Assessment Domain - Overview

Contents

Key Entities

The key entities in the Assessment domain are:

• An Assessment entity that describes assessments along with associated metadata.  The ObjectiveAssessment entity describes the structure of the assessment subcomponents or subtests that are separately scored or administered. The AsessmentItem represent the individual questions for cases when the student's score on each question is reported.
• A LearningStandard entity that drives the curriculum and the assessments.  The ObjectiveAssessment and AssessmentItem may reference the learning standards which they assess.
• A StudentAssessment entity that holds assessment scores/results and administration attributes for a student's attempt at the assessment.  

Key Concepts

The assessment model is flexible enough to convey results from complex tests such as the SAT and ACT exams, large summative assessments such as state-level standardized tests, formative assessments, and simple tests such as classroom benchmark assessments or quizzes.

The key concepts include the following:

• The overall domain is primarily designed to collect and exchange data that is useful in the interpretation of student assessment results at a classroom and school-level. The domain is therefore not designed to deliver on use cases such as:

  • to recreate assessment instruments fully or provide for assessment or assessment item portability across systems
  • to capture precise algorithmic or technical details of scoring systems
  • to gather all data generated from a student's interaction with an assessment instrument.
• On this last point, note that there is some focus on gathering student inputs during the assessment process, but those are generally limited: the assumption is that a  "replay" or more full account of student interactions (for example with a complex question or technology-enabled item) would be better accomplished by some facility on the host system and that therefore the focus is on providing for such linkages to other systems in these cases.
• The LearningStandard entity models the hierarchical organization of learning standards for skills and comprehension that are used to drive instruction and assessment.
• The ObjectiveAssessment entity reflects a subset of items on an assessment that may have its own score, performance levels, or academic subject and that can be tied to learning standards.  Objective assessments, when used, reflect the subcomponents or subtests of the assessment, and may be arranged hierarchically into multiple levels as required.

Parallel Assessment Metadata and Student Results Data

The Ed-Fi data model for assessments have parallel structures, depicted below, to:

1. Define the metadata for the assessment in terms of its structure and key attributes
2. Provide the results from students' taking the assessment

Every assessment will minimally have an Assessment instance providing the assessment metadata.  Similarly, every attempt by a student taking the assessment will minimally have a StudentAssessment instance, providing key attributes about the attempt and the highest-level assessment score(s).

If an assessment is organized into subcomponents or subtests that score specific skills or competencies, the structure is defined by ObjectiveAssessment instances that may be organized into a single sublevel, or may be organized hierarchically, as needed.  For each ObjectiveAssessment that is defined in the metadata, the student's results are reflected in  separate parallel instances of the complex attribute StudentAssessment.StudentObjectiveAssessment.

If an assessment reports results on individual questions, the metadata defining those questions are reflected as AssessmentItem instances.  For each AssessmentItem that is defined in the metadata, the student's results are reflected in parallel instances of the complex attribute StudentAssessment.StudentAssessmentItem.

Hierarchical Model

The Ed-Fi data model for assessments contains a number of entities and reference patterns. While its scope may make it a bit imposing to newcomers, understanding a few design principles enable it to be easily understood and adopted.

Assessment and ObjectiveAssessment Example

Many assessments are multi-tier in the sense that they provide multiple scores or result sets for each assessment. An example would be a single "reading" assessment that tested multiple skill areas, such as "Reading Comprehension," "Accuracy and Fluency," "Phonemic Awareness," and so forth.

In the Assessment domain, the top-level assessment is an Assessment entity and the skill areas are ObjectiveAssessment entities. This structure is recursive, so that there can be any number of levels of ObjectiveAssessments.

Once a student takes an assessment, the results can be modeled in the StudentAssessment entity with parallel StudentObjectiveAssessment complex attributes, each of which has references back to its parent or peer entities.

Support for Mapping to Local Learning Standards

Field implementation has shown that, while school systems will intake into their systems the results of student assessments in areas like "Reading Comprehension" (in other words, the ObjectiveAssessment entity with student results held in the StudentObjectiveAssessment entity), they also commonly need to map those "Reading Comprehension" results to learning benchmarks. The LearningStandard entity enables the ObjectiveAssessment to be mapped to an education standard. Further, the LearningStandard entity may be a locally defined benchmark or a formal, external learning standard, such as one provided by the state or Common Core State Standards.

The structure looks like this:

The source of the data is typically as follows:

Sample Mapping

For assessment vendors, it may be helpful to review a sample score report and see how the individual elements would map to the Ed-Fi Assessment model. The following is a fictitious math assessment score report, Math Whale followed by a breakdown of where we would expect the different data points to land within the Assessment model. 

For assessment vendors, it may be helpful to review a sample score report and see how the individual elements would map to the Ed-Fi Assessment model. The following is a fictitious math assessment score report, Math Whale followed by a breakdown of where we would expect the different data points to land within the Assessment model. Additional information for Assessment Vendors can be found in the

A Sample Score Report

A typical classroom score report (for a fictitious provider) might look like this:

Figure 1: a sample score report (click to expand)

A mapping of the main elements might look like this (you will see a few new elements appear here):

Figure 2: mappings for sample score report (click to expand)