Build a 4-point standards-based grading scale by anchoring Level 3 to the grade-level standard, then writing observable descriptors for the levels above and below it. Define each level by what students can do, keep the language consistent, and grade against a body of evidence rather than averaging.
What is a 4-point standards-based grading scale?
A 4-point standards-based grading scale is a fixed set of four levels that describe how fully a student has met a specific learning standard, rather than how many points they earned. In standards-based grading (SBG), each level, usually numbered 1 to 4, is tied to a written descriptor of observable understanding, so a score reports mastery of the target instead of a running tally of assignments.
The scale also separates what a student knows from behaviors like effort, neatness, or turning work in on time, which are reported elsewhere. And it decouples grades from percentages: a 2 means partial mastery, not a failing 50 percent. Because every teacher uses the same four descriptors for the same standard, scores become far more consistent across classrooms and easier for families to interpret.
What does each of the four levels mean?
Each level describes a different depth of understanding relative to the grade-level standard, with Level 3 defined as full proficiency. Level 4 shows understanding that reaches beyond the target through greater complexity or transfer; Level 3 meets the standard independently; Level 2 shows partial or developing command of simpler content; and Level 1 is a beginning attempt that still needs significant support.
The most important idea is that Level 3, not Level 4, is the goal. A student who consistently meets the standard has done exactly what the grade expects, so Level 3 should feel like a genuine success, not a near miss. Level 4 is reserved for reasoning or application you did not explicitly teach, which keeps the top of the scale meaningful rather than a reward for extra volume.
How do you write the level descriptors, step by step?
Start by writing Level 3, because it defines proficiency and anchors everything else. Unpack the standard into what a proficient student can actually do, then phrase it as a clear, observable statement. Once Level 3 is precise, the remaining levels are written in relation to it rather than invented separately.
From that anchor, describe Level 2 as partial success on the simpler or foundational parts of the standard, and Level 1 as an emerging attempt that depends on teacher support. Write Level 4 last, as deeper reasoning, transfer to a new context, or more sophisticated application, never simply more questions answered. Use consistent stems such as “I can” so students recognize the same structure on every scale.
- Unpack the standard into the specific skills and knowledge it requires.
- Write Level 3 first as an observable statement of full, independent proficiency.
- Define Level 2 as partial mastery of the foundational or simpler content.
- Define Level 1 as an emerging attempt that still needs significant support.
- Write Level 4 last as transfer, complexity, or reasoning beyond what was taught.
- Use student-friendly, consistent language, for example "I can" statements.
How do you keep Level 4 from turning into extra credit?
Define Level 4 as a different kind of thinking, not a larger amount of work. The trap is describing it as “completes all questions correctly” or “adds extra examples,” which rewards volume and speed rather than deeper understanding. Instead, Level 4 should capture transfer, applying the skill in an unfamiliar context, or reasoning about why and when a concept holds.
This distinction protects the meaning of proficiency. If Level 4 just means more of the same, students learn that Level 3 is second best and go back to chasing points, which is exactly the behavior standards-based grading tries to retire. A well-written Level 4 is something a student reaches through insight, not by doing double the assignments, and many students may finish a unit securely at Level 3.
How do you grade consistently against a proficiency scale?
Grade against a body of evidence and the most recent, most consistent performance, not a running average. Averaging punishes early struggle and hides real growth, so a student who ends the unit at Level 3 should be recorded at Level 3 even if they started at Level 1. Look for a pattern across several tasks before you assign a level.
Consistency also depends on calibration. Before grading, review a few sample responses with your team and agree on which level each one shows, so the same work earns the same score in every classroom. Keep the descriptors in front of you while you grade, and score one standard at a time rather than forming a single holistic impression of the whole assignment.
- Score against the descriptors, one standard at a time, not a holistic impression.
- Weight the most recent and most consistent evidence over early attempts.
- Avoid averaging, which buries growth and can turn a 2 into a failing grade.
- Calibrate with colleagues on sample work so a level means the same thing everywhere.
How do you report a 4-point scale to families and report cards?
Report the level with its descriptor, not a converted percentage, so families see what the number actually means. The common mistake is mapping 4/3/2/1 onto 100/75/50/25, which turns solid partial understanding (a 2) into a failing grade and reintroduces the point-chasing SBG was meant to remove. Instead, publish a simple key that pairs each level with its plain-language meaning.
If your school still requires a letter or percentage grade, decide the conversion deliberately and share it up front, treating Level 3 as clearly passing, proficient work. Pair every reported level with specific feedback on the standard so students and parents know the next step, not just the label. Reporting behaviors such as effort separately keeps the academic score clean and trustworthy.
How can a tool like JeddAI help you apply a proficiency scale consistently?
A tool like JeddAI helps by applying your own proficiency scale evenly across a stack of student work while you stay in control. JeddAI is a feedback and grading assistant, not an autopilot: you connect or upload student work, and it drafts feedback and grading aligned to your own rubric, success criteria, and comment banks, so the same Level 3 descriptor is applied the same way to every student.
That consistency is where a proficiency scale usually breaks down, because it is hard to hold four descriptors steady across a hundred papers. Working from a first-pass draft aligned to your levels, you spend your time reviewing and refining judgments instead of re-reading the rubric, and you save grading time without giving up the final call. If that fits how you grade, you can Get started with JeddAI and keep the scale consistent from the first paper to the last.
| Level | What it signals | What it looks like in student work |
|---|---|---|
| Level 4 — Advanced | Understanding beyond the grade-level standard | Transfers the skill to a new context or reasons about why it works |
| Level 3 — Proficient (the goal) | Meets the standard independently | Consistently does what the grade-level target asks, unaided |
| Level 2 — Developing | Partial mastery of simpler content | Handles foundational parts but not the full standard yet |
| Level 1 — Beginning | Emerging understanding | Attempts the task but needs significant support to progress |
Frequently asked questions
Should a 4-point scale include half-points like 3.5?
Some frameworks, such as Marzano's, add half-points to show partial progress between levels, which can sharpen feedback. Keep it optional and consistent; for many teachers, four clear levels are easier to score reliably than eight.
How is standards-based grading different from a traditional rubric?
A traditional rubric often scores a single assignment across several traits, while a proficiency scale tracks mastery of a standard across many pieces of evidence over time. The scale follows the learning target, not one task.
Can I use a 4-point scale in every subject?
Yes. The same four levels work across subjects because they describe depth of understanding relative to a standard. Only the Level 3 descriptor changes to match each subject's grade-level expectation.
What do I record when a student shows no evidence?
Record it as missing or "no evidence yet," not as a zero on the 4-point scale. A zero distorts the record; the aim is to note that the standard has not been demonstrated yet so the student can still show it later.
How many standards should one scale cover?
One. Each proficiency scale should describe a single standard or tightly related learning target, so a score points to a specific skill. Bundling several standards into one scale makes the resulting level ambiguous.
Get started with Jeddle
Jeddle gives teachers and students instant, syllabus-aligned feedback powered by JeddAI.
Looking for study material? Browse Jeddle's Australian-English subject resources, or explore more articles on Feedback & Assessment.
