Put two answers on Bandura’s aggression study side by side. The first is fluent, confident, well-written — and quietly wrong about who imitated whom under which condition. The second is drier, but every detail of the procedure is accurate. On a knowledge question, the second script should score higher, and a marker reading for fluency at 10pm can easily hand the marks the wrong way. Now turn the page to an evaluation question, where a student has listed five true things about the study without ever building them into a judgement — and the trap reverses: this time you’re rewarding a levels-of-response answer that never climbed a level. Cambridge International A Level Psychology (9990) marking lives in that gap, because a single paper asks you to switch between two entirely different marking styles, question to question.

This guide is about marking 9990 the way the Cambridge scheme actually works — crediting accurate knowledge and research-methods points precisely, judging evaluation and study design against level descriptors fairly — and where letting software hold the point-marked parts steady frees you to spend your attention on the judgement the scheme reserves for a human.

The two marking styles inside one 9990 paper

Cambridge assesses 9990 across its written papers — approaches and core studies, research methods, and at A Level the applied specialist options — and the marking blends two logics that behave very differently under a tired eye. (Treat the exact paper structure, durations and weightings as things to confirm against the current syllabus rather than numbers to quote from memory.)

• Point-marked knowledge and research-methods items. “Describe what the researchers found,” “outline one strength of the case study method,” “calculate the mean of these scores,” “state the sampling technique used.” These have creditable points defined in the scheme. A point is either made accurately or it isn’t, and some of the research-methods items carry a genuinely right-or-wrong element — a mean, a range, the correct reading of a scattergram, the correct identification of an independent variable.
• Levels-of-response items. “Evaluate the use of the laboratory experiment in this study,” “discuss the extent to which the study is reductionist,” “design a study to investigate…” These are marked against level descriptors, where the mark depends on the quality of the argument, the range of issues raised, and — crucially — whether points are used to build an evaluation rather than merely stated.

The skill in marking 9990 is not applying one scheme; it’s switching cleanly between the two, many times per script, without letting the habits of one contaminate the other.

Where 9990 marking drifts — and why it isn’t carelessness

Be honest about the 24th script. On the knowledge questions, fatigue makes you generous with a confident-sounding answer and stingy with an accurate but plainly written one — the halo of good expression bleeding into marks it shouldn’t touch. On the evaluation questions, the opposite creep sets in: by mid-pile you stop tracing whether a point is used and start rewarding presence, so a well-organised list of accurate observations drifts up into a band it hasn’t earned, while a rougher answer that genuinely reaches a judgement gets under-read.

None of this is a competence problem. It’s the predictable result of asking one human to run two marking logics across a stack of scripts in a single sitting. You can mitigate it — mark question-by-question across all scripts, keep the level descriptors open beside you, re-read borderlines — but you can’t fully remove it, because the limit is attention, not effort. It’s the same drift the generic parent guide describes for every subject, marking to the Cambridge mark scheme online for class consistency; 9990 just makes it sharper, because the two styles pull a marker in opposite directions on the same page.

What “marking to the scheme online” changes for 9990

When the point-marked parts of 9990 are marked online against the Cambridge scheme, the same creditable points are recognised on the last script as reliably as the first. An accurate description of a core study’s procedure scores on its merits, not on how well it reads. A correctly calculated mean or a correctly identified variable is marked right-or-wrong the same way every time. The research-methods knowledge — sampling techniques, types of validity and reliability, ethical guidelines, experimental designs — is exactly the kind of defined, point-based content where consistency beats tired hand-marking.

The honest scope: this is a consistent first pass, strongest on the knowledge and structured research-methods items. The levels-of-response questions — the evaluations, the “discuss” and “to what extent” items, and above all the design a study questions where a student proposes a method, a sample, controls and an ethical safeguard — still want your eyes. Judging whether an argument reaches a level, or whether a proposed design is actually workable and ethical, is a reading task, not a matching task. Treat automated marking there as a starting point you review and override. That review step is the whole difference between a tool you trust and one you don’t.

