Interpreting answers in context

Interpreting an answer means translating mathematics back into the situation that produced it. A bare number such as $3.4$ is rarely a complete interpretation. A good conclusion identifies the quantity, gives its units, explains its direction or practical meaning, and respects the assumptions and limits of the model.

This skill appears throughout A level Mathematics. You may need to interpret a root, gradient, integral, parameter, probability, test result, force or negative velocity. The calculation can be correct while the conclusion is not.

What you should know first

You should be able to:

• substitute into formulae and solve equations;
• work with graphs, gradients and rates of change;
• round to a stated accuracy;
• use basic probability and statistical language;
• recognise that a mathematical model simplifies reality.

Review mathematical language and notation, exact arithmetic and mathematical modelling if these ideas are uncertain.

A reliable interpretation routine

After obtaining a mathematical answer, ask three questions.

1. What quantity is this? Name the variable in ordinary language.
2. What does its value mean? Include units, direction, comparison or probability as appropriate.
3. Is it valid here? Check the domain, required accuracy, physical restrictions and modelling assumptions.

For instance, suppose $t$ is time in seconds and solving a model gives

$$t=-2 \quad\text{or}\quad t=5.$$

Both values may be algebraically correct. If the model starts at $t=0$, then $t=-2$ is outside the modelled time interval. The contextual conclusion is

$$\boxed{\text{The event occurs }5\text{ seconds after the start.}}$$

Do not write that the negative root is “impossible” without saying why. It may describe a time before the chosen origin, but it does not answer this question.

State the quantity and its units

Units distinguish quantities that happen to have the same numerical value. If $x=12$ represents a distance measured in metres, write $12\text{ m}$, not merely $12$.

Derived units require particular care:

$$\begin{array}{c|c} \text{Quantity} & \text{Typical unit} \\ \hline \text{area} & \text{m}^2 \\ \text{volume} & \text{m}^3 \\ \text{velocity} & \text{m s}^{-1} \\ \text{acceleration} & \text{m s}^{-2} \\ \text{force} & \text{N} \\ \text{rate of flow} & \text{m}^3\text{ s}^{-1} \end{array}$$

Worked example 1: interpret an area

A rectangular enclosure has length $x$ metres and width $(20-x)$ metres. Its maximum area is found to be $100$.

The interpretation is not “the maximum is 100”. It is

$$\boxed{\text{The greatest possible area of the enclosure is }100\text{ m}^2.}$$

The words identify the optimised quantity, and the squared unit identifies area.

Self-check 1

Water volume $V$, measured in litres, satisfies $V=600-25t$, where $t$ is in minutes. Interpret the coefficient $-25$.

Answer

The volume of water decreases at a constant rate of $25$ litres per minute. Equivalently,

$$\frac{dV}{dt}=-25\text{ L min}^{-1}.$$

The minus sign describes a decrease. It is clearer to say “decreases by $25$” than “changes by $-25$”.

Round for the situation, not just the display

An exact mathematical answer may not be a usable contextual answer. Counted objects usually require integers, money usually requires the smallest relevant currency unit, and measured quantities should not claim unjustified precision.

Worked example 2: a minimum whole number

A minibus seats $16$ passengers. A school must transport $93$ students. Division gives

$$\frac{93}{16}=5.8125.$$

Rounding to the nearest whole number gives $6$, but the reason is stronger: five minibuses carry only $80$ students, so the minimum is

$$\boxed{6\text{ minibuses}.}$$

When a question asks for a minimum capacity, round up. When it asks for the number of complete items that can be made from limited material, rounding down may be appropriate.

Worked example 3: premature rounding changes a decision

A component is acceptable only if its calculated length is at most $8.50\text{ cm}$. A calculation gives

$$L=8.5047\ldots\text{ cm}.$$

To $3$ significant figures, this displays as $8.50\text{ cm}$. It would still be wrong to call the component acceptable because the unrounded value satisfies

$$8.5047\ldots>8.50.$$

Make comparisons using the unrounded value. Round only the reported result. See checking answers for further accuracy checks.

Interpret roots and solutions within the model

Equations often produce several mathematical solutions. Context supplies extra restrictions such as

$$t\geq 0,\qquad x>0,\qquad 0\leq p\leq1,$$

or an explicit time interval.

Worked example 4: two roots, one relevant time

The height of a ball above the ground is modelled by

$$h(t)=1.5+14t-4.9t^2,$$

where $h$ is in metres and $t$ is in seconds after release. Find when the ball reaches the ground.

