testing. States and districts should seize the opportunity to create new assessment systems that empower students and educators alike.
Interested in more from Marzano Research about finding classroom solutions that improve outcomes for teachers and students? Visit our Classroom Practice page.
