A narrow-spectrum analysis between
preventative vaccinations and overall quality of health.
Corinne Hester, Brendan Jackson, Alex McPhillips, Shannon Otto
Master of Public Health Program, St. Louis University, St. Louis MO 63103
Introduction
Results
Vacci
MODEL SELECTION
Week 10
Model Selection
Basic principles
2 and Adjusted 2
Stepwise regression
test
AIC, BIC, lasso
Reproducibility and replication
Basic Principles
There is no one best way to choose a model
Many developments, still ongoing
Nearly
ANOVA: Analysis of Variance
Travis Loux
Setting
ANOVA (analysis of variance) compares the means of more than
two groups
I
Can also be two groups (same as independent samples t test)
I
Common notation: K groups, indexed k = 1, 2, . . . , K
I
Example:
I
I
I
Review Test Submission: Quiz 3A
Content
Course
Test
Started
Submitted
Due Date
Status
Attempt Score
Time Elapsed
Instructions
FL2015 BST-5000-01G, 01U, 02G, 02U
Quiz 3A
9/11/15 4:37 PM
9/11/15 4:57 PM
9/18/15 11:59 PM
Completed
6 out of 6 points
20 minute
A narrow-spectrum analysis between
preventative vaccinations and overall quality of health.
Corinne Hester, Brendan Jackson, Alex McPhillips, Shannon Otto
Master of Public Health Program, St. Louis University, St. Louis MO 63103
Introduction
Results
Vacci
LOGISTIC
REGRESSION
Week 11
Logistic Regression
The logistic model
Binary explanatory variable
Log odds ratios
Categorical explanatory variable
Single continuous explanatory variable
Separation
The Logistic Model
Instead of predicting the likely va
Jackson, Brendan D
BST 5100-01
Homework 6
1.
CH 7:5
According to our table, we see that Spearmans rho is 0.844, which is close to 1, meaning the
correlation is high. With a 95% confidence interval, we also see a p-value = 0.000, which is less
than 0.05 an
REPEATED MEASURES
ANOVA
Week 6
Repeated Measures ANOVA
Context
Sums of squares
ANOVA test
Sphericity
Context
For one- and two-way ANOVAs, we have assumed the
groups were sampled independently
Alternatively, one large sample could be classified by fa
CORRELATION
Week 7
Correlation
Correlation
Hypothesis test for
Confidence interval for
Spearmans rank correlation
Kendalls
Partial correlation
Correlation
Review of BST 5000
Pearsons
=
1 1
Measures the strength of the linear relationship b
REGRESSION
DIAGNOSTICS
Week 9
Regression Diagnostics
Linearity
Homoskedastic errors
Normally distributed errors
Collinearity and multicollinearity
Influential points
Outliers
Regression Assumptions
When we run a regression, we are assuming a true
u
Jackson, Brendan
BST 5100
Homework #8
1a.)
The relationship between Day 1 Hygiene and second day hygiene is statistically
significant. This is displayed by rendering a P-value of 0.001. The R2 is .310. This explains that
31% of the variances in Day 2 Hygi
WELCOME TO
BST 5100
Professor:
Travis Loux
Teaching Assistant: Daphne Lew
Describing Bivariate Data
Introduction
Scatterplots
Correlation
Simple Linear Model
Simple Linear Regression
Residuals
2
Bivariate Data
Bivariate data arises when we measure
Jackson, Brendan
Homework #5
BST 5100-01
1. CH 14:2
The significance value of 0.043 is less than the critical value of .05, so we reject the null
and say that sphericity does not hold. The Greenhouse-Geisser epsilon value of 0.558 is less than
0.75. The T
Brendan Jackson
BST 5010-01
Homework Assignment 1
1. CH 4:11
2. CH 4:12
3. CH 7:1 (hours and essay only, not grade)
Brendan Jackson
BST 5010-01
Homework Assignment 1
Based on the scatter plot and the correlations data,
there is a strong coorelation but it
BST 5100-01
Homework 7
For all regressions, write out the full regression equation and comment on the statistical significance of
each predictor. Also give the regression standard error and 2.
1. Ch 8: 1 (Include a scatterplot with the regression line dra
A narrow-spectrum analysis between preventative vaccinations and overall quality of health.
Project Proposal
Corinne Hester, Brendan Jackson, Alex McPhillips, Shannon Otto
Vaccine coverage rates among adults in the United States have slowly declined over
LOGISTIC
REGRESSION
Week 12
Logistic Regression
Multiple predictor variables
Interactions
Assessing model fit
Variable selection
Model assumptions
Reading SPSS
Multiple Predictor Variables
Just like with linear regression, we can include more than
DISCOVERING STATISTICS USING SPSS
Chapter 10: Moderation, mediation and more
regression
Smart Alexs Solutions
Task 1
McNulty et al. (2008) found a relationship between a persons Attractiveness and how
much Support they give their p
Brendan Jackson
BST 5100-01
Homework #3
1. CH 11:1
The output shows that the first contrast compares reward against punishment and
indifference. The second compares punishment against indifferences. Using a 95% confidence
interval, it can be observed that
Jackson, Brendan
BST 5100
Homework #9
1. CH 8.5
Model Summary
Change Statistics
Std. Error
Model
R
R
Adjusted
of the
R Square
F
df
Square
R Square
Estimate
Change
Change
1
Sig. F
df2
Change
1
.231a
.053
.050
.31125
.053
18.644
2
663
.000
2
b
.070
.066
.30
BST 5100-01
Homework 8
For all regressions, write out the full regression equation and comment on the statistical significance of
each predictor. Also give the regression standard error and 2.
1. Using the GlastonburyFestivalRegression.sav data set, run a
Running head: A NARROW-SPECTRUM ANALYSIS OF VACCINATIONS AND HEALTH.
A narrow-spectrum analysis between preventative vaccinations and overall quality of health.
Corinne Hester, Brendan Jackson, Alex McPhillips, Shannon Otto
Saint Louis University
Running
Linear Regression in SPSS
Regression -> Linear
General Linear Model -> Univariate
MODEL BUILDING
Indicator variables
Interactions
Regression coefficients
ANOVA table
Residual standard error
2
Create yourself
Create yourself
Automatic
Full model only
Stati
MULTINOMIAL
REGRESSION
Week 13
Multinomial Logistic Regression
Multinomial distribution
Generalizing from binary logistic
Multinomial logistic regression
Ordinal logistic regression
Multinomial Distribution
In BST 500 (or 502) you learned about the b
Brendan Jackson
BST 5100-01
Assignment #2
1. CH 11:3
The One-way ANOVA yields a P-value of .000, which is statistically significant
utilizing .05 significance. Therefore, there is sufficient evidence to support that there is a
difference between the four
Brendan Jackson
Homework #4
BST 5100-01
1. 13.1
Utilizing a 95% confidence interval, the graph shows that the variable Fugazi is rated
very low by older people and very high for younger people. Barf Grooks shows an inverse while
Abba showed high likeabili
Jackson, Brendan
BST 5100
Homework #7
1. According to our model summary, R2 = 0.006, and our equation is = 49.218 + 0.46x. Our
regression standard error is 9.860, and with a 95% confidence interval, our p-value = 0.038
and therefore significant, with a si