Set $h=0$:

$$1.5+14t-4.9t^2=0.$$

The quadratic formula gives

$$t=\frac{14\pm\sqrt{14^2+4(4.9)(1.5)}}{9.8} =-0.1035\ldots\quad\text{or}\quad2.9607\ldots$$

The model begins at release, so $t\geq0$. Therefore

$$\boxed{\text{The ball reaches the ground about }2.96\text{ s after release.}}$$

The negative root is not used because it lies before the model’s stated starting time.

Worked example 5: every root has contextual meaning

A company’s profit, in thousands of pounds, is modelled by

$$P(x)=-(x-2)(x-10),$$

where $x$ is the number of hundreds of units sold. The break-even points satisfy $P(x)=0$, so

$$x=2\quad\text{or}\quad x=10.$$

Since $x$ counts hundreds of units, the interpretation is

$$\boxed{\text{The company breaks even when it sells }200\text{ or }1000\text{ units.}}$$

Do not report “2 or 10 units”. The scale attached to the variable is part of the model.

Interpret gradients and rates of change

If $y$ depends on $x$, then

$$\frac{dy}{dx}$$

is the instantaneous rate of change of $y$ with respect to $x$. Its units are

$$\frac{\text{units of }y}{\text{units of }x}.$$

Its sign gives direction:

$$\frac{dy}{dx}>0 \Rightarrow y\text{ is increasing}, \qquad \frac{dy}{dx}<0 \Rightarrow y\text{ is decreasing}.$$

Worked example 6: interpret a derivative

The temperature $T$ of a liquid, in degrees Celsius, is modelled as a function of time $t$ in minutes. At $t=4$,

$$\frac{dT}{dt}=-3.2.$$

A complete interpretation is

$$\boxed{\text{After }4\text{ minutes, the temperature is decreasing at }3.2\ ^\circ\text{C per minute.}}$$

This is an instantaneous rate. It does not say that the temperature falls by exactly $3.2\ ^\circ\text{C}$ during every minute.

Worked example 7: interpret a connected rate

A circular oil patch has radius $r$ metres and area $A=\pi r^2$. At one instant,

$$r=5,\qquad \frac{dr}{dt}=0.4\text{ m min}^{-1}.$$

Differentiate with respect to time:

$$\frac{dA}{dt}=2\pi r\frac{dr}{dt}.$$

Hence

$$\frac{dA}{dt}=2\pi(5)(0.4)=4\pi\text{ m}^2\text{ min}^{-1}.$$

Therefore, when the radius is $5\text{ m}$, the patch’s area is increasing at $4\pi\text{ m}^2$ per minute. The rate belongs to that instant, not necessarily to the entire motion. Continue with connected rates of change.

Interpret model parameters

Parameters describe features of a whole model. Interpret them from the model’s structure and units, not from the letter used.

Worked example 8: exponential decay

The mass of a medicine in the body is modelled by

$$M(t)=80e^{-0.18t},$$

where $M$ is measured in milligrams and $t$ in hours.

At $t=0$,

$$M(0)=80,$$

so $80$ mg is the initial modelled mass.

The exponent must be dimensionless, so $0.18$ has units $\text{h}^{-1}$. It is the continuous decay constant, not ”$0.18$ mg lost per hour”. The percentage remaining after one hour is

$$100e^{-0.18}\%=83.53\ldots\%,$$

so the model predicts a decrease of about $16.5\%$ during the first hour.

For more parameter interpretation, see functions in modelling and series in modelling.

Self-check 2

The population of a colony is modelled by

$$N(t)=350(1.06)^t,$$

where $t$ is measured in years. Interpret $350$ and $1.06$.

Answer

$350$ is the modelled population when $t=0$. The factor $1.06$ means the population is multiplied by $1.06$ each year, corresponding to annual growth of $6\%$. It does not mean that $1.06$ organisms are added each year.

Interpret integrals and accumulated change

An integral accumulates a rate. If $v(t)$ is velocity, then

$$\int_a^b v(t)\,dt$$

is displacement, not necessarily distance travelled. Negative velocity contributes negative displacement.

Worked example 9: displacement is not distance

A particle has velocity

$$v(t)=2t-4\text{ m s}^{-1},\qquad 0\leq t\leq5.$$

Its displacement is

$$\begin{aligned} \int_0^5(2t-4)\,dt &=\left[t^2-4t\right]_0^5\\ &=5\text{ m}. \end{aligned}$$

Thus the particle finishes $5\text{ m}$ in the positive direction from its initial position.

