Question
Name the population, variable, and decision the analysis must support.
A 7-stage path beginning with Probability axioms and ending with Limit theorems, with recall checks between major ideas.
Build the underlying ideas in the Statistics subject guide, use the statistics calculator to check a completed setup, and return to Probability practice questions when you are ready to retrieve the method without a model.
Short cycles create evidence that the idea is becoming usable rather than merely familiar.
Name the population, variable, and decision the analysis must support.
Check how the data and assumptions permit the chosen method.
Return every numerical result to its units and context.
Inspect uncertainty, alternatives, and conclusions the evidence cannot support.
Later conclusions depend on how the question, data, assumptions, and model were established earlier.
Each unit names the actual skills, demonstrates one complete case, and identifies the mistake most likely to change the result.
You can lay out a sample space, assign probabilities that obey the basic rules, and compute the chance of an event. You can also use complements and unions correctly.
Interpretation risk: Counting unordered pairs only: finding 3 pairs out of 21 gives about 0.143, because it treats (2,5) and (5,2) as the same single outcome.
You can update a chance once you learn something new, and run Bayes' rule. You can also test whether two events really are independent.
Interpretation risk: Dividing by the wrong event: computing 0.2/0.4 = 0.5, which is the chance of B given A, not A given B.
You can use binomial, geometric and Poisson models for counts, and find their means and variances. You can also choose the right model from the story alone.
Interpretation risk: Leaving out the count of arrangements: reporting 0.03125, which is the chance of one exact order such as heads-heads-heads-tails-tails.
You can use density curves where probability is area under a curve, including uniform, exponential and normal shapes. You can also find percentiles with z-scores.
Interpretation risk: Subtracting without dividing by the standard deviation: using z = 8 puts the answer at essentially 100%, when the true figure is 84%.
You can handle two random variables at once, pull out each one's own distribution, and test independence. You can also compute a covariance.
Interpretation risk: Declaring independence from one matching cell: independence requires every cell to equal the product of its marginals, and one match proves nothing.
You can find means and variances of sums and rescaled variables without rebuilding the whole distribution. You can also add expectations even when the variables are linked.
Interpretation risk: Adding the 5 to the variance as well: reporting 3(4) + 5 = 17, when sliding a variable up by 5 does not change how spread out it is.
You can explain why sample averages cluster near the true mean and turn bell-shaped. You can also use that bell shape to approximate a count.
Interpretation risk: Dividing by n rather than the square root of n: reporting 0.2, ten times too small, which makes every interval look impossibly precise.
Start with the earliest unit whose definitions or notation you cannot explain from memory. Later units often depend on that foundation even when the current assignment looks different.
Move on after you can choose the method on a changed problem, complete it without copying a model, and check the answer independently.
Begin without a hint, record the first uncertain decision, repair only that connection, and then answer a fresh question to test whether the repair transferred.