CA Foundation Paper 3 common mistakes
CA Foundation Paper 3 · Understand the traps and how to avoid them.
In ratio-and-proportion or partnership problems, setting up the equation with the ratio terms in the wrong order (for example writing A:B as 3:5 when the question says A is to B as 5 is to 3), then solving flawlessly and losing the whole mark.
Candidates read the numbers and start computing before fixing which quantity is the antecedent. The English phrasing ('A is 3/5 of B', 'A exceeds B by 2:3') is translated hastily, and once a wrong equation is written the arithmetic feels correct so no self-check is triggered.
What to do instead
Before any algebra, write a one-line 'given' statement in symbols: A/B = 3/5, or A = (3/5)B, and underline the subject of the sentence. After solving, substitute the answer back into the original English sentence and check the direction of the inequality or ratio. Train with 10 mixed phrasing variants of the same ratio problem so the translation step becomes deliberate.
In simple and compound interest, annuity and effective-rate questions, using the wrong time unit - entering 18 months as t = 18 with an annual rate, or applying half-yearly compounding without doubling the periods.
Formulas are memorised as S.I. = PRT/100 and A = P(1+i)^n without internalising that i and n must match the compounding period. Under time pressure, the stated rate is plugged in directly and the period conversion is skipped.
What to do instead
Adopt a fixed ritual: circle the rate, circle the period, then write 'i = rate per period, n = number of periods' and convert explicitly (18 months at 12% p.a. compounded quarterly gives i = 3%, n = 6). Practise a set where the same data is asked under yearly, half-yearly and quarterly compounding and compare answers.
In permutations and combinations, choosing the wrong tool - using nPr when order is irrelevant (selecting a committee) or nCr when order matters (arranging letters, seating with positions) - and also mishandling 'at least one' conditions by direct counting instead of the complement.
The words 'select', 'choose', 'arrange' and 'form a committee' are not reliably mapped to C and P under exam pressure, and identical letters or repeated items in word-arrangement problems are overlooked. The complement method is under-practised.
What to do instead
Write a keyword-to-tool table and apply it mechanically: 'arrange/order/rank/seat/word' means permutation, 'select/choose/committee/group' means combination. For 'at least one', default to total minus none. For words with repeats, always divide by the factorial of each repeated letter count. Do 15 rapid classification drills where you only decide P or C, without computing.
In sets, relations and functions, confusing the number of subsets, proper subsets and elements of a power set, or misreading 'A is a subset of B' as 'A is an element of B', leading to answers that are off by one or structurally wrong.
The notation is visually similar and candidates learn 2^n as a formula without tracking what n counts. Proper-subset and non-empty-subset variants are treated as the same question.
What to do instead
Build a small reference card: a set with n elements has 2^n subsets, 2^n − 1 proper subsets, 2^n − 2 non-empty proper subsets, and its power set has 2^n elements. For each practice question, first state n explicitly, then state which of the four counts is being asked, then compute.
In syllogisms, accepting conclusions that are only 'probably' true, or reversing the direction of a statement (treating 'All A are B' as 'All B are A'), and drawing a conclusion from two particular premises.
Candidates reason from real-world plausibility rather than from the formal form of the statements, and the standard rules for two-premise syllogisms (a valid conclusion needs at least one universal premise and the middle term distributed) are not internalised.
What to do instead
Draw Venn diagrams for every syllogism, including the possibility diagram with the 'some' region overlapping. Apply the fixed checks: a conclusion is valid only if it holds in every possible diagram; two particular premises yield no conclusion; 'All A are B' does not license 'All B are A'. Do 20 items where you only draw diagrams and mark valid or invalid without timing.
In seating arrangement and blood-relation puzzles, fixing the facing direction wrongly (treating a row facing north as facing south) or losing track of gender and generation while drawing the family tree, so the final answer is a mirror image or a wrong relationship.
The initial orientation is set once and never re-verified, and family-tree information is held in working memory instead of being drawn. Negative information ('C is not the wife of D') is dropped.
What to do instead
Draw the row and write N or S at both ends before placing anyone. For blood relations, always draw a tree with separate symbols for male, female and marriage, and mark each generation on a horizontal level. Record every negative clue as a crossed-out option in the diagram rather than in your head.
In coding-decoding and series questions, assuming the shift is uniform (all letters +2) when the pattern is alternating, positional, or based on the reverse alphabet, and then forcing later terms to fit the first pattern seen.
The first two examples usually fit a simple shift, so the pattern is fixed prematurely. Reverse-alphabet position (A = 26) is rarely practised, and number series are attacked by trial arithmetic instead of checking differences and ratios systematically.
What to do instead
For codes, test at least three hypotheses before committing: constant shift, alternating shift, and position-based shift; write the alphabet with both forward and reverse numbers at the top of the rough sheet. For number series, compute first differences, then second differences, then ratios, in that order, and only then consider squares, cubes and primes.
In measures of central tendency and dispersion, computing the mean of grouped data without using class midpoints, or mixing up the coefficient of variation with variance and reporting a percentage that is really a ratio.
The grouped-data formula is recalled as a plain average, and the distinction between absolute and relative measures of dispersion is blurred because only the formulas were memorised, not their purpose.
What to do instead
For grouped data, always write the midpoint column (x) and the frequency column (f) before touching a formula, then compute Σfx/Σf. Learn dispersion as a hierarchy: range, quartile deviation, mean deviation, standard deviation are absolute; coefficient of variation = (SD/mean) × 100 and coefficient of range = (L − S)/(L + S) are relative. State which type the question asks for before computing.
In correlation and regression, interpreting r as a percentage of causation, reporting r = 1.4 or similar out-of-range values without noticing, and using the regression of y on x to predict x.
The interpretation of r is taught loosely and the two regression lines are memorised as interchangeable equations. Arithmetic slips that produce impossible values pass unchecked because no plausibility filter is applied.
What to do instead
Memorise two guardrails: r always lies between −1 and +1, and correlation never proves causation. Label each regression equation at the moment you derive it (y on x estimates y from x) and check that the product of the two regression coefficients equals r². After every correlation answer, ask 'is this between −1 and 1?' before moving on.
In probability, adding probabilities of non-mutually-exclusive events without subtracting the intersection, and treating dependent draws (cards without replacement) as independent.
The addition and multiplication rules are recalled without their conditions, and the phrase 'without replacement' is read as a minor detail rather than a change to the sample space.
What to do instead
Before writing any rule, label the events: mutually exclusive or not, independent or not. Then apply P(A or B) = P(A) + P(B) − P(A and B) whenever overlap is possible, and for without-replacement draws reduce both numerator and denominator on each successive draw. Practise card and ball problems where you explicitly state the revised denominator at each step.