Applied statistics (CPH: BD)

Frontpage Lecture plan Vidcasts Datasets R install help

Estimation

Literature

[A] chapter 5, except section 5.5.

Lecture material

This lecture as: slideshow (html), Rmarkdown (Rmd), notes (pdf).

The entire module: notes (pdf).

Answers to exercises in the book

Exercises

  1. Solve Agresti exercise 5.7, 5.17, 5.19 (use RStudio), 5.31, 5.37 and 5.41 (hint: which standard deviation is necessary for the normal distribution with mean 9 to cover the interval 0-18 with probability 99.7%?)

  2. Confidence intervals:

    If you work on your own machine you should start by installing the package TeachingDemos (already installed on RStudio server):
    install.packages("TeachingDemos")
    Then you load it:
    library(TeachingDemos)

    Try the function call ci.examp() to visualize 50 confidence intervals.
    You can study the arguments of ci.examp by the command ?ci.examp.
    Try to change some of the arguments of the function and use method="t". Try e.g. to generate 40 random samples (reps=40) and plot the corresponding 40 confidence intervals with a confidence level of 90% (conf.level=0.90). Convince yourselves that we expect 4 of these confidence intervals not to contain the population mean.
    How does this fit with what you see?

  3. Go through any Rmarkdown exercises of the previous two lectures that you did not yet finish.