SOCR Simulated HELP Data Activity

SOCR Simulated HELP Data

See the SOCR Simulated HELP Data first. These data can be copy-pasted using the mouse from the HTML table into a plain text file "help_data.csv".

R examples

These simulated HELP data can be used to demonstrate (using SOCR and R)a number of different statistical, modeling, inferential and data analytic techniques.

Data I/O, summaries, visualization

options(digits=2)  # decimal precision
options(width=80)  # narrows output to stay in the grey box

hemp_sim_data <- read.csv("http://socr.umich.edu/data/SOCR_HELP_SIm_Data_2014.csv", na.strings=c("",".","NA"))
# note that we specify all of these values as indicating missing data ("",".","NA")
attach(hemp_sim_data)

rownames(hemp_sim_data)  # print row and column names
colnames(hemp_sim_data)

summary(hemp_sim_data)
fivenum(hemp_sim_data$\$ $mcs)

mean(hemp_sim_data$\$ $mcs, na.rm=TRUE); median(hemp_sim_data$\$ $mcs, na.rm=TRUE); range(hemp_sim_data$\$ $mcs, na.rm=TRUE); sd(hemp_sim_data$\$ $mcs, na.rm=TRUE); var(hemp_sim_data$\$ $mcs, na.rm=TRUE)

quantile(hemp_sim_data$\$ $mcs, seq(from=0, to=1, length=11), na.rm=TRUE)

no_mis_hemp_sim_data_mcs <- na.omit(hemp_sim_data$\$ $mcs)

hist(no_mis_hemp_sim_data_mcs, main="", freq=FALSE)
lines(density(no_mis_hemp_sim_data_mcs), main="MCS", lty=2, lwd=2)
xvals <- seq(from=min(no_mis_hemp_sim_data_mcs), to=max(no_mis_hemp_sim_data_mcs), length=100)
lines(xvals, dnorm(xvals, mean(no_mis_hemp_sim_data_mcs), sd(no_mis_hemp_sim_data_mcs)), lwd=2)

cor_mat <- cor(cbind(hemp_sim_data$\$ $mcs, hemp_sim_data$\$ $i11, hemp_sim_data$\$ $pcs1))
cor_mat
cor_mat[c(2, 3), 2]

plot(hemp_sim_data$\$ $mcs[hemp_sim_data$\$ $female==0], hemp_sim_data$\$ $cesd[hemp_sim_data$\$ $female==0], xlab="MCS", ylab="cesd", type="n", bty="n")

text(hemp_sim_data$\$ $mcs[hemp_sim_data$\$ $female==0& hemp_sim_data$\$ $substance=="alcohol"],
   hemp_sim_data$\$ $cesd[hemp_sim_data$\$ $female==1& hemp_sim_data$\$ $substance=="alcohol"],"A")

text(hemp_sim_data$\$ $mcs[hemp_sim_data$\$ $female==0& hemp_sim_data$\$ $substance=="cocaine"],
   hemp_sim_data$\$ $cesd[hemp_sim_data$\$ $female==0& hemp_sim_data$\$ $substance=="cocaine"],"C")

text(hemp_sim_data$\$ $mcs[hemp_sim_data$\$ $female==0& hemp_sim_data$\$ $substance=="heroin"],
   hemp_sim_data$\$ $cesd[hemp_sim_data$\$ $female==1& hemp_sim_data$\$ $substance=="heroin"],"H")

rug(jitter(hemp_sim_data$\$ $mcs[hemp_sim_data$\$ $female==0]), side=2)
rug(jitter(hemp_sim_data$\$ $mcs[hemp_sim_data$\$ $female==0]), side=3)


table(hemp_sim_data$\$ $homeless, hemp_sim_data$\$ $female)


OR <- (sum(hemp_sim_data$\$ $homeless==0 & hemp_sim_data$\$ $female==0 , na.rm=TRUE)*
       sum(hemp_sim_data$\$ $homeless==1 & hemp_sim_data$\$ $female==1 , na.rm=TRUE))/
      (sum(hemp_sim_data$\$ $homeless==0 & hemp_sim_data$\$ $female==1 , na.rm=TRUE)*
       sum(hemp_sim_data$\$ $homeless==1 & hemp_sim_data$\$ $female==0 , na.rm=TRUE))
OR


chisq_val <- chisq.test(hemp_sim_data$\$ $homeless, hemp_sim_data$\$ $female, correct=FALSE)
chisq_val


fisher.test(hemp_sim_data$\$ $homeless, hemp_sim_data$\$ $female)


ttres <- t.test(hemp_sim_data$\$ $age ~ hemp_sim_data$\$ $female, data=hemp_sim_data)
print(ttres)


wilcox.test(hemp_sim_data$\$ $age ~ as.factor(hemp_sim_data$\$ $female), correct=FALSE)

ksres <- ks.test(hemp_sim_data$\$ $age[hemp_sim_data$\$ $female==0], hemp_sim_data$\$ $age[hemp_sim_data$\$ $female==1], data=hemp_sim_data)
print(ksres)

