Skip to main content

SAT Math · Problem-Solving and Data Analysis

SAT Inference from Sample Statistics

These questions ask what a random sample can tell you about a larger population, and what a margin of error means for an estimate.

What to watch for

  • A result from a random sample applies only to the population it was drawn from.
  • A margin of error gives a range of plausible values, not a guarantee.
  • Larger samples usually give smaller margins of error.

Worked examples

Example 1

A water utility selected 240 households at random from all the households it serves. The utility used the sample to estimate that the mean daily water use of all households it serves is 412 liters, with an associated margin of error of 18 liters.

Which of the following is the most appropriate conclusion?
  1. A.The mean daily water use of all households the utility serves is exactly 412 liters.
  2. B.It is plausible that the mean daily water use of all households the utility serves is between 394 liters and 430 liters.
  3. C.It is plausible that most households the utility serves use between 394 liters and 430 liters of water per day.
  4. D.It is plausible that most households the utility serves use between 412 liters and 430 liters of water per day.
Show the answer and explanation

Answer: B

Choice B is correct. Subtracting the margin of error from the estimate and adding it to the estimate gives 412−18=394412 - 18 = 394 and 412+18=430412 + 18 = 430. These are the bounds of the interval of plausible values for the actual mean, so it is plausible that the mean daily water use of all households the utility serves is between 394 liters and 430 liters. Choice A is incorrect because it treats the estimate as exact. Choice C is incorrect because the interval describes the mean, not the water use of individual households. Choice D is incorrect because it also describes individual households, and it uses only the half of the interval above the estimate.

Example 2

A city selected a random sample of 500 households in its north district and a separate random sample of 500 households in its south district. Based on the samples, it is estimated that 41% of north district households and 46% of south district households compost food scraps. The margin of error for each of these two estimates is 5%.

Which of the following is the most appropriate conclusion?
  1. A.The data do not establish that the percents of households that compost food scraps differ in the two districts.
  2. B.The percent of all south district households that compost food scraps is greater than the percent of all north district households that do.
  3. C.It is plausible that fewer than 35% of north district households compost food scraps.
  4. D.The two estimates cannot be compared, because the two districts may have different numbers of households.
Show the answer and explanation

Answer: A

Choice A is correct. Plausible values are 41−5=3641 - 5 = 36 to 41+5=4641 + 5 = 46 percent for the north district and 46−5=4146 - 5 = 41 to 46+5=5146 + 5 = 51 percent for the south district. The intervals overlap from 41% to 46%, so the same percent is plausible for both districts. Therefore, the data do not establish that the percents of households that compost food scraps differ in the two districts. Choice B is incorrect because the overlap means a difference is not established. Choice C is incorrect because 35% is below the north district's interval. Choice D is incorrect because each margin of error comes from its own random sample.

Practice inference from sample statistics

More SAT inference from sample statistics questions at every difficulty, with explanations in your language.