For distance, first find when the direction changes:

$$2t-4=0\Rightarrow t=2.$$

Therefore

$$\begin{aligned} \text{distance} &=-\int_0^2(2t-4)\,dt+\int_2^5(2t-4)\,dt\\ &=4+9\\ &=13\text{ m}. \end{aligned}$$

Displacement includes direction; distance is total path length. Review calculus in kinematics.

Interpret signs in mechanics

A negative answer usually means motion or force opposite to the chosen positive direction. It does not automatically mean an error.

Worked example 10: negative velocity

Take upward as positive. A stone’s velocity after $3$ seconds is calculated as

$$v=-7.4\text{ m s}^{-1}.$$

The correct interpretation is

$$\boxed{\text{After }3\text{ s, the stone is moving downwards at }7.4\text{ m s}^{-1}.}$$

Velocity is $-7.4\text{ m s}^{-1}$; speed is $7.4\text{ m s}^{-1}$. Always state the positive direction before forming signed equations. See kinematics language and modelling in mechanics.

Worked example 11: a negative force candidate

Suppose a calculation that assumed a string was taut gives tension

$$T=-3\text{ N}.$$

A string cannot push, so it cannot exert negative tension. The result shows that the assumed taut-string model is invalid in that situation. The physical conclusion is that the string would become slack, and the motion must be reconsidered using a different model.

This is more informative than simply changing $-3$ to $3$.

Interpret probabilities and expected values

A probability is a long-run modelled proportion or a measure of uncertainty. It is not a guarantee about one trial.

Worked example 12: probability is not a prediction of certainty

If the probability that a manufactured item is defective is $0.03$, then in a batch of $500$ the expected number of defective items is

$$500(0.03)=15.$$

Interpretation:

$$\boxed{\text{The model predicts }15\text{ defective items on average in batches of }500.}$$

It does not claim that every batch contains exactly $15$. The actual count varies from batch to batch.

Worked example 13: conditional probability

Suppose

$$P(R\mid S)=0.72,$$

where $R$ is “passes the road test” and $S$ is “attended the revision session”. This means:

Among candidates who attended the revision session, the modelled probability of passing is $0.72$.

It does not mean that $72\%$ of those who passed attended the session. That would concern $P(S\mid R)$, which is generally different. Review conditional probability.

Interpret correlation and regression cautiously

For a product moment correlation coefficient $r$:

• the sign gives the direction of linear association;
• $|r|$ indicates the strength of linear association;
• correlation does not by itself establish causation.

Worked example 14: what a correlation does and does not say

For a sample of towns, the correlation between mean daily temperature and ice cream sales is $r=0.86$.

A justified conclusion is:

$$\boxed{\text{There is a strong positive linear correlation between temperature and ice cream sales in the sample.}}$$

It is not justified to conclude that increasing ice cream sales causes temperature to rise. A lurking variable, season, may influence both.

If a regression model is fitted, interpolation within the observed data range is generally safer than extrapolation beyond it. A good fit over one interval need not continue outside that interval. See correlation and regression.

Interpret hypothesis tests in the language of evidence

A hypothesis test does not prove that a hypothesis is true or false. It measures how compatible the data are with the null hypothesis $H_0$ under the model.

Use conclusions such as:

$$\text{"There is sufficient evidence at the }5\%\text{ significance level to suggest that..."}$$

or

$$\text{"There is insufficient evidence at the }5\%\text{ significance level to suggest that..."}$$

Worked example 15: translate a test decision

A company claims that $40\%$ of customers choose option A. A test uses

$$H_0:p=0.40, \qquad H_1:p>0.40.$$

The calculated $p$-value is $0.018$. At the $5\%$ significance level,

$$0.018<0.05,$$

so reject $H_0$. A complete conclusion is

$$\boxed{\text{There is sufficient evidence at the }5\%\text{ level to suggest that more than }40\%\text{ of customers choose option A.}}$$

This conclusion matches the direction of $H_1$ and refers to customers, not merely to $p$.

If instead the $p$-value were $0.18$, we would fail to reject $H_0$. We would not say that the data prove $p=0.40$. The data would provide insufficient evidence for $p>0.40$. Review hypothesis testing language and binomial hypothesis tests.

