The first statPREP kickoff workshops will be held in June 2017.

Here is an interactive document to give a taste of statPREP computing. The workshop tutorials, example lessons, and other materials will be delivered in this format. Note that you need only an ordinary browser to use these materials, but all the commands can also be directly used in the usual R computing environment.

The statPREP workshops are divided into two main parts. The first centers on building skills in data computing: mastering the remarkably small set of commands that are used in the statPREP lessons and activities (and which would suffice for the computing content of an entire introductory statistics course). The second on a lesson plan, selected by the workshop participant, that focusses on a single lesson that can be introduced to your statistics course to help you and your students benefit from the data-centric teaching that is the heart of the statPREP approach.

## Skill Building

Please note that parts of these tutorials are still in draft form.

Introduction to the tutorial system

- Introduction to data tables
- The basic vocabulary of R
- Formulas and the modeler’s tilde
- Graphics
- Classical statistical calculations
- Stats on quantitative variables
- Counts and proportions
- Hypothesis tests: t, chi-squared, ANOVA

- Sampling variation and confidence intervals
- Sampling and resampling
- Confidence intervals (not yet available)

## Example Lessons

### Displaying variation ✓

This tutorial introduces graphics to show the distribution of a variable, in this case birth weight.^[Still to be added: apps for adjusting smoothing parameters, exercises. Source file on GitHub.]

### How big a sample? ✓

This tutorial starts with real ballot-by-ballot data from which we can draw random samples. By comparing results from a sample to the actual results, we’ll determine whether and to what extent a sample can be representative and predictive.

### Small n and big t* ✓

This lesson introduces the idea behind Student’s t, that when working with small n the standard error of the mean has a different distribution than with large n. You and your students can explore how large n needs to be to justify working with t* = 2. Conversely, you can see how very small n calls for inflating t*.

### What’s normal?

Starting with the everyday meaning of “normal,” we’ll use data to construct an operational definition of “normal.” One example will make use of a census of all 4-million registered births over a year in the US, including birth-by-birth measures of pregnancy risk factors, maternal and newborn outcomes, and demographics. This will segue into more abstract measures of “normal,” such as the shape and parameters of the famous normal distribution.

### Which way did the probability go?

Using medical diagnosis and outcomes data, you’ll show students the distinction between the likelihood of an outcome given a diagnosis and the accuracy of the diagnosis itself.

Materials for the workshop are under active development. A snapshot of the materials are at our development site.