Reflect on the CLT

Schwab

Explore the CLT

1 Distributions

2 Confidence Intervals

3 Confidence Interval (activity)

Data Distribution
- How does the sample data look?
Sampling distribution
- A theoretical distribution (math model) of proportions from many samples
- centered around the true propotion p.
Bootstrapped Distribution
- If parametric its center is the same as the Null Distribution, center at p.
- If non-parametric it is centered around \(\widehat{p}.\)

This is how we build the sampling distribution.

If we consider many proportions from many samples and the conditions are satisfied, then the distribution of the sample proportions will be Normal.

If we make a 95% CI from 100 point estimates of \(\widehat{p}\) from 100 samples we expect 95 of those to contain the true proportion.