When a Reasoning Question Has Two Plausible Patterns

Start by separating the observations from the story

A pattern question gives you observations and asks you to infer an operation. The observations might be a sequence of values, changes in a grid, rotations of a shape, or relationships between quantities. Your first explanation is a hypothesis, not a fact. That distinction matters because a short set of observations can often be fitted by more than one rule.

Read the prompt and record only what is actually present. Note the order of the elements, their position, direction, size, colour, count, or arithmetic value. Then describe the change in plain language before naming an operation. For example, “the increase gets larger each time” is more useful at first than immediately writing a formula. It helps you see which parts of the display the rule must explain.

Also check the task boundary. Does it ask for the next term, the missing cell, the odd one out, or the rule itself? A rule that explains a sequence may not answer a question about position or symmetry. The testing Standards describe a construct as the characteristic an assessment is intended to measure, and they emphasize that score interpretations depend on the specified use. The same discipline applies at item level: solve the task that was asked, not a more entertaining task you invented.

Generate two rules, then make them earn their place

When you see a second plausible pattern, write both candidates down. Do not keep them as vague impressions such as “it seems to alternate.” State the order and operation: add successive even numbers; rotate a quarter-turn and increase the count; combine the first two cells to produce the third. An explicit rule can be checked. A hunch cannot.

Use four tests for each candidate. First, coverage: does it explain every supplied element, or only the part that caught your eye? Second, prediction: does it produce one answer for the missing element? Third, economy: does it need fewer special cases, hidden resets, or arbitrary choices? Fourth, task fit: does it use the kind of relation the instructions or layout make relevant? These are practical comparison tools, not a claim that nature always rewards the shortest formula.

A rule that fits the visible data but requires an exception is weaker than one that fits the data directly. A rule that depends on an unannounced fact, such as a particular calendar convention or an imagined three-dimensional view, is also weaker unless the prompt supplies that context. Simplicity is a useful tie-breaker because every extra condition creates another way for the rule to fail, but simplicity cannot repair a rule that ignores part of the item.

This is close to the distinction between an item being plausible and being proper. In multiple-choice assessment research, ambiguity becomes a construction problem when more than one response can be justified under the item’s information. For a solver, the practical signal is simple: if two explanations survive the same checks and lead to different answers, stop searching for a magical interpretation and consider that the question has not determined a unique solution.

Check the whole structure, not one convenient slice

Many reasoning items contain several possible directions of analysis. In a sequence, inspect changes between terms and changes between those changes. In a grid, test rows, columns, and, where justified by the layout, diagonals. In a spatial problem, track each property separately: orientation, number of parts, location, and shading. Do not combine all of them into one elaborate explanation before you know which property changes.

Imagine a grid in which each row appears to rotate a symbol one step, while each column appears to add a mark. Test whether the proposed missing cell satisfies both relationships. If it satisfies only the top row, it is not yet a grid rule. If one interpretation explains rotation and another explains mark count, ask whether the item intends a two-feature rule or whether one feature is decorative. The answer should be supported by the full structure, not by the first regularity you notice.

Answer choices can help you test a rule, but they should not become the source of the rule. Start with the supplied information, predict the result, and then compare your prediction with the options. Looking at options too early can encourage elimination by wording, position, or visual resemblance. A study of multiple-choice reasoning tasks in physics found that students sometimes relied on reading comprehension, knowledge of the answer choices, and test-wise strategies rather than the intended reasoning process. That finding does not show that every multiple-choice item is weak, but it does show why an attractive option is not proof of a sound rule.

If your candidate predicts an answer that is absent, check your arithmetic and representation first. If it predicts several possible answers, the rule is incomplete. If a different candidate predicts one listed answer only by adding a special condition, label that condition instead of quietly assuming it. Good solving makes assumptions visible.

