Predicting Obesity from Trauma and Race

author: Michael A. Erickson date: 8 March 2014 autosize: true

Introduction

## Loading required package: lattice
## Loading required package: MASS
## Loading required namespace: car
## 
## Attaching package: 'memisc'
## 
## The following objects are masked from 'package:stats':
## 
##     contr.sum, contr.treatment, contrasts
## 
## The following object is masked from 'package:base':
## 
##     as.array

Read in the Data

## read.spss() is from the foreign library -- note the errors rto.raw <-
## read.spss('../TRAUMA_FEB_25_2014_final merge_Edman_WATSON.sav', to.data.frame=TRUE,
## use.value.labels=TRUE) rm(rto.raw)

## spss.system.file() is from the memisc library
hate <- spss.system.file("Hate_Study 2.sav")
hate.ds <- as.data.set(hate)
hate.df <- as.data.frame(hate.ds)
## description(rto.ds) # does not work with knitr because of unicode values

I don't feel like I'm looking at the right thing, but the p-value does not rely on the theoretical \( \chi^2 \) distribution.

hg <- droplevels(subset(hate.df, select = grep("Gender|Haterelati", names(hate.df), value = TRUE)))
(hgt <- with(hg, table(Haterelati1, GenderPers)))
##                     GenderPers
## Haterelati1          Male Female
##   Family                1     10
##   Friend                8     15
##   Ex                    2      8
##   Coworker/Student     12     31
##   Step Family Member    0      3
##   Authority Figure      1      4
##   Acquaintance          0      4
##   NoHate                3      3
##   Other                 0      1
round(prop.table(hgt, 2), 3)
##                     GenderPers
## Haterelati1           Male Female
##   Family             0.037  0.127
##   Friend             0.296  0.190
##   Ex                 0.074  0.101
##   Coworker/Student   0.444  0.392
##   Step Family Member 0.000  0.038
##   Authority Figure   0.037  0.051
##   Acquaintance       0.000  0.051
##   NoHate             0.111  0.038
##   Other              0.000  0.013
chisq.test(hgt, simulate.p.value = TRUE)
## 
##  Pearson's Chi-squared test with simulated p-value (based on 2000 replicates)
## 
## data:  hgt
## X-squared = 7.612, df = NA, p-value = 0.4858