Confidence Intervals Explained With My Numbers
Understand what an interval does and does not say.
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
Explain confidence intervals using my own numbers. Sample statistic and size: [SAMPLE_STATISTIC_AND_N]. Variable type (mean, proportion, difference): [VARIABLE_TYPE]. The decision I want to make: [DECISION]. Do this in order: 1. Compute the interval step by step with the formulas written out: for a mean, estimate plus or minus t x s / sqrt(n); for a proportion, p plus or minus 1.96 x sqrt(p(1-p)/n) as a rough guide. State the assumptions and whether they hold for my data. 2. Explain in plain words what a 95 percent interval means, using the idea of repeating the study many times, and list three common misreadings (95 percent chance the true value is inside this interval; 95 percent of data lies inside; non-overlap needed for a difference). 3. Show how the width changes if my sample were four times larger (it halves), with arithmetic. 4. Draw the interval as a text diagram with the estimate and a decision threshold, and say what each position of the threshold would imply for my decision. 5. Give the code to reproduce it, plus a bootstrap interval as a check. 6. Write two sentences I can say to a manager, and one sentence to avoid.
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.