"Correlation does not imply causation" might be the most repeated phrase in statistics, but the SAT will still try to trick you with it. These questions ask whether a study's design actually supports the conclusion being made. Focus on three things: random assignment (needed for causation), random selection (needed for generalization), and sample size. If a study only surveys volunteers from one school, you can't generalize to all teenagers. Watch for answer choices that overreach what the data allows.

Outliers love to hide in SAT data sets and wreck your mean while barely touching the median, something that many students miss. One-variable data problems reward sharp attention to measures of center and spread. You'll find mean, median, mode, range, and interquartile range from raw data, frequency tables, or box plots. Know when an outlier distorts the mean but not the median, pick the right statistic for skewed distributions, and avoid confusing IQR with full range.

Successive 20% discounts do not equal a 40% discount, yet many students fall for that trap and lose points. Percentages on the SAT reward precision across realistic scenarios. Expect questions on percent change, compound percentage applications, determining what percent one number is of another, and reversing a percentage to find the starting value. Translate "percent" to "/100" immediately, apply changes step-by-step instead of just adding them, and test your answer by working backward to catch common errors.

Probability questions on the SAT can feel straightforward until a conditional twist or "at least one" phrase turns your quick guess into a wrong answer. Expect questions on simple probability, complementary events (1 minus the probability), independent events (multiply probabilities), dependent events (adjust after each draw), and conditional probability from frequency tables or Venn diagrams. Master interpreting "at least," "exactly," and "or" scenarios correctly. Two-way tables are frequent, test your ability to find joint, marginal, and conditional probabilities fast without mixing rows and columns.

A ratio is just a relationship between two quantities, but the SAT finds creative ways to make that simple idea feel complicated. You'll see ratios hidden inside tables, word problems, and unit conversions. The core skill is always the same: set up a proportion and solve. Watch out for questions that give you a part-to-part ratio when you need part-to-whole, or vice versa. If the ratio of cats to dogs is 3:4, the fraction of cats isn't 3/4. It's 3/7. Mixing this up is one of the most common errors, and the SAT will always include that wrong answer as a choice.

Statistical inference sounds advanced, but on the SAT it usually means reading margin of error and confidence statements without doing heavy calculations, so don't overthink! Expect questions on interpreting margin of error in polls, understanding what a confidence interval or 95% confidence level really means, and recognizing that larger samples reduce error but never prove causation. Watch for distractors claiming surveys prove cause-and-effect or misstating what the margin covers. Focus on literal reading of survey results and sample-size effects to score these reliably. Key reminders: margins shrink with bigger samples, confidence levels describe long-run success of the method, not certainty for one poll, and surveys show association, never proof of cause.

SAT loves slipping two-variable scatterplots into what looks like a one-variable stats question, catching students who forget the difference. One-variable data means single lists or distributions: find center (mean/median), spread (range/IQR), outliers, and read dot plots or histograms. Two-variable data involves paired values: analyze scatterplots for positive/negative correlation, clustering, or outliers, plus lines of best fit. One-variable is about describing a group; two-variable is about how two groups relate.