D638 Monitoring Student Learning Through Assessment, catalog number EDUC 5078, is the two-CU graduate assessment course in the WGU Master of Arts in Teaching sequence. It covers formative and summative assessment, benchmark testing, progress monitoring and alternative assessment, and it covers the same ground as the undergraduate D659 (EDUC 2221). The graduate addition is the scoring tool. At this level you are expected not only to choose an assessment but to build the instrument that turns student work into a defensible judgment.
Where EDUC 5078 goes past the undergraduate course
Alternative assessment is the phrase that changes the work. Once a task is a performance, an exhibition, a portfolio or a project, the correctness of a response stops being obvious, and the burden shifts onto the criteria. Graduate assessment work is scored substantially on whether your criteria would produce the same judgment in two different hands.
That is a harder problem than it looks. Criteria written as adjectives fail it immediately. Excellent analysis, thorough explanation and strong evidence all describe an impression rather than a feature, so two markers using them will disagree. Criteria written as observable features survive: identifies at least two causes and explains the relationship between them, or cites the source of every figure used in a claim. The second kind can be checked; the first can only be felt.
The second graduate expectation is systems thinking. An undergraduate plan can treat assessment lesson by lesson. A graduate plan is expected to describe how classroom checks, unit tests and periodic benchmarks fit together without duplicating each other, and what each layer is for. Candidates who write three assessments that all measure the same thing at the same grain have described redundancy rather than a system.
The third is fairness under scrutiny. Graduate work is expected to reason about bias in items and tasks, about the difference between measuring content and measuring reading speed, and about what accommodation preserves the construct being measured. Those are technical questions with technical answers, and they are where thin submissions become visible.
A worked example makes the distinction concrete. Suppose the target is whether students can justify a claim with evidence, and the task is a written argument. A dense passage of source text now sits between the student and the target, so for a student two years below grade level in reading the task measures decoding as much as reasoning. Reading the sources aloud, or supplying them at an accessible level while keeping the reasoning demand identical, preserves what you meant to measure. Shortening the number of sources to be compared does not, because comparison across sources was part of the reasoning itself. Two accommodations, superficially similar, and only one of them leaves the score interpretable. Writing that difference down is close to the whole of a fairness aspect.
Converting the rubric into an assessment document
Open your Course of Study and count the scored aspects before drafting. Each is scored alone against a three-point scale and each needs a 2 to pass. Assessment courses tend to have aspects that pair a build obligation with a justify obligation, and the justification half is the one candidates truncate.
If D638 is a performance assessment for you, use the rubric's nouns as your headings and answer in the rubric's order. Where an aspect asks you to create something, the created artifact goes in the document, in full, and the explanation of it goes next to it. An attachment referenced but not integrated tends to be scored as though it were missing.
A worked word budget. Assume six scored aspects and directions pointing at roughly 1,750 words of narrative plus the instruments themselves. Take 110 words for the class and unit context, and 90 for the close. That leaves 1,550, or about 258 words per aspect before adjustment.
Adjust so the reasoning aspects are funded. The aspect on validity or fairness deserves close to 350, and the aspect on interpreting results and responding deserves the same, because both are arguments rather than descriptions. Take the difference from the aspects that ask you to identify the assessment type or list its features, which sit comfortably at 160. The instruments themselves usually sit outside the narrative count, so check the directions before assuming your tables are free.
A shape that fits an assessment system
Where directions specify sections, use theirs. Where they do not, this arrangement shows the system rather than a pile of instruments.
| Layer | Purpose it serves | Where candidates slip |
|---|---|---|
| Targets | The specific outcomes the system is accountable for | Targets broad enough that nothing could fail to measure them |
| Classroom checks | Fast, frequent, actionable within the same lesson | Checks that generate scores nobody uses |
| Unit assessment | Aggregated evidence at the end of a sequence | Format chosen before the target was settled |
| Alternative task | Performance evidence that a selected-response test cannot produce | A task that is engaging and measures nothing specific |
| Scoring tool | Observable criteria at each level, usable by another marker | Adjective ladders that describe impressions rather than features |
| Benchmark and monitoring | Comparable measures over time with a decision rule | Repeated tests with no rule about what a trend triggers |
| Fairness provisions | Accommodations that preserve what is being measured | Supports that quietly remove the skill under assessment |
Test your scoring tool before submitting it. Score two imagined pieces of student work with it, one strong and one middling, and see whether the tool forces the judgment or merely permits it. If you could justify two different scores for the same imagined paper, the criteria still describe impressions.
