Which feature permits a cause-and-effect conclusion?
Collecting Data. Best move: Random assignment
Random assignment balances lurking variables and supports causal inference.
Use a short diagnostic to expose the first decision that needs repair.
Use the statistics subject guide, statistics practice questions, the statistics formula sheet, the Probability and Statistics course to move from an exam signal to the explanation, formula, practice set, or course sequence that repairs it.
Answer from your current habits. The feedback points to a stronger exam-preparation move.
Collecting Data. Best move: Random assignment
Random assignment balances lurking variables and supports causal inference.
Inference. Best move: Results at least this extreme assuming the null
A p-value is conditional on the null hypothesis and measures extremeness.
Regression Analysis. Best move: Observed minus predicted
Residual = observed response - predicted response.
The published format is 42 questions on multiple choice and 4 questions on free response. Multiple choice - About 2.1 minutes each. Reading the scenario is most of the work; the arithmetic is usually small. Free response - About 22 minutes per question. These are long, multi-part scenarios - Read all parts before writing part (a).
Start with the heaviest published weightings: One-Variable Data and Collecting Data (20–30%), Probability, Random Variables, and Distributions (15–25%), Inference for Categorical Data: Proportions (15–25%). Then repair whichever earlier unit those depend on.
Only where the exam allows one, and practising otherwise builds the wrong habit. Multiple choice: Use it for normalcdf, invNorm, and binomial probabilities instead of the tables when speed matters. Free response: The calculator gives the number; the rubric pays for the name of the procedure and the conditions. Provided for the whole exam: Formulas and the normal, t, and chi-square tables are given. Do not spend study time memorising them - Spend it on when each applies.
Never write 'accept the null hypothesis' or 'this proves'. You either reject or fail to reject, and neither proves anything.