AGES 9-14 - GUIDED STATISTICS
The Resampling Bag
A guided bootstrap demo: turn one original sample into a bag, draw the same number of values with replacement, and see how the sample mean can wiggle.
Bootstrap resampling
same n, draw with replacementOriginal sample becomes the bag
Use the sample as the bag
The unknown population is not available here.
Put each draw back
The same observed value can appear more than once.
Draw exactly n values
Each resample has the same size as the original sample.
Look at the means
The spread estimates the standard error of the sample mean.
Means from bootstrap resamples
Latest bootstrap resample
original sample
bootstrap means
middle interval
original mean
Controls
each resample uses exactly n draws
Jump ahead
Live stats
!
Bootstrap idea. We usually do not know the whole population.
The bootstrap treats the original sample as a stand-in bag.
Each resample draws exactly the original sample size, with replacement, so repeats are allowed and some original values may be left out.
Repeating this many times shows how the sample mean could wiggle if this observed sample were our best picture of the population.
The SD of the bootstrap means estimates the standard error of the original mean, not the SD of the raw values.
good original sample needed