A small example: fitting is not the same as deciding

Consider an example that is not taken from a test. The sequence is 2, 4, 8, 14, and you must propose the next value. One rule looks at the differences: add 2, then 4, then 6, so the next increase is 8 and the next value is 22. Another rule alternates operations: add 2, multiply by 2, add 6, multiply by 2, giving 28 next. Both descriptions can reproduce the four values shown.

The difference rule is usually the better answer because it uses one consistent operation on the gaps and extends the visible progression without switching operation types. The alternating rule is not impossible, but it has a more complicated structure and its next operation is inferred from only one completed alternation. If the question gave another term, the alternatives could separate more clearly. If an answer choice or instruction explicitly indicated alternating operations, that additional information could change the decision.

The important lesson is not “always choose arithmetic differences.” It is to compare how much evidence each rule uses and how much it assumes. Ask: Does the rule apply uniformly? Does it explain all the displayed values? Does it produce one next value? Does it need an arbitrary starting point, direction, or exception? If the only reason to prefer one candidate is that it appears in the answer choices, your confidence should remain modest.

For a visual item, use the same method without pretending that a picture is self-explanatory. State the feature you are tracking, test the relation in every available direction, and see whether the missing part is forced. Never reconstruct or share proprietary items from a named professional instrument. A fresh example is enough to learn the operation; protected material is not needed.

Know when the problem is the item, not your reasoning

Some uncertainty belongs to the solver: a missed sign, an overlooked rotation, or a rule that was never checked against the final row. Some uncertainty belongs to the item: vague wording, a decorative feature that appears meaningful, or answer choices that allow two interpretations. These causes should not be treated as the same kind of evidence.

Item quality is part of measurement quality. ETS describes its item-development process as including checks for alignment with the intended objective, technical quality, accessibility, fairness, grammar, and accuracy. The testing Standards likewise treat item wording and format as part of the test content that must be related to the construct being measured. In plain terms, a question should not make you use reading tricks or an accidental visual cue when the stated aim is pattern reasoning.

A multiple-choice result also hides the path you took. Selecting the correct option can reflect the intended operation, elimination, familiarity with a format, or a lucky guess. Selecting the wrong option can reflect a reasoning error, a misread instruction, time pressure, or an ambiguous item. Research comparing multiple-choice and constructed-response versions of reasoning tasks has found that the formats are not automatically interchangeable and that the skills used to answer may differ. One item therefore gives very little basis for a broad claim about ability.

If you are reviewing an item after the test, write a short audit: my rule, the evidence it explains, the assumption I used, and the exact point where the official explanation differs. If the official answer requires information not present in the prompt, ask the publisher for clarification. A correction or explanation is more useful than silently turning your doubt into a low score.

Make the next decision proportionate to the result

During a timed quiz, use a short stopping rule. Spend a first pass identifying the task and the obvious features. If two rules remain, compare them across the whole item and choose the one with broader coverage and fewer exceptions. Mark the question if the format permits, move on, and return only if time remains. Endless searching can turn a small ambiguity into a large timing cost.

Afterward, classify the experience rather than labelling yourself. Was the difficulty caused by calculation, tracking several visual properties, reading the instruction, time pressure, or genuine underdetermination? The classification tells you what to practise. Reworking a familiar item can improve familiarity with its format or operation. That is a useful near-transfer outcome, but it does not justify claiming that a short practice session raises general intelligence or predicts future performance in unrelated settings.

Test IQ Free’s free assessment is an adaptive educational reasoning assessment with 50 original questions across pattern, quantitative, and spatial reasoning. It reports an estimated IQ range and domain performance; it is not a diagnosis or a professional assessment. If you use it, take the free 50-question assessment and review your reasoning profile. The optional report can organize that review and a practice plan, but it is practice support, not a promise of a higher IQ.

Return to the original question with a changed interpretation: two plausible patterns are a reason to make your assumptions explicit, not a reason to decide what kind of thinker you are. The practical next step is to audit the rule against every observation, record unresolved ambiguity, and practise the specific operation that the evidence actually points to.