Unit 3: Collecting Data

AP Statistics: 51 practice questions with detailed explanations.

Unit Study Guide

Executive Summary

How data is collected decides what conclusions are valid. Random sampling enables generalization; random assignment enables causation.

Sampling

A census measures everyone; a sample measures a subset. Simple random samples give every individual an equal chance. Stratified samples take random draws within groups; cluster samples measure everyone in randomly chosen groups; systematic picks every kth. The sampling frame is the list you draw from — undercoverage misses groups not on it.

Bias

Convenience and voluntary-response samples skew. Nonresponse bias: those who answer differ from those who don't. Response bias: wording or lying distorts answers. Bigger samples reduce variability but never fix bias.

Experiments

Experiments impose treatments; observational studies only watch. Random assignment spreads lurking variables across groups, enabling causal conclusions. Control groups provide comparison; placebos and blinding prevent expectation effects; blocking groups similar subjects; matched pairs use each subject as its own control.

Inference

Random sampling → generalize to the population. Random assignment → cause-and-effect. Confounding: two variables' effects can't be separated. Replication (many subjects) reduces chance variation.

Exam traps

Sampling and assignment solve DIFFERENT problems — don't mix them up. Larger samples do not remove bias. Observational studies cannot prove causation.

Top 5 Concepts to Master

  1. 1Match designs to bias types.
  2. 2Choose the right sampling method.
  3. 3Explain what random assignment achieves.
  4. 4State what conclusions a design supports.

Key Terms & Definitions

Practice with Flashcards
Census

Data from the entire population.

Simple random sample

Every individual equally likely to be chosen.

Stratified sample

Random draws within each subgroup.

Cluster sample

Random groups, everyone inside measured.

Undercoverage

Groups missing from the sampling frame.

Confounding

Two variables’ effects cannot be separated.

Random assignment

Chance allocation of treatments.

Double-blind

Neither subjects nor evaluators know the treatment.

Common Misconceptions: Exam Traps

A bigger sample fixes biased sampling.

Correct: Size reduces variability; bias is about selection.

Observational studies can prove causation if large enough.

Correct: Only experiments with random assignment support cause.

Random sampling and random assignment do the same job.

Correct: Sampling generalizes; assignment enables causation.

Voluntary response surveys represent the population.

Correct: Self-selection attracts strong opinions.

Question Bank Breakdown

By difficulty

easy 21medium 30

By topic

Experimental Design 14Inference from Studies and Experiments 9Introduction to Data Collection 8Bias in Sampling 8Sampling Methods 7Random Assignment 5

All Questions in this Unit