|
Activity Number:
|
196
- SPEED: Teaching Statistics: Strategies and Applications
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2018 : 10:30 AM to 11:15 AM
|
|
Sponsor:
|
Section on Statistical Education
|
|
Abstract #333034
|
|
|
Title:
|
Introducing Forecast Intervals with a Confidence Game
|
|
Author(s):
|
Robin Lock*
|
|
Companies:
|
St. Lawrence University
|
|
Keywords:
|
prediction interval;
time series;
games;
statistics education
|
|
Abstract:
|
Most statistics students deal with confidence intervals for estimating population parameters but get less experience with the more natural question of finding an interval that is likely to contain a future value from the population. The latter type of forecast interval is particularly relevant in a time series setting. We describe an in-class game where students are given some past history and asked to provide intervals that are likely to contain future values of a series. The scoring system rewards smaller intervals that successfully capture the value but imposes a penalty for intervals that miss the mark. We discuss how students do at the game and give audience members a chance to try playing.
|
Authors who are presenting talks have a * after their name.
