How can secondary teachers cut marking time without weakening feedback?

Why does marking take up so much of a secondary teacher's week?

Marking dominates the week because a normal secondary load is four or five classes, each producing drafts, extended responses and assessment tasks that need individual, criteria-based comment. One senior English class can generate a full set of essays in a single week. Multiply that across year levels, and often across two subjects, and the pile stops being an event and becomes a permanent condition.

The expectation is not a tick and a grade either. Under the Australian Curriculum, work is judged against achievement standards, and senior courses add authority-specific rubrics from NESA, VCAA, QCAA, SACE and SCSA. Naming the criterion, explaining the gap and pointing to the next step is skilled writing. Doing it fairly for thirty students in a row is what consumes the hours.

This matters because the answer is almost never to mark faster by force of will. Speed applied to an unchanged process produces thinner comments and a standard that slides. The habits behind expert essay marking are mostly decisions taken before the pile arrives, and most of them cost nothing to adopt.

Which decisions before the task shape how long marking takes?

Write the rubric before you write the task. If you cannot describe what a top-band response does differently from a middle-band one, marking will be slow, because you will be inventing the standard while applying it. Drafting the criteria first also exposes questions that are unmarkable as written, which is much cheaper to discover in week one than in the marking pile.

Keep the number of criteria small. Four or five criteria with clear descriptors are faster and more reliable than a grid of twelve rows that all say roughly the same thing in different words. Every extra row is another judgement per student, multiplied by the class, and rows that overlap force you to make the same decision twice.

Then give students the criteria and an annotated exemplar before they start. Work written against a known standard is quicker to mark because fewer responses miss the task outright. Handing out the rubric, exemplars and any study guides in advance shifts effort from your marking to their drafting, which is where the effort does more good anyway.

What does a faster marking routine actually look like?

Mark one criterion at a time across the whole class rather than one script at a time. Marking vertically keeps a single standard in your head, so you compare like with like and reach each decision faster. It also makes borderline cases obvious, because near-identical responses end up sitting next to each other instead of forty minutes apart.

Read the class quickly before you write anything. A five-minute skim of all thirty scripts gives you the range before you commit a mark to the first one, so you are far less likely to re-mark the early scripts once your standard settles. The skim also surfaces the errors that repeat, which are the ones worth handling once for everybody.

Timebox the session. Decide in advance what a Year 9 report or a Year 11 essay is worth in minutes, time the first five scripts, and use that average to plan the rest. The first few always take longer than the rest. Knowing this by measurement rather than by feel is what stops a marking session from expanding to fill the entire evening.

• Shuffle the order each task, so the same students are not always marked last when your patience is thinnest.
• Skim the full set before marking any of it, and write down the three errors that repeat.
• Mark criterion by criterion, not script by script.
• Keep the comment bank open and paste from it, editing only the sentence that is specific to that student.
• Moderate a sample with a colleague once per assessment, rather than once per term.

How do comment banks and whole-class feedback save real time?

Most of what a teacher writes in a marking session is not unique. The same three or four problems recur: the thesis does not answer the question, evidence is quoted but never explained, the conclusion restates the introduction. Composing that explanation from scratch thirty times is the largest avoidable cost in the whole task, and it is the part students value least, because they all receive roughly the same paragraph.

A comment bank fixes it. Write the best version of each recurring comment once, with the criterion named and a next step attached, then reuse it. The bank should be in your words and it should grow every term. Ten well-written comments will cover most of a class set, and the two minutes you then spend on the student-specific sentence is the part they actually remember.

Anything that applies to more than a third of the class belongs in whole-class feedback, not in thirty margins. Teach the fix in ten minutes at the start of the next lesson, put a weak and a strong example side by side on the board, and mark only the individual variation on the page. Students act on a correction that was taught far more often than on one they read alone.

How do you keep standards consistent when you mark faster?

The real risk in a large pile is drift, where script fifty is judged on slightly different terms from script five. Drift is not caused by speed. It is caused by holding the standard in memory instead of on paper. A written rubric applied one criterion at a time is the main defence, and it works whether you mark in one sitting or across a fortnight.

Anchor the scale before you commit to anything. Pull three scripts at random, place them as top, middle and bottom, and mark those three slowly and carefully. They become the reference points for everything else. Re-read them at the end of the session, and if your standard has moved, adjust the whole set rather than pretending the first ten were marked on the same basis as the last ten.

Then moderate a sample. Swap five scripts with a colleague teaching the same course and compare marks against the same descriptors. Twenty minutes of moderation buys more reliability than an extra hour of individual marking, and it is the evidence you want in hand when a student or a parent queries a result.

Where can an AI marking assistant help, and where should it not?

Everything above works with a pen and a printed rubric. Software earns its place only on the repetitive layer: applying one set of criteria evenly across a class, flagging which success criteria a response meets or misses, and drafting the routine comments so you edit rather than compose. That is the honest scope. No tool can settle a borderline band, weigh a student’s progress over the term, or know that a student sat the task the week after a family emergency.

That repetitive layer is what Jeddle was built to handle. You load the rubric, success criteria and comment bank you already use, JeddAI drafts a first-pass mark and feedback against them, and you review, edit and confirm every result before a student sees it. Because the criteria are yours, the drafts follow your wording and your standards, and the same setup carries across NESA, VCAA, QCAA, SACE and SCSA courses rather than assuming one national template.

Two rules are worth keeping whatever tool you use, or whether you use one at all. Never let a mark reach a student without a teacher having read it, and never adopt a rubric you did not write or check line by line. Assessment authorities hold the teacher accountable for the judgement, not the software. Students can also tell the difference between a comment written for them and one that was not, which is exactly why the time you save on the routine layer should go back into the comments that only you can write.

• The tool drafts: a first-pass mark against your rubric, and routine comments in your comment-bank wording.
• The tool flags: which success criteria a response meets or misses, so you scan the evidence instead of hunting for it.
• You decide: the final mark, every borderline case, and any comment that needs context about the student.
• You own: the rubric, the standards applied, and the tone of what the student finally reads.