SSU ITB Jalur Tes common mistakes
SSU ITB Jalur Tes · Understand the traps and how to avoid them.
In Penalaran Umum syllogism and logical-deduction items, the candidate accepts a conclusion because its content sounds true in the real world instead of checking whether it follows necessarily from the stated premises, so they pick the plausible real-world option over the strictly valid one.
Everyday knowledge is automatic and fast, while formal validity checking is slow; under time pressure the brain defaults to judging truth of content rather than the form of the argument, and items are deliberately built with true-but-invalid and false-but-valid distractors.
What to do instead
Before reading the options, rewrite the premises in stripped symbols (A, B, C) so content cannot bias you, then test each option only against those symbols. Train with mixed sets where half the correct answers are factually absurd; mark any option you accepted on real-world grounds and re-derive it formally.
On Penalaran Umum quantitative-comparison and number-sequence items, the candidate computes both quantities fully and then forgets to compare them, or treats 'equal' as never being the answer, so they lose the Pilihan C (sama besar) cases.
School habit rewards getting a single number, not judging a relation; candidates also carry a folk belief that comparison questions must have a 'bigger' winner, so equality is psychologically filtered out.
What to do instead
Adopt a fixed two-column scratch layout (P | Q) and end every item by writing one of the three relation symbols before touching the options. Drill blocks of comparison items engineered so roughly a third are equal, and count how many times you chose 'equal' to confirm you are not suppressing it.
In Pengetahuan dan Pemahaman Umum, the candidate answers synonym, antonym, and relation items from a single remembered meaning of a word, ignoring that Indonesian words are often polysemous and that register (formal, colloquial, technical) changes the correct pairing.
Vocabulary is stored as one strong association per word, and the first meaning retrieved feels definitive; exam items exploit secondary senses and near-synonyms whose nuance differs, so the strongest association is often the trap.
What to do instead
For each stem word, list two or three senses and a register note before looking at options, then eliminate options that match only a wrong sense. Build a personal log of word pairs you missed with the specific nuance that separated them, and review it weekly rather than rereading generic word lists.
In Kemampuan Memahami Bacaan dan Menulis, the candidate chooses the option that 'sounds better' for a sentence-completion or effective-sentence item without checking the grammar of the surrounding clause, so they accept a fragment, a dangling subject, or a mismatched conjunction.
Fluency illusion: text that reads smoothly is judged correct even when structurally broken, and candidates rarely parse the sentence's subject and predicate under time pressure, so agreement and conjunction logic go unchecked.
What to do instead
Physically bracket the subject and verb of the target sentence and confirm they agree and that the conjunction matches the logical relation (contrast, cause, addition). Practise on items where the fluent-sounding option is grammatically wrong, and state aloud why each rejected option fails structurally.
In reading-comprehension sets across Bahasa Indonesia and English literacy, the candidate answers from prior knowledge or personal opinion about the topic rather than from what the passage states, and picks an option that is true in general but unsupported by the text.
Comprehension questions feel like general-knowledge questions, and background knowledge activates automatically; options that agree with the reader's worldview feel safer than text-supported options that are bland or counterintuitive.
What to do instead
For every option, point to the line or sentence that supports it; if you cannot locate one, eliminate it regardless of how true it seems. Do sets on topics you know well and deliberately choose the text-supported answer over your own view, then check against the key to break the habit.
On inference and main-idea items, the candidate selects an option that is too broad or too narrow, or that restates one detail, instead of the option that matches the passage's exact scope and the author's claim.
Distractors are written to be partially correct, and a striking detail or sweeping statement is more memorable than the accurate middle-scope answer; candidates also confuse 'what the passage is mostly about' with 'what is interesting in it'.
What to do instead
Write a one-line main idea in your own words before reading the options and reject any option that adds scope ('all', 'always') or drops the central claim. For inference items, verify the option is a necessary consequence, not merely a possible one, and label each rejected option as too broad, too narrow, or off-topic.
In Literasi dalam Bahasa Inggris, the candidate skims the passage for words that echo the question and picks the option containing those words, falling for lexical-match distractors and missing negation, hedging, or contrast markers.
Keyword matching is a fast, low-effort search strategy that feels productive, and under time pressure the visual familiarity of repeated words is mistaken for evidence; negation and concessive markers are small words easily skipped.
What to do instead
Read the question stem first, then read the relevant paragraph fully and paraphrase it before looking at options; reject any option that merely recycles passage vocabulary. Circle negation and contrast words (not, except, although, however, only) during reading and drill items whose correct answer contains none of the stem's keywords.
In Penalaran Matematika context problems, the candidate extracts numbers and immediately computes with them without defining what each quantity represents or what the question actually asks, producing a correct calculation of the wrong quantity.
Word problems trigger a rush to arithmetic, and the final sentence containing the actual question is often read last or skimmed; multi-step contexts also hide unit changes and intermediate conditions that a direct computation ignores.
What to do instead
Before calculating, write one line: given quantities with units, the target quantity with its unit, and the operation chain linking them. After computing, restate the answer as a sentence answering the question and check the unit and order of magnitude; if either looks odd, re-read the final sentence.
Across all domains, the candidate spends disproportionate time on one hard item, then rushes the remaining questions, leaving easy items unanswered or misread; they also fail to use the elimination-and-guess policy on items they cannot finish.
Sunk-cost attachment to a nearly-solved problem plus anxiety about skipping makes candidates stay put, and many do not know whether their exam penalises wrong answers, so they leave blanks out of caution or guess randomly without eliminating.
What to do instead
Set a per-item time cap in practice (roughly one minute for TPS items), mark and move on when exceeded, and return only after the section is complete. Confirm the scoring rule for your exam path, then practise a fixed guessing protocol: eliminate at least two options and commit, so timed mocks reproduce real pacing.
In Pengetahuan Kuantitatif, the candidate substitutes specific numbers or estimates from a sketch and concludes a general statement is true, instead of testing whether it holds for all permitted values, including zero, negatives, and fractions.
One confirming example creates strong confidence, and candidates rarely search for counterexamples; properties that hold for positive integers silently fail for zero or fractions, and diagrams are treated as exact when they are only illustrative.
What to do instead
For 'must be true' items, deliberately test the boundary set: 0, 1, a negative, and a fraction between 0 and 1. Treat any figure as not-to-scale unless measurements are given, and for 'which could be false' items hunt for a single counterexample rather than verifying the statement.