GNFS Aptitude Test common mistakes
GNFS Aptitude Test · Understand the traps and how to avoid them.
Spending a fixed amount of time on every numerical question instead of banking the easy percentage and ratio items first and returning to multi-step rate problems later.
Candidates treat the section as a uniform test, but GNFS-style numerical papers mix quick arithmetic with longer multi-stage calculations; the clock pressure makes them commit to a hard item and lose three or four easy marks.
What to do instead
Before the timer starts, decide a per-question ceiling (for example 60–75 seconds). Do one full pass answering only items you can finish inside the ceiling, flag the rest, then use the remaining time on flagged items in ascending difficulty. Practise with a visible timer so the triage becomes automatic.
Estimating instead of computing when a numerical answer option set is tightly clustered, then selecting the nearest-looking value.
Under time pressure candidates round intermediate figures (for example 17.4% to 17%) and the distractors are built precisely around that rounding, so the estimate lands on a wrong option.
What to do instead
Check the spread of the options first. If any two options differ by less than about 5%, do exact arithmetic and keep one extra decimal place through the intermediate steps, rounding only at the final line. If options are widely spaced, estimation is legitimate and faster.
Answering verbal reasoning items from general knowledge or personal opinion rather than from the passage as written.
The passages are often on familiar topics (business, science, public policy), so candidates import outside facts and mark a statement 'true' because it is true in the world, not because the passage supports it.
What to do instead
For every true/false/cannot-say item, locate and underline the exact sentence that would have to exist for the statement to be true. If no sentence supports or contradicts it, the answer is 'cannot say'. Do not accept inferences that require a single extra step beyond the text.
Confusing 'false' with 'cannot say' on verbal items, treating anything not explicitly stated as false.
Everyday reasoning treats absence of evidence as evidence of absence; the cannot-say category has no counterpart in normal conversation, so the distinction feels artificial and gets collapsed.
What to do instead
Apply a two-question test. First: does the passage contradict the statement? If yes, false. Second: if not, could the statement be true given the passage? If yes, cannot say. Only mark false when you can point to a contradicting sentence.
Solving abstract reasoning series by looking for a single rule (for example 'the dot moves one place clockwise') and stopping once one pattern fits.
Abstract items usually layer two or three rules at once (rotation plus alternation of shading plus element count). Finding one rule produces a confident but wrong answer because the distractor options satisfy that rule alone.
What to do instead
After identifying a candidate rule, verify it against every element in every panel, then explicitly ask what a second and third element are doing. Track each element separately (position, colour, count, size) on scratch paper before comparing options.
Assuming the answer options are mutually exclusive and that finding one plausible option ends the search.
Candidates read options top to bottom and stop at the first that 'works', but several options may fit the visible rule while only one fits all rules; the first plausible option is often a deliberate trap.
What to do instead
Read all options before deciding, and eliminate actively: for each option, state the rule it violates. Choose only the option that survives every rule you have identified, not the first one that survives one rule.
Mismatching rows when copying figures from a table or comparing two lists side by side, especially where similar codes or names appear in adjacent rows.
Attention-to-detail items deliberately place near-identical entries (for example AB-1234 and AB-1324, or two similar surnames) next to each other; the eye skims and the brain fills in the expected pattern.
What to do instead
Use a physical or on-screen pointer (cursor, ruler edge, folded paper) to track one row at a time, and compare character by character rather than word by word. For long lists, work in blocks of five rows and re-check the block boundary before moving on.
Rushing the final verification step on accuracy items, submitting a transcription or arithmetic result without re-reading it against the source.
The task feels trivial once understood, so candidates skip the check to save time; the errors that cost marks are almost always introduced in the last step, not in the reasoning.
What to do instead
Build a mandatory five-second check into every accuracy item: re-read the source value and your answer once, comparing digits in reverse order (right to left), which breaks the pattern-completion effect that hides transposition errors.
Choosing the first workable solution to a decision-making scenario without testing it against the stated constraints and priorities.
Scenario items present an obviously sensible action, but the question stem contains constraints (budget cap, deadline, delegated authority, safety policy) that rule it out; candidates answer from instinct about 'good management' rather than from the stated rules.
What to do instead
Before reading the options, extract the constraints into a short list (limits, priorities, who decides). Then score each option against that list and eliminate any that breaches a hard constraint, even if it is otherwise attractive. Answer what the scenario permits, not what you would personally do.
Reading a chart's axis labels, units, or baseline carelessly and drawing a conclusion from the visual slope or height rather than the plotted values.
Truncated axes, mixed units (thousands versus millions), and dual-axis charts exaggerate or invert apparent trends; the visual impression is processed faster than the labels and dominates the judgement.
What to do instead
Before answering any data-interpretation item, state the units of each axis, the starting value of the baseline, and which series belongs to which axis. Then read the actual data values for the points in question and compute the change rather than judging it by eye.