A 9990-specific marking workflow

1. Let it mark the point-marked knowledge to the scheme. Descriptions of core studies’ aims, procedures, findings and conclusions; definitions of methods terms; the right-or-wrong research-methods items including calculations and data reads. These get the same points every time, across the class.
2. Check accuracy is being rewarded over fluency. Spot-check a few knowledge answers that read well to confirm they were actually correct, and a few plain ones to confirm accurate detail scored. That’s where students feel marking is fair or unfair.
3. Read the evaluation and design questions yourself. Take the automated mark as a first pass, then judge against the level descriptors — is the point used to evaluate, is the study designable and ethical — and override where needed.
4. Glance at every total near a grade boundary. A couple of misjudged evaluation marks can move a grade. Consistency on the point-marked bulk makes these rarer; never skip the check.

Why consistent 9990 marking matters beyond the time saved

The time argument is real but secondary. The bigger payoff is that your data becomes trustworthy. When the knowledge and research-methods items are marked to the same standard across the class, a weakness your analytics surface — say, the group consistently losing marks on validity, or muddling repeated-measures and independent-measures designs — is signal, not an artefact of you marking that question last and hardest. You can re-teach with confidence. Inconsistent marking adds noise that sends you chasing problems that aren’t there.

It also makes your marks defensible. When a parent asks why one child scored below another on similar-looking evaluation answers, “the same level descriptors were applied to both, and the difference is that one built an argument and one listed points” is a stance you can hold. For giving that feedback at class scale, see examiner-style feedback to 30 students at once.

How this looks on the platform

Tutopiya’s Cambridge A Level Psychology 9990 resources mark the point-based knowledge and research-methods questions against the Cambridge scheme — consistently across every script — while flagging the evaluation and design-a-study items as a first pass you review and finalise. Because the point-marked bulk is level across the class, the topic-level analytics built on it are trustworthy. It’s free to start with one class, no school sign-up. You can also see the whole teacher platform these guides put to work.

This is one of four 9990 guides for teachers. The others cover the 9990 past-paper question bank, building a 9990 mock exam from past papers, and 9990 lesson resources mapped to the syllabus.

FAQ

Can automated marking handle both the knowledge and the evaluation questions in 9990? It handles them differently. The point-marked knowledge and structured research-methods items — descriptions of core studies, methods definitions, calculations, data interpretation — are a strong fit for consistent automated marking. The levels-of-response evaluation and “design a study” questions are best treated as a first pass you read and finalise, because judging the quality of an argument or the workability of a proposed design is a reading task.

How is marking Psychology different from marking a numeric subject online? 9990 isn’t a method-mark subject. Most of the credit is either point-marked knowledge or levels-of-response judgement, with only a limited amount of right-or-wrong calculation in the research-methods items. So there’s no method/accuracy framing as in maths — the split that matters is point-marked versus level-descriptor marking, and the value of consistency is holding both steady across a class.

Does it mark the research-methods calculations and data interpretation? The genuinely right-or-wrong parts — a mean, a range, identifying a variable or sampling technique, reading a scattergram — mark cleanly and consistently. Where a research-methods item asks students to justify a choice or design an investigation, that shifts into levels-of-response territory and belongs in your review.

Will it reward a fluent but inaccurate description of a core study? Marking to the scheme credits accurate points, not polished writing — which is exactly the drift a tired human marker is prone to. You should still spot-check well-written answers to confirm the detail is actually correct, since that’s where students most feel marking is fair.

Do I lose control of the marks? Only if you choose a tool without a review step. The model that works is consistent-first, teacher-final: the point-marked knowledge and methods items marked uniformly to the scheme, and you reading and overriding the evaluation, discussion and study-design answers.

The bottom line

Marking 9990 well means running two marking logics cleanly — crediting accurate knowledge and research-methods points precisely, and judging evaluation and study design against level descriptors — which is precisely what a tired marker can’t sustain across a full class set. Let consistent online marking hold the point-marked parts steady, keep your judgement for the evaluation and design questions, and your marks become both fairer to students and trustworthy as data.

Mark your 9990 class to the scheme — consistently, free with one class →

Ready to Excel in Your Studies?

Get personalised help from Tutopiya's expert tutors. Whether it's IGCSE, IB, A-Levels, or any other curriculum — we match you with the perfect tutor and your first session is free.

Book Your Free Trial