How to Check an Answer

Precise definition

To check an answer is to seek independent evidence that it satisfies the original problem, not merely repeat the same steps. Verification can establish exact correctness for a candidate; reasonableness checks detect many errors but are not always proofs.

Notation and mathematical language

Useful checks include substitution into the original equation, differentiation of an antiderivative, multiplication of a proposed inverse, recalculation by a different method, estimation, bounds, dimensions or units, domain restrictions, and boundary cases. A residual measures discrepancy but must be interpreted relative to scale and precision.

Conceptual picture

A strong check reverses the solution process or uses another representation. Repeating calculator entries can reproduce the same transcription error. Exact and approximate answers need different language: equality for proven exact values, approximation with precision and error awareness for numerical results.

Conditions and key results

Substitution may be necessary but not sufficient when the problem asks for all solutions, a maximum, or a proof. A candidate can work while another is missing. Numerical agreement at sample points cannot prove a universal identity.

A reliable strategy

1. Return to the original prompt and list every requirement, restriction, unit, and requested form.
2. Use the strongest direct check available, such as substitution or differentiation.
3. Apply an independent magnitude, sign, graph, unit, or alternative-method check.
4. Confirm completeness, domain, precision, and interpretation before declaring the work finished.

Fully worked example

Interpretation and application

Verification improves mathematics, code, measurement, and modelling. An algebraically correct result can still use the wrong model or units, so contextual validation remains separate from symbolic correctness.

Common mistakes

Verification and reasonableness

• Substitute into the unsimplified original.
• Use an inverse relation or second method.
• Check scale, sign, units, restrictions, edge cases, and requested precision.

Practice

1. How do you check a linear-equation solution?
2. How do you check an antiderivative?
3. Does one working example prove an identity?

Further deduction

For an optimization answer, check more than the derivative equation: feasibility, endpoints, nondifferentiable points, classification, and units. For a probability, verify it lies in $[0,1]$ and that disjoint totals sum correctly. For a matrix inverse, multiply in both orders. Match the check to the mathematical object.

Independent checks should use a different representation from the original method. After solving a quadratic by factoring, substitute roots or compare their sum and product with the coefficients. After differentiating symbolically, estimate a centred secant slope $[f(x+h)-f(x-h)]/(2h)$ at a safe point. After computing a probability, enumerate a small sample space or simulate only as a plausibility check; simulation does not prove an exact fraction. For a matrix solution, multiply $A\hat x$ and compare with $b$, while remembering that a small residual may be misleading for an ill-conditioned numerical system. A check also tests the question's constraints: a negative length, probability above 1, excluded denominator value, wrong units, or missing endpoint can invalidate otherwise clean arithmetic. Selecting a check tailored to the mathematical object is stronger than simply repeating the same calculator entry.

Related topics

• How to Show Your Work
• How to Learn from Mistakes
• Common Algebra Errors