Missing Values

sum(is.na(pcs1))  # count the missing values in the variable pcs1, 208
sum(!is.na(pcs1)) # count the non missing values in the variable pcs1, 246

sum(pcs1==49, na.rm=T) # Count the occurrence of 49 in pcs1, (omitting any missing values)
which(!complete.cases(pcs1)) # Find cases (row numbers) that are incomplete
# pcs1[pcs1==99] = NA You can re-map all 49 values in pcs1 as NA (missing) 
# pcs1 = pcs1[!is.na(pcs1)] # you can remove all NA (missing) data from pcs1

Sorting and subsetting

new_cesd = sum(hemp_sim_data$\$ $f1a-hemp_sim_data$\$ $f1t, na.rm=TRUE);
new_cesd

impute_mean_cesd = mean(hemp_sim_data$\$ $f1a - hemp_sim_data$\$ $f1t, na.rm=TRUE) * 20;
sort(hemp_sim_data$\$ $cesd)[1:4]
sum(is.na(hemp_sim_data$\$ $drinkstatus))
table(hemp_sim_data$\$ $drinkstat, exclude="NULL")

gender <- factor(hemp_sim_data$\$ $female, c(0,1), c("male","Female"))
table(hemp_sim_data$\$ $female)

Exploratory data analysis

newhemp_sim_data <- hemp_sim_data[hemp_sim_data$\$ $female==1,]
attach(newhemp_sim_data)
sub <- factor(substance, levels=c("heroin", "alcohol", "cocaine"))
plot(age, i1, ylim=c(0,40), type="n", cex.lab=1.4, cex.axis=1.4)
points(age[substance=="alcohol"], i1[substance=="alcohol"], pch="A")
lines(lowess(age[substance=="alcohol"], 
  i1[substance=="alcohol"], delta = 0.01), lty=1, lwd=2)
points(age[substance=="cocaine"], i1[substance=="cocaine"], pch="C")
lines(lowess(age[substance=="cocaine"], 
  i1[substance=="cocaine"], delta = 0.01), lty=2, lwd=2)
points(age[substance=="heroin"], i1[substance=="heroin"], pch="H")
lines(lowess(age[substance=="heroin"], 
  i1[substance=="heroin"], delta = 0.01), lty=3, lwd=2)
legend(44, 38, legend=c("alcohol", "cocaine", "heroin"), lty=1:3, 
  cex=1.4, lwd=2, pch=c("A", "C", "H"))

options(show.signif.stars=FALSE)
lm1 <- lm(i1 ~ sub * age)
lm2 <- lm(i1 ~ sub + age)
anova(lm2, lm1)

summary(lm1)

names(summary(lm1))
summary(lm1)$sigma

names(lm1)

lm1$\$ $coefficients
coef(lm1)
vcov(lm1)
pred <- fitted(lm1)
resid <- residuals(lm1)
quantile(resid)

Bivariate relationship

subst <- as.factor(substance)
genfem <- as.factor(ifelse(female, "F", "M"))
interaction.plot(subst, genfem, cesd, xlab="substance", las=1, lwd=2)

subs <- character(length(substance))
subs[substance=="alcohol"] <- "Alco"
subs[substance=="cocaine"] <- "Coca"
subs[substance=="heroin"] <- "Hero"
gend <- character(length(female))
library("lattice")
bwout <- bwplot(cesd ~ subs + genfem, notch=TRUE, varwidth=TRUE, col="gray")
bwout 
boxmeans <- tapply(cesd, list(subs, genfem), mean)

References

• Evaluation and Linkage to Primary (HELP) Care study
  • Data formats: help.csv and help.Rdata
  • Study and Data specifications
  • HELP study documentation
• SAS and R Data Management, Statistical Analysis, and Graphics, Kleinman / Horton, 2009

---

• SOCR Home page: http://www.SOCR.ucla.edu

---

---

Translate this page:

(default)	Deutsch	Español	Français	Italiano	Português	日本語	България	الامارات العربية المتحدة	Suomi	इस भाषा में	Norge
한국어	中文	繁体中文	Русский	Nederlands	Ελληνικά	Hrvatska	Česká republika	Danmark	Polska	România	Sverige