Evidence craft when you are building the instrument
Graduate assessment work is evaluated partly on the technical vocabulary it uses correctly, so sourcing and precision matter together.
- Use measurement terms in their real senses. Validity is about the inference, reliability is about consistency, and neither is a synonym for quality.
- Support fairness claims with published guidance rather than intuition, especially anything touching multilingual learners or documented accommodations.
- Where you adapt an existing instrument or rubric, cite the original. Adapted is honest and unattributed is not.
- Present sample results as illustrative when they are constructed, and let them include a result that complicates your story.
- Keep all student references anonymous and non-identifying, including combinations of details that would identify a class.
- APA for every external source, including state assessment guidance and any commercial screening tool.
The line that lifts these papers is a statement about what your system will miss. Every assessment design is a sampling decision, and naming what the sample leaves out, then saying which layer would catch it later, reads as measurement literacy rather than as a hole.
What separates Competent from a returned assessment plan
Because each aspect is scored independently, returns are narrow. In this course the recurring one is a scoring tool whose levels cannot be told apart by an outsider.
- Every criterion in the scoring tool names something observable in the student work.
- Adjacent performance levels differ by a stated feature, not by an intensifier.
- Each layer of the system measures something the other layers do not.
- Every accommodation preserves the construct rather than replacing it.
- Every interpretation of results is attached to a rule set before the results appeared.
Work is recorded as Competent or Not Competent, without letter grades or a conventional grade point average, and performance assessment work can be revised and resubmitted with no grade penalty. So the price of a return is calendar time inside a six-month flat-rate term, and since closing more courses in a term is what lowers your effective cost per course, avoidable rework is the most expensive thing in the program.
If your D638 includes a proctored objective assessment, our role is preparation only: drilling the measurement distinctions, running practice interpretation items, and giving an honest readiness read. We never sit assessments and never request credentials.
Six mistakes that cost time in D638
- Writing rubric levels as adjective ladders. Excellent, good, fair and poor is not a scoring tool, and this is the single most common graduate assessment return.
- Building an alternative task with no defined evidence. If the task is fun and the criteria are vague, the assessment measures compliance.
- Duplicating layers. Three instruments measuring the same target at the same grain is redundancy, and one of them is probably the wrong tool.
- Accommodating the construct away. Reading a reading test aloud measures listening. Say what the accommodation preserves.
- Treating benchmarks as bigger unit tests. Benchmarks exist to compare across time and groups, which places different demands on their design.
- Leaving the scoring tool untested. Score two imagined papers with it before you submit. Most weak tools reveal themselves in five minutes.
How support works on D638
Send the rubric out of your Course of Study, the task directions, and the subject and grade band you want the system built for. The draft returns aspect-mapped, with a scoring tool written in observable features, layers that each do distinct work, and constructed results that behave like real ones.
If you are taking the graduate planning course near this one, send both rubrics together. Assessment and unit design are the same argument seen from two ends, and building them in one pass usually saves a week over building them separately.
Three questions MAT candidates ask about D638
Is D638 the same course as EDUC 5078?
Is D638 the graduate version of D659?
What makes a rubric good enough for a graduate assessment course?
Building a scoring tool that would survive a second marker?
Send the rubric and task directions. You get an aspect-mapped draft with criteria written as observable features, layered instruments that do distinct work, and realistic sample results.
Where D638 sits in WGU's programs
The July 2026 catalog places this code in 9 current WGU programs. Open a program page for the complete standard path and term positions. The live Degree Plan remains authoritative after transfer credit, substitutions, and mentor planning.
The assessments, one by one
The public catalog does not publish this course's PA/OA identity or task count. WGU Tutors publishes at most one PA manual per course and only from a WGU-controlled public rubric. Until that source exists, PA help begins from the student's real Course of Study and OA support remains preparation only.