Applied maths 9.1: Sampling distributions
Why an average from a sample wobbles from sample to sample, and how much: the sampling distribution and the standard error. A poll of 1,000 voters says 52 % back a plan. Ask a different 1,000 andโฆ
9 posts match
Why an average from a sample wobbles from sample to sample, and how much: the sampling distribution and the standard error. A poll of 1,000 voters says 52 % back a plan. Ask a different 1,000 andโฆ
What makes a good estimate (bias and spread), and the maximum likelihood method that finds one for almost any model. A bent coin lands heads 7 times in 10. Most people say the chance of heads isโฆ
Turn an estimate and its standard error into a range of believable values, and say correctly what 95 % confidence means. A news report says 52 % ยฑ 3 %, 19 times out of 20. Few readers know whatโฆ
Test a claim with data: the null hypothesis, the test statistic, the p-value, and what a p-value does and does not mean. A cereal maker claims 500 g per box. You weigh 36 boxes and the average isโฆ
The two ways a test can be wrong, the power to detect a real effect, and how many observations a study needs. Many studies are too small to find the effect they are looking for, and then report 'noโฆ
Estimate the uncertainty of any statistic by resampling your own data: bootstrap standard errors and percentile intervals. There is a neat formula for the standard error of a mean. For a median, aโฆ
Combine what you believed before with what the data say: prior, likelihood, posterior, and the Beta model for proportions. A coin you know nothing about lands heads 3 times out of 3. Maximumโฆ
Fourteen questions across statistical inference, with new numbers on every attempt. Pass at 80 % to move on to unit 10. This is where you find out what you can do without help. Each question is newโฆ
Bipolar and MOSFET transistors, used as switches for loads and as amplifiers for signals. A microcontroller pin can supply a few milliamps. A car's headlights need 10 amps. A three-legged componentโฆ