Normal
Error Model
No matter what distribution the error terms i have (and hence the Yi),
the least squares method provides unbiased point estimates of 0 and 1.
These estimates have minimum variance among all unbiased linear
estimators.
However, to do con

Regression Models
A regression model is a formal means of expressing two essential ingredients of a statistical
relationship:
1)
A tendency of the response variable Y to vary with the predictor variable X in a systematic
fashion.
2)
A scattering of points

Estimation of the
Survival Function
We now wish to estimate the survival function. Note, unless stated
otherwise, we assume recorded values of time are continuous and
subject only to right censoring.
Assume we have a sample of n independent observations o

SDS 380D Statstcal Methods II
UTC 4.112 TTH 12:30-2:00 Spring 2016
Professor:
Matt Hersh
matt.hersh@austn.utexas.edu
Office: GDC 7.508F
Office Hours: Th 2:00-3:00
Teaching Assistant:
Teaching Assistant:
Kejin Lee
Kejin53@gmail.com
Office: PAR 210
O hours:

Mixed Models
So far, we have been using a cell means model for the single-factor random
effects model. We can also use a random factor effects model. The two
models are equivalent.
To do so, we express each factor level mean i as a deviation from its
expe

Generalized Linear Models
Generalized linear models, GLMs, describe patterns of association and
interaction.
The models help us evaluate which explanatory variables affect the
response, while controlling for effects of possible confounding variables.
For

Logistic Regression Models
The odds of success is the probability of success divided by the
probability of failure:
1
And
log
1
x
exp( x ) exp( ) exp( x )
1
Logistic Regression Models
An odds ratio is the ratio of two odds. Here we look at the o

Mixed Models
Typical ANOVA models have fixed factor effects. Consider studying the
effects that gender and drug (Tylenol, Advil, control) have on level of
headache.
These are fixed factor levels. Our interest centers on the specific factor
levels chosen.

Randomized Complete
Block Designs
In a randomized complete block design, the experimental units are first
sorted into homogenous groups, called blocks, and all treatment
combinations are assigned at random to experimental units within the
blocks.
This req

Introduction
Survival analysis is time to event analysis. Time to event is the time it
takes until a pre-defined event occurs. The times till events are called
survival times.
The term comes from biomedical sciences were there is an interest in
observing