# case1801: Obesity and Heart Disease In Sleuth3: Data Sets from Ramsey and Schafer's "Statistical Sleuth (3rd Ed)"

## Description

To better understand whether the relationship between heart disease and obesity could be due to the social stigma associated with obesity, researchers examined cardiovascular disease rates of obese and non-obese women in American Samoa, where obesity was considered socially desirable. 3,112 women were categorized according to whether they were obese or not and whether they died from cardiovascular disease (CVD).

## Usage

 `1` ```case1801 ```

## Format

A data frame with 2 observations on the following 3 variables.

Obesity

a factor with levels `"NotObese"` and `"obese"`

Deaths

the number of women who died from CVD

NonDeaths

the number that died from other causes

## Source

Ramsey, F.L. and Schafer, D.W. (2013). The Statistical Sleuth: A Course in Methods of Data Analysis (3rd ed), Cenage Learning.

## References

Crews, D.E. (1988). Cardiovascular Mortality in American Samoa, Human Biology 60: 417–433.

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42``` ```str(case1801) attach(case1801) ## EXPLORATION myTable <- cbind(Deaths,NonDeaths) # Form a 2 by 2 table of counts row.names(myTable) <- Obesity # Assign the levels of Obesity as row names myTable # Show the table ## INFERENCE (4 methods for getting p-values and confidence intervals) prop.test(myTable, alternative="greater", correct=FALSE) # Compare 2 proportions prop.test(myTable, alternative="greater", correct=TRUE) # ...with cont. corect. prop.test(myTable,correct=TRUE) # 2-sided alternative (default) to get CI chisq.test(myTable) # Pearson's Chi-Squared Test fisher.test(myTable, alternative="greater") # Fisher's exact test fisher.test(myTable) # 2-sided alternative to get CI for odds ratio myGlm1 <- glm(myTable ~ Obesity, family=binomial) # Logistic reg (CH 21) summary(myGlm1) # Get p-value-- 0.734 beta <- myGlm1\$coef exp(beta[2]) #Odds of death are estimated to be 17% higher for obese women exp(confint(myGlm1,2)) # 95% confidence interval ## GRAPHICAL DISPLAY FOR PRESENTATION myTable # Deaths NonDeaths #Obese 16 2045 #NotObese 7 1044 prop.test(16,(16+2045)) #For one proportion, est: 0.0078 95% CI: 0.0046 to 0.013 prop.test(7,(7+1044)) #For one proportion, est: 0067 95% CI: 0.0029 to 0.014 pHat <- c(0.007763222, 0.006660324)*1000 # Get estimated deaths per 1,000 women lower95 <- c(0.00459943, 0.002921568)*1000 upper95 <- c(0.01287243, 0.014318321)*1000 if(require(Hmisc)) { # Use Hmisc library myObj <- Cbind(pHat,lower95,upper95) Dotplot(Obesity ~ myObj, # Draw a dot plot of estimates and CIs xlab="Estimated CVD Deaths Per 1,000 Women (and 95% Confidence Intervals)", ylab="Weight Category", ylim=c(.5,2.5), cex=2) } detach(case1801) ```

Sleuth3 documentation built on May 29, 2017, 11:28 a.m.