A Mistake List Is Not a Diagnosis
“Grammar error”, “poor time management”, “weak vocabulary” and “off-topic” describe visible outcomes. They do not yet tell us what failed first. This page owns the diagnosis of recurring PSLE English mistakes; the separate Improve PSLE English Performance guide owns the longer baseline → intervention → retest → transfer cycle after the likely cause has been identified.
A useful diagnostic sequence is:
Symptom → Earliest Cause → Direct Test → Targeted Repair → Retest.
1. Start With the Failure Signature
| Visible symptom | Test first |
|---|---|
| Answers the wrong thing | Task interpretation |
| Knows passage but cannot answer inference | Evidence-to-inference link |
| Composition wanders | Plan selection and causality |
| Good ideas, unclear sentences | Sentence encoding |
| Vocabulary disappears in writing | Meaning-to-word retrieval under load |
| Repeated late grammar errors | Automation under sustained load |
| Runs out of time | Where drift begins, not just final speed |
| Strong at home, unstable in tests | Transfer and examination load |
2. Test the Earliest Plausible Cause
Suppose a student writes a weak composition ending. The ending may not be the real problem. The plan may have become too large ten minutes earlier, the climax may not create a consequence, or time may have drifted before the final paragraph began.
Move upstream until you find the first layer that breaks.
3. Separate Knowledge Failure From Load Failure
A student may know a grammar rule in isolation and still fail to use it while simultaneously planning, retrieving vocabulary and writing under time pressure.
Test the same capability at several loads:
Isolated → Short Task → Integrated Task → Timed Paper.
The point at which accuracy falls tells you more than the error label alone.
4. Separate One-Off Errors From Recurring Signatures
One spelling mistake is noise. The same tense drift appearing in four compositions is a signature. One slow paper may be fatigue. Repeated unfinished papers with the same early stall point indicate a timing or decision bottleneck.
Track recurrence before prescribing a large intervention.
5. Do Not Diagnose the Learner From an AL Band
Two students with the same overall Achievement Level can arrive there through different combinations of strengths and weaknesses. One may need foundation repair; another may need transfer or examination stabilisation.
Use observed output, not a grade stereotype.
6. Common Cross-Paper Failure Families
| Failure family | Typical signs | Repair direction |
|---|---|---|
| Representation | Misreads prompt, question or passage relationship | Decode and reconstruct |
| Retrieval | Knows during review, cannot produce in task | Active recall and contextual retrieval |
| Structure | Ideas exist but relationships are unclear | Sentence / paragraph / causal architecture |
| Transfer | Works on practised surface, fails on unfamiliar one | Changed-surface retesting |
| Load | Accuracy collapses when tasks combine | Reduce and rebuild integration |
| Time | Completion fails predictably | Find drift point and reallocate |
| Self-check | Same correctable errors survive to submission | Personal error signature and risk-based checking |
7. Repair One Layer at a Time
If a student is simultaneously told to improve grammar, vocabulary, planning, handwriting, speed, creativity and confidence, the diagnosis has not narrowed enough.
Choose the smallest intervention that can test the diagnosis. If it works, retest. If it does not, move to the next plausible cause.
8. The Failure Signature Check
SYMPTOM — What failed visibly?
EARLIEST CAUSE — What upstream weakness could generate it?
TEST — How can I isolate that capability?
REPAIR — What is the smallest targeted intervention?
RETEST — Does it survive a changed task and higher load?
The Exit Condition
A failure signature is useful when it changes the next action. The learner, parent or tutor should be able to move from “something went wrong” to one testable cause and one bounded repair rather than adding a generic pile of more work.