Skip to content

How a total becomes a grade and a pass

The rule, written down

StudyDrome Exam Manager turns a total into a grade with a grading scale: an ordered list of levels, each with a minimum and its own pass flag. The scale compares raw points or percentages, rounded to two decimal places. The first level whose minimum is met wins, and that level decides the pass.

Book a pilot

What is a grading scale?

A grading scale is an ordered list of levels. Each level carries four things.

Part

What it holds

Name

The label the result is reported under

Minimum

The value a candidate must reach for that level

Pass flag

Whether reaching this level counts as a pass

Position

Where the level sits when the levels are compared

The scale itself is one of two types. A points scale compares marks. A percentage scale compares a percentage.

Scales are yours to define. They live on the Grading Scales screen in the manager console, and one of them can be set as the default. An exam then names the scale it is graded on.

Which number does the scale compare?

The scale type decides which number is compared.

Scale type

The value compared

Points

The points the candidate earned

Percentage

Earned points as a share of the total, rounded to two decimals

Anything that is not a points scale takes the percentage rule. So the percentage path is the normal behavior rather than a special case.

A paper worth zero points compares as zero. The arithmetic never divides by nothing, and the attempt still resolves to a level.

Which questions count toward the total?

The total is defined by what it leaves out.

Left out

Why

Likert items

They collect opinion, so they carry no key and no points

Voided questions

A flawed item excluded after delivery stops counting on every attempt of that sitting

Ungraded answers

An answer still waiting for a marker has no score to add

Likert items are excluded when the attempt is created, so they never enter the total in the first place. Voiding is a toggle on the published exam. Switch it back, and the earlier score returns exactly.

Partial credit belongs to the question, not to the scale. Question types cover which types award part marks.

How is a level chosen?

The levels are compared in the order you listed them.

  1. Take the comparison value for the scale type.
  2. Walk the levels in their configured order.
  3. The first level whose minimum is met wins.
  4. If no level matches, the lowest level applies.

Step four matters. A candidate under every minimum still gets a named result rather than a blank one.

An exam with no scale attached, or a scale with no levels, returns no grade at all. Nothing is guessed. A missing setting reads as missing, which is the outcome you can act on.

How is pass or fail decided?

The matched level carries the answer. Pass is read from that level's own flag.

So the pass rule is data, not code. It sits in a table you define, it travels with the exam, and it can be produced on request. Nobody has to remember what the threshold was in March.

Two other exams decide differently, and each is its own mechanism.

Exam

What decides the result

An OSCE circuit

The circuit's own conditions, in standard setting

A difficulty-weighted paper

Its thresholds and domain minimums, in difficulty-weighted scoring

On a weighted paper, the scale still names the level. The pass decision moves to the scoring policy.

When is a grade calculated, and when is it recalculated?

The system finalizes grading itself. After every grading action, the finalizer counts the answers that are still awaiting a manual mark.

Answers waiting

What happens

None, on a submitted attempt

The total is computed, the level and the pass are set, and the attempt becomes graded.

None, on an attempt already graded

The scores are recalculated in place

One or more, on an attempt already graded

The attempt returns to submitted

The last row is the safeguard. Un-grade one answer and the attempt steps back with it, so a published level always reflects a complete mark. The teaching-staff view covers re-marking from the marker's perspective.

Difficulty weighting runs after finalization, on exams that have it switched on.

What does publishing freeze?

Publishing copies the grading scale into the exam's snapshot. Every level goes with it, and so do the level translations.

That is what makes a grade reproducible. Edit the scale for next semester, and leave last semester's results where they are because they were graded against the copy.

It also means the rule is retrieved rather than reconstructed. Ask what counted as a pass on a paper sat two years ago, and the answer is a stored record. How publishing freezes an exam covers the wider snapshot, and defending a result covers the records an appeal asks for.

Which settings sit around grading?

The grading side of an exam carries these settings. Each one is editable on a live exam from the Grading & Results drawer, with no republishing.

Setting

Options

Ships as

Teacher grading

Disabled, Always Allowed, Time-Based

Disabled

Grading deadline

After Student Submits, After Test Availability Ends

—

Grading Days Limit

Several days

—

Teacher attempt visibility

Latest Attempt Only, All Attempts

Latest Attempt Only

Student names shown to graders

On or off

Off

Exam Manager for teaching staff covers how each one behaves for the people marking.

Notes for a reviewer

A grade is only as defensible as the rule behind it. These are the properties that decide whether that rule can be produced later.

  • The pass rule lives in a table you define, and it travels with the exam.

  • A grade level can be reconstructed from the published scale and total.

  • The total is defined by what it excludes, and every exclusion is published here.

  • Exam Type is a reporting label. It changes nothing about how a paper is scored.

  • A missing scale produces no grade rather than a guessed one.

  • Three mechanisms can determine the result, and each one is identified in the report.

Frequently asked questions

Where is the rule that decides a pass?

In the grading scale attached to the exam. Each level in that scale carries its own pass flag, and the level a candidate reaches supplies the answer. The scale is copied into the exam's snapshot at publish, so the rule that applied to one sitting stays retrievable long after the scale itself has been edited.

Does editing a grading scale change grades already awarded?

No. Publishing copies the scale and all of its levels into the exam's snapshot, and grading runs against that copy. Edit the scale afterward, and the change applies to exams published thereafter. Results already awarded keep the levels they were graded against, which is what makes an old grade reproducible.

Can a grading scale compare raw points rather than percentages?

Yes. A points scale compares the marks a candidate earned, so each level minimum is written in points. A percentage scale compares earned points as a percentage of the total, rounded to 2 decimal places. The type is set on the scale itself, so every exam using that scale compares the same way.

Which questions are left out of the score total?

Likert items, voided questions, and answers are still waiting for a mark. Likert items have no key or points and are excluded when the attempt is created. Voiding removes a flawed published question from every attempt of that sitting, and switching the toggle back restores the earlier score exactly.

Book a pilot

Bring the grading scale you use today. We will build it in a sandbox and grade a past paper using it so that you can check each level against your own results.

Book a pilot

Related