Self-check 3

A researcher tests whether the mean battery life is less than $10$ hours. At the $1\%$ significance level, the null hypothesis is not rejected. Which conclusion is valid?

1. The mean battery life is exactly $10$ hours.
2. The mean battery life is at least $10$ hours.
3. There is insufficient evidence at the $1\%$ level to suggest that the mean battery life is less than $10$ hours.

Answer

Statement 3 is valid. Failure to reject the null hypothesis is not proof that it is true, so statements 1 and 2 claim more than the test establishes.

State limitations without dismissing the model

Models are deliberately simplified. A useful interpretation distinguishes what the model predicts from what reality must do.

Common limitations include:

• a population cannot grow exponentially forever because resources are finite;
• constant acceleration may be reasonable only over a short interval;
• a particle model ignores an object’s dimensions and rotation;
• regression beyond the observed data range assumes the trend continues;
• a normal model may assign tiny probabilities to physically impossible values;
• sampled data may not represent the intended population.

Worked example 16: challenge an unrealistic prediction

A plant’s height in centimetres is modelled by

$$H(t)=12+4.5t,$$

where $t$ is weeks after planting. The model predicts

$$H(100)=462\text{ cm}.$$

The calculation follows the formula, but the contextual conclusion should be cautious:

The linear model predicts a height of $462\text{ cm}$ after $100$ weeks, but this is a long extrapolation. Real plant growth will not remain constant indefinitely, so the prediction may be unreliable.

Do not say simply that the answer is “wrong”. It is a correct prediction from a model whose assumptions are doubtful at that time.

Common misconceptions

”A negative answer must be rejected”

Not always. Negative velocity gives direction, negative acceleration refers to the chosen axis, and negative displacement indicates final position relative to the start. Reject a negative value only when the quantity or model forbids it, such as mass, time after the start or string tension.

”The calculator value is the final answer”

The display lacks meaning until you name the quantity, attach units, choose suitable accuracy and check contextual restrictions.

”Expected means what will happen”

An expected value is a long-run mean under the probability model. It need not be a possible outcome. For a fair six-sided die,

$$E(X)=3.5,$$

although a single roll can never equal $3.5$.

”Rejecting the null hypothesis proves the alternative”

A test provides evidence at a stated significance level. It does not deliver deductive proof, and its conclusion depends on the model and sampling process.

”Correlation explains why”

Correlation measures association, not mechanism. Causation requires suitable study design and further evidence.

Mixed self-check

Question 1

A car’s stopping distance is modelled by $d=0.006v^2+0.3v$, where $d$ is in metres and $v$ is in kilometres per hour. Solving $d=35$ gives $v=55.39\ldots$ and $v=-105.39\ldots$. Interpret the answer to the nearest kilometre per hour.

Answer

Speed cannot be negative in this model, so reject the negative root. The model predicts that a stopping distance of $35\text{ m}$ corresponds to a speed of approximately

$$\boxed{55\text{ km h}^{-1}}.$$

Question 2

The gradient of a distance against time graph is $6$ at $t=10$, with distance in kilometres and time in hours. Interpret the value.

Answer

At $t=10$ hours, the distance is increasing instantaneously at $6\text{ km h}^{-1}$. This is a speed at that instant. It is not necessarily the average speed over the first $10$ hours.

Question 3

A regression line based on data for children aged $5$ to $12$ predicts a height of $196\text{ cm}$ at age $20$. What should you say?

Answer

$196\text{ cm}$ is the regression model’s prediction, but using the line at age $20$ is extrapolation beyond the observed range. The relationship between age and height is unlikely to remain linear, so the prediction is unreliable.

Question 4

With east positive, a particle’s displacement after an interval is $-18\text{ m}$. Interpret this result.

Answer

The particle’s final position is $18\text{ m}$ west of its initial position. This does not reveal the total distance travelled.

Final checklist

Before leaving a contextual question, check:

• Quantity: Have I named what the number represents?
• Units: Are they present and correctly powered?
• Sign: Have I translated direction or decrease correctly?
• Domain: Is the answer within the model’s allowed interval and physically meaningful?
• Accuracy: Did I compare using unrounded values and report suitable precision?
• Statistics: Have I avoided claims of certainty, proof or causation?
• Model: Have I separated a model prediction from reality and noted relevant limitations?
• Wording: Does my final sentence answer the precise question asked?

Next, practise selecting an approach in choosing a method, making your working persuasive in communicating mathematical reasoning, and spotting errors with common A level Mathematics exam mistakes.