OSAT Middle Level/Intermediate Mathematics (225) common mistakes
OSAT Middle Level/Intermediate Mathematics (225) · Understand the traps and how to avoid them.
When simplifying or combining radical and rational expressions, the candidate distributes an exponent or root across a sum, writing √(a² + b²) = a + b or (a + b)² = a² + b², and then builds the rest of the solution on that false identity.
The linear-looking structure of the expression invites the same reflex used with multiplication, and the candidate has practiced distributing over sums so often that the habit overrides the rule that roots and powers do not distribute over addition. On a multiple-choice test the resulting answer is often present as a distractor, which confirms the error instead of exposing it.
What to do instead
Before any simplification, test the proposed identity with concrete numbers (a = 3, b = 4). Require every step to cite an operation that is actually valid: factoring, multiplying both numerator and denominator, or applying an exponent to a product or quotient, never to a sum. Drill a short set of 'valid or invalid?' items where the candidate must justify each transformation in one sentence.
The candidate converts between fractions, decimals, and percents by moving the decimal point the wrong number of places or in the wrong direction, for example writing 0.4% as 0.4 rather than 0.004, and then reports a final answer that is off by a factor of 10 or 100.
Percent work is usually done by memorized procedures rather than by reasoning about magnitude, so under time pressure the candidate recalls 'move it twice' without the anchor of what a percent means. Answers that are wrong only by a power of ten look plausible, so nothing triggers a check.
What to do instead
Adopt a single written conversion ritual: percent to decimal means divide by 100, and check the result against a benchmark such as 50% = 0.5 and 1% = 0.01. Estimate the answer before computing, then compare the computed result to the estimate; a mismatch signals a place-value slip. Practice with mixed items such as 0.4%, 4%, 40%, and 400% in one set so the contrast becomes automatic.
When solving a linear equation or inequality, the candidate multiplies or divides by a negative coefficient and keeps the inequality symbol unchanged, or clears a denominator without tracking the restriction that the denominator cannot be zero, producing an extraneous solution.
Equation-solving is learned first and practiced far more, so the inequality-specific rule and the domain restriction are the last things recalled and the first things dropped. The candidate is focused on isolating the variable and treats the symbol as decoration.
What to do instead
Make sign-flipping an explicit written step: circle the coefficient before dividing, and if it is negative, draw the flip arrow on the symbol. For rational equations, list excluded values before solving and check each candidate against that list at the end. Include a verification step where a test value from the claimed solution set is substituted into the original inequality.
The candidate treats a function transformation as if it behaved like ordinary arithmetic, shifting y = f(x) right when seeing f(x + 2) or stretching horizontally when seeing 2f(x), and misidentifies the vertex or asymptote of the transformed graph.
Inside-the-parentheses changes operate in the opposite direction of intuition, and the candidate has no reliable routine for separating inside from outside changes. Without a graph, the error is invisible until the answer is graded.
What to do instead
Use a fixed anchor function, typically y = x² or y = |x|, and compute a small table of three points for both the parent and the transformed rule before answering. Classify each change as inside (horizontal, opposite direction) or outside (vertical, same direction) and write that label next to the expression. Verify the vertex or key point numerically rather than by inspection.
When working with slope, parallel and perpendicular lines, or distance and midpoint, the candidate plugs coordinates into the wrong slots, computes slope as run over rise, or gives the perpendicular slope as the negative of the original instead of the negative reciprocal.
The formulas look similar and are memorized as strings rather than as relationships, so under pressure the candidate reconstructs them approximately. The negative-reciprocal rule is easy to half-remember because the negative sign is the more memorable part.
What to do instead
Label the coordinates (x₁, y₁) and (x₂, y₂) on the page before substituting, and write the slope formula with the labels attached. For perpendicularity, verify by multiplying the two slopes and confirming the product is −1. Practice a mixed set where each item requires stating which relationship (parallel, perpendicular, neither) is being used and why.
The candidate applies a surface-area formula when the question asks for volume, or forgets that a real-world measurement problem requires unit conversion before the formula is applied, mixing inches and feet or centimeters and meters within one calculation.
Geometry items often embed the conversion in the stem rather than stating it, and the candidate reads for the shape and the numbers while skimming past the units. Surface area and volume formulas for prisms, cylinders, and spheres are similar enough to be confused when recall is rushed.
What to do instead
Rewrite the given quantities with units attached and convert everything to a single unit in a dedicated first step, before selecting any formula. Decide explicitly whether the question asks about covering (area), filling (volume), or enclosing (perimeter or circumference), and write that word at the top of the work. Keep a one-page formula reference that pairs each formula with the quantity it measures.
The candidate computes a mean, median, or standard deviation correctly for the data given but answers a question about a different quantity, for example reporting the mean when the item asks which measure is most resistant to an outlier, or reading the median off a histogram by taking the middle bar rather than the middle value.
Statistical items test interpretation as much as calculation, and the candidate defaults to the computation that was most recently practiced. Histograms and box plots require reading cumulative position, which is a different skill from computing a summary statistic.
What to do instead
Restate the question in the candidate's own words before computing, naming the target quantity explicitly. For graphical displays, find the total frequency first and locate the median position by cumulative count, not by visual midpoint. Practice items that ask 'which measure and why' so the reasoning, not just the number, is rehearsed.
The candidate confuses correlation with causation or misreads a scatterplot's line of best fit as a prediction guarantee, concluding that because two variables are strongly associated one must cause the other, or extrapolating far beyond the observed data range.
Everyday reasoning treats pattern as proof of cause, and the candidate has seen the caution so often that it reads as boilerplate rather than as a testable claim. Extrapolation errors arise because the arithmetic of substituting a value into the regression line is easy and feels safe.
What to do instead
For each association item, write down a plausible confounding variable and check whether the item's conclusion would survive it. Mark the observed data range on the axis and refuse predictions outside it unless the item explicitly asks about extrapolation. Distinguish the three claims (association, prediction, causation) and match the answer to the claim actually made.
The candidate adds probabilities of overlapping events without subtracting the intersection, or multiplies probabilities for events that are not independent, so P(A or B) is overstated and P(A and B) is understated.
The addition and multiplication rules are taught together and the conditional structure is easy to skip when the stem does not use the word 'given'. The candidate also assumes independence by default because most textbook examples are independent.
What to do instead
Before computing, draw a two-way table or Venn diagram and fill in counts, then read the probability off the diagram. Ask explicitly whether the events can occur together (overlap) and whether knowing one changes the other (dependence), and write the answers down. Use the conditional formula P(A|B) = P(A and B) / P(B) whenever the stem supplies a condition.
The candidate answers a counting or arrangement problem with the wrong model, using permutations where order does not matter or combinations where it does, or forgets to divide by repeated-factor factorials when identical items are arranged, and reports a count that is too large by a whole factor.
The words 'arrange' and 'select' are the only cue most candidates use, and many stems phrase selection in ways that sound sequential. The identical-items adjustment is the least practiced part of the topic and is dropped under time pressure.
What to do instead
Ask one diagnostic question before computing: would swapping two chosen items produce a different outcome? If yes, order matters and use permutations; if no, use combinations. For arrangements with repeats, count total positions first and divide by the factorial of each repeated group. Sanity-check by listing all outcomes for a small version of the same problem.