lecture17.md

Lecture 19 (Week 9, Wednesday)

Sampling Distributions to Confidence Intervals (continued)

Comparing simulated and randomized sampling distributions of b₁

• why do you think the simulated sampling distribution looks more normal?
• simulated sampling distribution came from a normal distribution
• we don't know the shape of the actual data

Confidence Intervals

• we can construct a new copy of the sampling distribution, centered at our sample mean

Using the Randomized Sampling Distribution

• how can we calculate the confidence interval from this distribution?
  • move this distribution so that it's centered at 1.9
  • then find the lowest and highest 2.5%
  • or, calculate the standard error based on this distribution

Finding the 95% Cut-off Points

• how do we get from this to a confidence interval?

confint(lm(Games~Condition, data = ScienceData))