BS821
Class 1
1
Introduction to Categorical Data Analysis
This class will focus on the analysis of categorical data
Categorical data is defined as having a measurement scale that consists of categorie
BS821
Class 2
1
The Binomial Distribution
Up until now, youve dealt primarily with the normal distribution as a data
distribution.
When dealing with categorical data, this is hardly ever appropriate
T
Monte Carlo Approximations
Gheorghe Doros
2017-09-28
Gheorghe Doros
Monte Carlo Approximations
2017-09-28
1 / 51
Need
Often of interest is the calculation of
Z
Z
E [g()] =
g()dp(|data) =
g()p(|data)d
BS821
Class 5
Logistic Regression
The situation
Regression situation with a dichotomous dependent variable
Independent variables are measurement variables, or dummy variables
representing categorica
EB821
Class 4
1
Analyzing Stratified Categorical Data
We will often have categorical data that is stratified by one or more
variables that may affect the relationship between our main exposure and
out
Review of Probability
Probabilities as models for uncertainty
Gheorghe Doros
2017-09-12
Gheorghe Doros
Review of Probability
2017-09-12
1 / 56
Belief functions
Let F , G, and H be three statements abo
BS821
Class 3
1
MEASURES OF EFFECT SIZE IN 2x2 TABLES
Problems with notation: There are eight ways that you can construct a 2 x 2 table:
#1
Outcome
#2
Exposure
Yes
No
9
5
Yes
45
No 34
#4
Outcome
Outco
Why Bayes
Bayesian Approach and Probability
Gheorghe Doros
2017-09-05
Gheorghe Doros
Why Bayes
2017-09-05
1 / 27
What is Bayesian approach to statistics?
Most of the stats you have learnt is most like
Comparison of PROC FREQ and PROC LOGISTIC Approaches
to Confounding and Effect Modification
Step 1: Checking for effect modification
0
1
Group 1
0
1
35
115
6
73
data one;
input group row col wt;
datal
Single - parameter models
Binomial, Poisson, Normal and Exponential data
Gheorghe Doros
2017-09-19
Gheorghe Doros
Single - parameter models
2017-09-19
1 / 54
Addendum to last lecture
1
The posterior d
Normal Model
Gheorghe Doros
2017-10-03
Gheorghe Doros
Normal Model
2017-10-03
1 / 33
Why Normal Model?
0.8
The most utilized statistical models has at its core the normal distribution
The primary reas
BS 805
Class 4
COMBINING SAS DATA SETS
Adding observations to data sets - set statement
Adding variables to data sets - merge statement
Modifying information in a data set - update statement
Optio
BS805
Class 2
Arrays and Two Factor ANOVA
Arrays
o array name
o size of the array
o variables inside the array
do loops
o index of the loop
o bounds of the loop
Useful with repeated measurements on
BS 805
Class 9
Changing Levels in SAS Data Sets
Overview
1) Converting one record per subject to n records per subject - output statement
2) Converting n records per subject to one record per subject
BS 805
Class 5
Multiple Linear Regression
Linear Regression
Dependent Variable: measured (continuous) variable. Also called the outcome
variable.
Independent Variables: measured (or binary) variables
BS 805
Class 7
Regression Diagnostics and Goodness of Fit
Situation
We want to fit a multiple regression model to a data set but are unsure that
assumptions used in the model are satisfied by the data
BS 805
Class 1
Data Sets, Advanced Input
1.
Overview:
A.
Data Set structure
B.
Temporary SAS Data Sets:
Made in SAS data step work.dsname
used by SAS during ordinary SAS programs
not for storage bey
BS 805
Class 11
Macros
Purpose
1. Macros are used to repeat a series of SAS statements.
a. Useful if a series of statements is used often.
b. Example would be a series of checks to run on all variable
BS805
Class 8
COLLINEARITY DIAGNOSTICS, PIECEWISE LINEAR
MODELS
I.
Collinearity Diagnostics
When one is analyzing data using a multiple regression model, a problem can arise
such that an independent v
BS 805
Class 3
DATES AND FUNCTIONS
MULTIPLE COMPARISONS IN TWO FACTOR
ANOVA
Using Dates in SAS
o Informats (read in dates from raw data file)
o Formats (control how dates are presented in output)
o I
BS704 Assignment 4
Goals of the Assignment
Use R to generate and summarize descriptive statistics on a sample of patients participating
in a study to evaluate the impact of an interactive web based a
BS704 Assignment 8
Goals of the Assignment
Conduct and interpret analysis of variance in R, and
Conduct and interpret pairwise post-hoc tests in R.
Dataset for Analysis
This homework uses a masked d
BS704 Assignment 5
Goals of the Assignment
Compute confidence intervals for a mean and proportion by hand,
Compute confidence intervals for a mean and proportion using R,
Create new continuous vari
BS704 Assignment 10
Goals of the Assignment
Perform and interpret sample size calculations by hand and using R.
Description of the Study
A group of researchers are interested in estimating the mean a
BS704 Assignment 9
Goals of the Assignment
Conduct and interpret chi square goodness of fit tests by hand and using R, and
Conduct and interpret chi square tests of independence by hand and using R.
BS704 Assignment 12
Goals of the Assignment
Interpret slope and y-intercept of simple linear regression model with a categorical
predictor,
Interpret slopes of multiple linear regression models,
In
BS704 Homework 11
This assignment focuses on correlation and linear regression. R commands for computing the
correlation and running a linear regression model are described in Section 4.1 of the cours
BS704 Assignment 2
Goals of the Assignment
Distinguish variable types for analysis,
Create new variables using R,
Use R to generate descriptive statistics for different variable types, and
Summari
BS704 Assignment 6
Goals of the Assignment
Conduct and interpret tests of hypothesis to compare groups with respect to
continuous and dichotomous outcomes by hand and using R.
Dataset for Analysis
A