Interpolation and extrapolation on ACT Science: estimating a value between known data points and extending a trend beyond the measured range, while flagging the greater uncertainty of extrapolation.

What this topic is asking

Not every ACT Science value sits exactly on a labelled gridline or in a table row. Sometimes you must estimate between known points (interpolation), and sometimes you must predict beyond the measured data (extrapolation). Both are Interpretation of Data skills, and both come down to following the trend the data already show. The ACT also tests whether you understand that extending a trend too far is less certain.

Interpolation: estimating between points

Interpolation fills the gap between data you have. The method:

1. Identify the two known points that bracket the value you need.
2. Judge where your target sits between them (a quarter of the way, halfway, three-quarters).
3. Estimate the value the same fraction of the way between the two known results.

For a roughly linear trend, a target halfway between two x-values has a y-value about halfway between the two y-values, that is, their average. For example, if a quantity is 10 at x = 2 and 20 at x = 4, then at x = 3 it is about $\frac{10 + 20}{2} = 15$.

Extrapolation: predicting beyond the data

Extrapolation extends the trend past the last data point. The method:

1. Establish the pattern in the measured range (for a line, the constant step or rate).
2. Continue that pattern to the x-value you need.
3. Read off the predicted y-value.

For a straight line that rises by a fixed amount each step, keep adding that amount. If distance rises by 100 m every 10 s through 30 s, then at 40 s it is $300 + 100 = 400$ m.

Why extrapolation is less certain

The key conceptual point the ACT tests is that extrapolation carries more uncertainty than interpolation. Inside the data, you have evidence the trend holds. Outside it, you are assuming the trend continues, and many real relationships level off, reverse, or change shape beyond the measured range. A reaction rate that rises with temperature will eventually fall once an enzyme denatures; a cooling curve flattens as it approaches room temperature. When a question asks you to judge the reliability of a prediction, "less certain because it extends beyond the data" is often the right idea.

How the ACT frames these

• Interpolation questions often say "estimate," "approximately," or "closest to," with the target value between two table entries or gridlines.
• Extrapolation questions often say "if the trend continues" or "predict," with the target value beyond the last data point.
• Reliability questions ask whether such a prediction can be trusted, where the answer turns on whether the point is inside or outside the data.

Try this

Q1. A quantity is 12 at x = 10 and 24 at x = 14, rising steadily. Estimate its value at x = 12. [2 points]

• Cue. x = 12 is halfway between 10 and 14, so the value is about $\frac{12 + 24}{2} = 18$.

Q2. Explain why a prediction made by extrapolation is less reliable than one made by interpolation. [2 points]

• Cue. Interpolation estimates within the measured range, where the trend is known to hold; extrapolation assumes the trend continues beyond the data, where it may level off, reverse, or change shape.