NOTES FOR STATISTICS
William Knight
University of New Brunswick
=======
Under Construction
=======
DESCRIPTIVE STATISTICS
Preface
to descriptive statistics
One Dimension
Frequencies and Histograms
Shapes of distributions
A common shape ("bell curve")
Where is the middle? Median and Average
Quantiles -- Quartiles, percentiles, etc.
Five point summary
and
Box plot
Spread
Conversion: /
Temperature
/
Nonlinear
|O|
Exercises, In class work, Other supplementary material ...
Other web sites for univariate descriptive statistics
(( may be obsolete ))
HyperStat Online
--
"http://davidmlane.com/hyperstat/desc_univ.html">
U. of California at Berkely
#2
#3
Two Dimensions:
Scatter Plots and Lines
Intro to scatterplots at
Some scatter plot types
Scatter plot examples
Summarizing a scatterplot by a line
other material
http://www.ruf.rice.edu/~lane/hyperstat/A115370.html
Practice with plots and lines
, Illustrated with growth of young Toronto males
The "Least Squares" criterion
Dangers and misuse of summing a scatter plot with a line:
Example
of data for which a a line
should not
be fitted
See also "Regression warnings" at the
bottom
of
http://www.mathcs.duq.edu/larget/math225/9-2.html#reg-line
Other web sites for bivariate descriptive statistics
Presenting data for two continuous measurements - SurfStat Australia
--
http://surfstat.newcastle.edu.au/surfstat/main/1-3-1.html
Probability & Statistics, Univ. of the Pacific -- Lecture 23
--
http://jarl.cs.uop.edu/cop/math/chance/coursenotes/lecture23.html
Analogies
Between One and Two Dimensions
BAD DATA
All Data Has Errors
Examples
SAMPLING A POPULATION
Population and sample
How to get a (simple) random sample
The sample
IS NOT
the population
The Square Root Law
Other web pages on sampling
INFERENCE
Confidence intervals -- How close to the
real
number?
Confidence intervals --- General
Confidence interval for
Population Median
Confidence interval for
Population Mean
-------
( a refinement)
Confidence interval for
Population Total
{
New page
}
Confidence intervals for
sums and differences of means
{
New page
}
Cheating on sample size
SCIENTIFIC EXPERIMENT
Seasickness remedy
, a parable in scientific method.
((
associated calculations
))
Sunburn
a lesson in experimental design.
Other web pages
Scientific method at
http://www.anu.edu.au/nceph/surfstat/surfstat-home/2-1-3.html
Practice examples at
http://www.isds.duke.edu/~erik/sta110/notes/design_examples/design_examples.html
See "Experiments" under
http://www.stat.berkeley.edu/~stark/SticiGui/Text/index.html
DOES TREATMENT REALLY BEAT CONTROL?
Sign Test
U-Test
Key to some common tests
EPILOG
APPENDICES or POSTSCRIPTS
Tables
A few public domain tables
Techniques - not ideas
Using normal distribution tables
Copyright Notice -
See title page
William Knight
