ST 411/511: Methods of Data Analysis
Winter 2016
Quiz 1
Friday January 15, 2016
1. Give a one-sentence definition of a p-value.
The p-value is a function of the observed sample results (a statistic) that is used
for testing a statistical hypothesis.
2. TR

Stat 411/511
SOME OTHER TWO SAMPLE PROCEDURES
Oct 28 2015
Charlotte Wickham
stat511.cwick.co.nz
Your turn
Fill in the column for Wilcoxon Rank
Sum on the test summary
worksheet (posted on web).
Wilcoxon Rank Sum
Wilcoxon Rank Sum test
Null hypothesis*
The

Stat 411/511
ALTERNATIVES TO THE T-TOOLS
Oct 26 2015
Charlotte Wickham
stat511.cwick.co.nz
Two independent samples
Two sample t-test
Randomization test
Wilcoxon Rank Sum
Today
Doesn't assume Normality and is resistant to outliers
Levene's test
Test for

Stat 411/511
EXTRA SS F-TEST
Nov 6 2015
Charlotte Wickham
stat511.cwick.co.nz
Quiz #3
Today noon - Monday noon
Same format, two sections, 30mins
each
Short answer you need to be able to
do exp(x)
The ANOVA table
a convenient way to lay the calculations ou

Stat 411/511
COMPARISON OF MANY GROUPS
Nov 2 2015
Charlotte Wickham
stat511.cwick.co.nz
Chapter 5 Sleuth
More groups
So far we have only looked at two groups, now
we'll look at multiple groups.
In this chapter two big questions:
1. How can we compare two

Stat 411/511
SIGN TESTS
Oct 30 2015
Charlotte Wickham
stat511.cwick.co.nz
Quiz #2
Quiz #2
Researchers are interested in the eect of using a bike for
transport to the grocery store on the amount people spend
at the store for people in Corvallis.
They sta

Stat 411/511
LOG TRANSFORM
Oct 27 2014
Charlotte Wickham
stat511.cwick.co.nz
Log transform
Sometimes assumptions can be met by
transforming the data.
A particularly useful transformation is the
logarithmic transformation.
Useful, when variation increases

Stat 411/511
ASSUMPTIONS OF THE T-TOOLS
Oct 19 2015
Charlotte Wickham
stat511.cwick.co.nz
Announcements
Quiz #2 this weekend.
Same format, same timing.
Study guide posted.
Participation Task
You have been assigned a Participation Number in canvas grades:

Stat 411/511
ASSUMPTIONS, OUTLIERS & LOG
Oct 21 2015
Charlotte Wickham
stat511.cwick.co.nz
Your turn
We are going to do a two sample t-test.
Put these datasets in order from:
I would be very worried about the Normality assumption to
I would not be worrie

Stat 411/511
THE RANDOMIZATION TEST
Oct 16 2015
Charlotte Wickham
stat511.cwick.co.nz
Today
Review randomization model
Conduct randomization test
What about CIs?
Using a t-distribution as an
approximation to the randomization
distribution.
Display 1.1
p.

Stat 411/511
TWO-SAMPLE T
Oct 12 2015
Charlotte Wickham
stat511.cwick.co.nz
Today
The two sample model
The two sample t-test and CI
When sampling isnt random
Two sample sampling model
Two Populations
Every member has
one number
associated with it,
but we

Stat 411/511
ONE WAY ANOVA
Nov 4 2015
Charlotte Wickham
stat511.cwick.co.nz
Summary from Weds
Comparing two groups
A single comparison of two group means
When you have multiple groups:
Do the usual two-sample t-test but,
use all groups to get the pooled
s

Stat 411/511
ANOVA ASSUMPTIONS
Nov 9 2015
Charlotte Wickham
stat511.cwick.co.nz
DA #1 followup
Population inference to all opposite sex married
couple households in Oregon.
(Response rates for ACS ~ 97.9%)
No causal inference.
Gender was not randomly as

ST 411/511: Methods of Data Analysis
Winter 2016
Midterm Review
Scientific questions:
Identify population of interest
Identify variable of interest
Understand (be able to define and use) the population distribution of the variable of
interest
Identify

ST 411/511: Methods of Data Analysis
Winter 2016
Quiz 2
Friday January 22, 2016
1. Explain, in one sentence, why it is difficult (impossible?) to infer a causal relationship
from an observational study.
(1 point) It is difficult or impossible to infer a c

ST 411/511: Methods of Data Analysis
Winter 2016
Quiz 3
Friday January 29, 2016
1. Describe what it means for a method of constructing confidence intervals to produce
valid intervals.
2. List the assumptions necessary for an equal-variance two-sample t-te

ST 411/511: Methods of Data Analysis
Winter 2016
Quiz 5
Friday February 19, 2016
1. State the null hypothesis that is tested by a one-way ANOVA with M groups (be sure to
define your notation).
2. State the alternative hypothesis for a one-way ANOVA with M

