April 5
our story so far
Any questions about confidence intervals or anything we've
done so far?
thinking about projects
I've put up an assignment for you to turn in for Thursday,
telling me what you have in mind for a term project.
new material
H0 = null hypothesis (No result: data is due to randomness.)
HA = alternative explanation (Typically what we want to show.)
alpha = significance level = 1 - confidence (from confidence level)
= chosen threshold of "yes we have a non-random result", fixed before experiment
= 0.05 (typical)
p-value = probability of measured experimental result, if H0 (null hypothesis) is true.
... source of much confusion and error in science
Then the possibilities are given in a 2x2 table :
reality / decision do_not_reject_H0 reject_H0
----------------------------------------------------------------------------------
H0 | OK : no effect type 1 error (find effect incorrectly)
|
not H0 | type 2 error (miss real effect) OK : find real effect
update:
I have a (fixed) example in sleepers.R that's similar to one of the examples in the text.
Modify the hypothesis & data in sleepers.R to illustrate all four of the possible outcomes,
and discuss.
Discuss the "legal system jury" version of this: innocent or guilty, and the "guided practice"
4.23 & following.
Do some exercises from the text:
But don't treat p-values as a magic bullet ...
P-values have become somewhat controversial lately. Not because the ideas
behind them are wrong, but because it is dangerously tempting to misuse
them unintentionally, by for example reading too much into a low p-value,
or by publishing only when the p-value looks good (and not publishing a negative result).
Here are a few related articles.
how do we know ... and related topics
The notion of a "hypothesis testing framework" is part of
a larger topic, of how we (statistically) know what we think we know.
skeptical inquirer
wikipedia articles
![[paper clip]](/courses/source/wiki_images/paper_clip_tilt.png)
name last modified size