Assessment Domain - Best Practices

Contents

The Ed-Fi assessment specification is purposefully non-restrictive when it comes to the types of result data than can be sent to the Assessment Outcomes API, so that vendors with different types of data can be accommodated.  While differences exist between assessment platforms, there are enough similarities to warrant a suggested method of returning data to the API for the most common use cases. These suggested methods will help assure consistent data types for the most common use cases.  This document addresses six of those use cases, for typical, point-based assessments.

An assessment’s overall maximum and minimum score

For a typical point-based assessment’s overall maximum and minimum score we suggest an AssessmentReportingType descriptor value of "Number score" and a ResultDataType descriptor value of "Decimal".

The score elements should reflect the total number of points available.  We recommend you round to two decimal places which are always present.  For instance, for an assessment that has a maximum of 50 points, the suggested formulation for the “scores” element in json is:

An assessment’s performance levels

For a typical point-based assessment’s performance levels we suggest an AssessmentReportingMethodType descriptor value of "Raw score" and a ResultDataType descriptor value of "Percentage". We recommend no decimal places.

A unique consideration with performance levels is that the level cutoffs can be either inclusive or exclusive.  We suggest that the MinimumScore be inclusive while the MaximumScore not be considered part of the performance band.  In the example json shown below, this would indicate that a score of 80 would be considered “Proficient”.  In addition, if it is possible for your students to achieve better than 100% on a performance level, we suggest that a MaximumScore value for the top performance level be omitted.

An objective’s scores designation

For a typical point-based assessment objective’s scores designations we suggest an AssessmentReportingType descriptor value of "Raw score" and for the ResultDataType descriptor value we recommend "Percentage", as shown in the example json below.

An objective’s performance levels

For a typical point-based objective’s performance levels we suggest an AssessmentReportingMethodType descriptor value of "Raw score" and a ResultDataType descriptor value of "Percentage". We recommend no decimal places.

A unique consideration with performance levels is that the level cutoffs can be either inclusive or exclusive.  We suggest that the MinimumScore be inclusive while the MaximumScore not be considered part of the performance band.  In the example json shown below, this would indicate that a score of 80 would be considered “Proficient”.  In addition, if it is possible for your students to achieve better than 100% on a performance level, we suggest that a MaximumScore value for the top performance level be omitted.

A student’s final score for the overall assessment

For a student’s final score on a typical point-based assessment we suggest an AssessmentReportingType descriptor value of "Number score" and for the ResultDataType descriptor value we recommend "Decimal".  

For the result, we recommend you round to two decimal places which are always present. The result should reflect the raw score of the student’s final result. For instance, for a score of 37 out of 50 points, the result would be shown as 37.00.  

A student’s final score on an objective

For a student’s score on a typical point-based objective’s performance levels we suggest an AssessmentReportingMethodType descriptor value of "Raw score" and a ResultDataType descriptor value of "Percentage".

For the result we recommend two decimal places, reflecting the percent correct the student achieved for the objective.