Drill 3

library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(readr)
library(ggplot2)
library(knitr)

NHDUH_Data <- read_csv("~/Downloads/Skills Drill 3 NSDUH Data.csv")

## Parsed with column specification:
## cols(
##   MDScore = col_double(),
##   sexident = col_double()
## )

head(NHDUH_Data)

## # A tibble: 6 x 2
##   MDScore sexident
##     <dbl>    <dbl>
## 1      NA       99
## 2      NA        1
## 3       4        1
## 4      NA       99
## 5      11        1
## 6       8        1

NHDUH<-NHDUH_Data%>%
  mutate(MDCategory = ifelse(MDScore<5,"Low Risk",
                      ifelse(MDScore>=5 & MDScore<=12,"Moderate Risk",
                      ifelse(MDScore>=13,"High Risk",NA))),
         sexident = ifelse(sexident==1,"Straight",
                    ifelse(sexident==2,"GayLesbian",
                    ifelse(sexident==3,"Bisexual",NA))))%>%
  select(MDScore,MDCategory,sexident)

table(NHDUH$MDCategory,NHDUH$sexident)%>%
  prop.table(2)%>%
  round(2)

##                
##                 Bisexual GayLesbian Straight
##   High Risk         0.22       0.15     0.06
##   Low Risk          0.33       0.52     0.64
##   Moderate Risk     0.45       0.34     0.30

chisq.test(NHDUH$MDCategory,NHDUH$sexident)

## 
##  Pearson's Chi-squared test
## 
## data:  NHDUH$MDCategory and NHDUH$sexident
## X-squared = 1057.7, df = 4, p-value < 2.2e-16

chisq.test(NHDUH$MDCategory,NHDUH$sexident)[7]

## $expected
##                 NHDUH$sexident
## NHDUH$MDCategory  Bisexual GayLesbian  Straight
##    High Risk      122.6741   63.07335  2817.253
##    Low Risk      1079.1074  554.82702 24782.066
##    Moderate Risk  529.2185  272.09964 12153.682

chisq.test(NHDUH$MDCategory,NHDUH$sexident)[6]

## $observed
##                 NHDUH$sexident
## NHDUH$MDCategory Bisexual GayLesbian Straight
##    High Risk          383        132     2488
##    Low Risk           573        459    25384
##    Moderate Risk      775        299    11881