EDPT common mistakes
EDPT · Understand the traps and how to avoid them.
Working arithmetic problems entirely in your head and writing only the final answer, so a single slip in carrying or regrouping silently corrupts the result with no visible trail to catch it.
The EDPT arithmetic section is timed aggressively, and candidates feel that skipping scratch work is the only way to finish. Mental arithmetic feels faster because errors are invisible until the answer is already committed.
What to do instead
Force a written intermediate step for every operation: write the carry, write the regrouped value, write the product before adding. Practice timed sets where an answer with no scratch work is scored as wrong regardless of correctness, so the habit is built under time pressure rather than only in untimed practice.
Reading a word problem once and immediately computing with the first two numbers that appear, without identifying what quantity is actually being asked for.
Under time pressure the brain pattern-matches to a familiar operation ('total' means add, 'left' means subtract) and starts calculating before the question sentence is parsed. Distractor numbers placed in the setup reward this shortcut with plausible wrong answers.
What to do instead
Adopt a two-pass routine: first pass, underline the final question sentence and circle the unit requested; second pass, select only the numbers that feed that quantity and explicitly discard the rest. Do drills where each problem contains one deliberately irrelevant number, and require yourself to state in writing why each used number is needed.
Losing track of place value or decimal alignment when adding, subtracting, or comparing numbers of different lengths, producing answers off by a factor of ten.
Handwritten columns drift when the numbers have unequal digit counts, and the eye aligns the leftmost digits rather than the decimal point. This is a mechanical habit, not a conceptual gap, so it survives even when the candidate understands the math.
What to do instead
Before computing, rewrite the problem with the decimal points vertically aligned and pad with trailing zeros so every number has the same number of decimal places. For comparisons, convert to the same number of decimal places first. Build a checklist item: 'align before you operate' and apply it even to single-line problems.
Spending a long block of time on one stubborn item while later, easier items go unanswered.
Sunk-cost pressure: having invested two minutes, abandoning the item feels like wasting that effort, and the fixed time limit is not being actively monitored. Candidates also underestimate how many remaining items are quick wins.
What to do instead
Set a hard per-item ceiling (for example, 60 to 90 seconds) and a visible timer. On hitting the ceiling, mark the item, select a provisional answer, and move on. Practice with a rule that any item revisited later must be re-entered from scratch rather than resumed mid-thought, which removes the incentive to keep grinding.
Treating a number series as arithmetic by default and testing only constant differences, then guessing when that fails.
Constant-difference series are the most familiar pattern, so the first hypothesis consumes the available time and alternatives are never systematically tried. Candidates also fail to compute the differences themselves, trying to eyeball the pattern instead.
What to do instead
Use a fixed ladder of hypotheses, written down: (1) first differences, (2) second differences, (3) ratios, (4) alternating or interleaved subsequences, (5) operations on position index. Compute each stage on paper rather than guessing, and abandon a stage only after it demonstrably fails.
Assuming a letter or symbol series follows alphabetical order when it actually alternates direction, skips a fixed number of letters, or interleaves two independent sequences.
Reciting the alphabet forward is automatic, so the mind anchors on forward progression and does not consider reversal or interleaving. Candidates also count letters inclusively versus exclusively without deciding which convention the sequence uses.
What to do instead
Immediately write the alphabet with index numbers above each letter on your scratch sheet. For each series, test forward step, backward step, and split-into-two-subsequences as three explicit hypotheses, and verify the rule against every given term before predicting the next one.
Solving analogies on surface association ('both are animals', 'both are tools') rather than on a specific, stated relationship, then choosing an option that matches the topic but not the relation.
Associative memory fires faster than relational analysis, and the distractor options are usually built to be topically related to the stem. The candidate never articulates the relation, so there is nothing to check the answer against.
What to do instead
Write the relation as a sentence with a blank verb before looking at the options ('A is a part of B', 'A is the tool used by B', 'A is the immature form of B'). Then test each option by substituting it into that same sentence, and reject any option where the sentence requires a different verb.
Reading a logic or syllogism problem and answering from intuition about what is probably true in the real world, instead of from what the stated premises strictly entail.
Real-world knowledge is automatically activated by familiar terms, and 'some' and 'all' are interpreted conversationally rather than formally. Candidates also reverse the direction of conditional statements.
What to do instead
Draw the premises as circles or arrows before evaluating any conclusion, and for conditionals write the contrapositive explicitly. Apply a single test to every candidate conclusion: is there a drawing consistent with all premises in which this conclusion is false? If yes, reject it, regardless of how plausible it sounds.
Treating a symbol-substitution or coding item as a memory task, trying to hold the mapping in your head while decoding, and losing the mapping partway through a multi-character string.
The mapping table is short enough that it feels memorizable, so no written key is made. Working memory is then consumed by the mapping rather than by the decoding, and errors appear in the middle of long strings.
What to do instead
Transcribe the mapping into a compact written key before touching the target string, then decode character by character with the key in view. For multi-step transformations, decode one layer at a time and write the intermediate string rather than composing transformations mentally.
Failing to verify a decoded answer by re-encoding it back through the rule, and submitting the first plausible-looking option.
Options in coding and decoding items are designed to differ by a single character or a single step of the transformation, so a partially correct decode matches a distractor. Time pressure discourages the extra pass.
What to do instead
Make reverse verification mandatory: take your chosen answer and apply the inverse of each transformation step, confirming you recover the original string exactly. Budget the final seconds of each item for this check, and practice sets where every distractor differs from the key by exactly one character so the check becomes habitual.