ST 411/511: Methods of Data Analysis
Winter 2016
Quiz 6
Friday February 26, 2016
NAME: _
1. Suppose we are comparing five different fertilizer treatments for tomatoes: A, B, C, D,
and E. Each of 30 potted tomato plants is randomly assigned to one of the t

ST 411/511: Methods of Data Analysis
Winter 2016
Quiz 4
Friday February 12, 2016
1. Summarize the differences between the equal-variance two-sample t-test and Welchs
two-sample t-test:
a. How are the test statistics different?
b. How are the reference dis

ST 411/511: Methods of Data Analysis
Winter 2016
Quiz 7
Friday March 4, 2016
1. What assumptions are necessary to produce valid inference for a simple linear regression
model?
2. How do you estimate the subpopulation variance 2 in simple linear regression

ST 411/511 Fall 2016 Outline 1
Reading assignment: Chapter 1. This will be largely review. You should have covered
most of the material in your introductory statistics class.
Chapter 1 Drawing Statistical Conclusions
Case Study 1.1.1 What are the effects

ST 411/511 Fall 2016 Outline 2
Reading assignment: Chapter 2. This chapter reviews the details of one- and twosample t-tests.
Chapter 2 Inference Using t-Distributions
Case Study 2.1.2 Twin study. How big is the difference in volumes of left hippocampus
w

ST 411/511 Fall 2016 Outline 3
Reading assignment: Chapter 3. This chapter explores the consequences of violation
the assumptions of t-tests.
Chapter 3 Assumptions of the t-Tools
Three assumptions needed for t-test and t confidence interval:
Case Study 3.

Stat 411/511
THE ONE SAMPLE T-TEST
Oct 2 2015
Charlotte Wickham
stat511.cwick.co.nz
Today
Recap
We dont know the population SD,
Z -> t
A little bit of R
Hypothesis testing & p-values
from last time
Recap
Population inference is using a sample to learn

Stat 411/511
MORE ON THE RANDOM SAMPLING MODEL
Sep 29 2015
Charlotte Wickham
stat511.cwick.co.nz
Announcements
My oce hours:
Mondays 11am 255 Weniger
Thursdays 3-5pm 3003 Cordley
Help with Statistics Classes
Kidder M111, Fall Term 2015
Students in ST 201,

ST 511
HW 1
PRASANNA VENKATESH RAJARAMAN
932-663-230
18)https:/engrprn.engr.oregonstate.edu:9192/app;jsessionid=1pdf3n3zt9hj2?
service=direct/1/UserReleaseJobs/
$ReleaseStationJobs.release&sp=S206%7C6729d7d0d4a0d7ca07627ebcc802f1f7
A set of random numbers

Two independent samples
Null
hypothesis*
Assumptions
Randomization test
on (Y2 - Y1)
Two sample t-test
The treatment effect
is zero.
The difference in
population means is
zero.
The difference in
population medians
is zero.
OR
OR
OR
The treatment effect
is

Stat 411/511
ANOVA & REGRESSION
Nov 28 2012
Charlotte Wickham
Wednesday, November 28, 12
stat511.cwick.co.nz
Randomized experiment
Insulating Fluid Case Study
Breakdown times for electrical insulating fluid at various
voltages.
n = 76
I=7
Wednesday, Novem

Stat 411/511
ASSUMPTIONS OF REGRESSION
Nov 25 2012
Charlotte Wickham
Sunday, November 25, 12
stat511.cwick.co.nz
Remember these?
What are the assumptions of linear regression?
Your turn
Sunday, November 25, 12
The usual analysis procedure
1. Plot response

Stat 411/511
INFERENCE IN SIMPLE LINEAR
REGRESSION
Nov 21 2012
Charlotte Wickham
Wednesday, November 21, 12
stat511.cwick.co.nz
Hubway data
Check website for info on dates and
times.
Wednesday, November 21, 12
Three types of inference
Inference on the slo

Examples of two group comparison hypotheses and statistical summaries
Schizophrenia case study
The outcome variable is:
the left hippocampus volume (cm^3)
Group 1:
Twin with schizophrenia
Group 2:
Twin without schizophrenia
Is the study a control

Stat 411/511 covers statistical tools for dealing with one-, two-, and k-sample comparisons of
quantitative responses, and the modeling of a quantitative response as a function of a quantitative
explanatory variable (simple linear regression).
After takin

Hypotheses for two group comparisons
Identify the outcome, and the groups (or treatments), and whether the groups/treatments were randomly assigned.
The outcome variable is:
Group 1: !
!
!
!
Group: 2
Substitute in the identified outcome and groups where i

Stat 411/511
Learning more R
(optional)
In class, I will teach you all the R you need for ST411/511. However, this will barely scratch the surface
of the R language. If you are interested in learning more R I've included a few resources here.
I will try t