Biostatistics

{Student Name}

{Date}

COH 602: SAS Data Analysis Assignment #1 – Heart Data Set

Using the SAS program to analyze the Heart Data set I chose to use the three variables of

systolic blood pressure, diastolic blood pressure and cause of death. The data that I was able to

collect from using the UNIVARIATE procedure for systolic blood pressure reported a mean

value of 136.91, median value of 132, a mode of 120, with a standard deviation of 23.74, and

interquartile range of 28. Systolic blood pressure has a skewness value of 1.49 and the frequency

distribution is positively skewed. What these numbers show from the 5,209 recorded sets of data

is the value of systolic blood pressure that showed up the most was 120. This should not come

as a surprise since the accepted “normal” blood pressure is 120/80, where the 120 is systolic and

80 is diastolic. The lowest systolic value was 82 and the highest value was 300. Some of the

extremely high systolic values observed within this data set positively skews the data and also

raises the average (mean) value from a “normal” 120 to 136.91.

The second numerical variable that I chose was diastolic blood pressure. The data

collected using the UNIVARIATE procedure in SAS reported a mean value of 85.36, median

value of 84, a mode of 80, with a standard deviation of 12.97, and interquartile range of 16.

Diastolic blood pressure has a skewness value of 0.88 and the frequency distribution is positively

skewed. The same kind of explanation as I presented in the previous paragraph is also noticed in

regards to diastolic blood pressure. The diastolic value that presented the most was 80, which

again does not come as a surprise since the accepted “normal” diastolic is 80. The lowest

diastolic value was 50 and the highest was 160. Again, some of the higher diastolic values

observed positively skews the data and raises the average (mean) value from “normal” 80 to

85.36.

The third variable that I used from the Heart Data set was cause of death. Since cause of

death is a categorical variable, I will interpret data that what collected using the FREQ procedure

in SAS. There were five listed causes of death: cancer, cerebrovascular disease, coronary heart

disease, other, and unknown. There was a total cumulative frequency of 1,991. What this means

is there are 1,991 data entries for the variable cause of death. Cancer had a frequency of 539,

accounting for 27.07% of deaths. Cerebrovascular disease had a frequency of 378, accounting

for 18.99% of deaths. Coronary heart disease had a frequency of 605, accounting for the largest

percentage of deaths in 30.39%. Other had a frequency of 357, accounting for 17.93% of deaths.

The final cause of death category was “unknown” which had a frequency of 112, accounting for

5.63% of deaths.

{Student Name}

{Date}

COH 602: SAS Data Analysis Assignment #2 – Heart Data Set

For the second data analysis assignment, I choose to analyze cholesterol and weight. The

mean for cholesterol was 227.417 the standard deviation was 44.935 the median was 223, the

interquartile range was 59, and the mode was 200. These values and the histogram chart are

positively skewed indicating the tail to be longer on the right side. This tells us this sample

population has higher levels of cholesterol than they do lower levels. Specifically the

interquartile range illustrates to us where the bulk of our values lie in cholesterol levels and the

standard deviation indicates how spread out the values are within these cholesterol levels. The

histograms provide a visual of how sex slightly altered these descriptive statistics. For example

cholesterol levels in men had slightly higher percentages between the 200 and 300 levels than the

women. This was shown as almost 25% of men were in the 245 range whereas only about 20%

of women were in this same range.

The second variable I analyzed was weight. The mean for weight was 153.08, the

standard deviation was 28.915, the median was 150, the interquartile range was 40, and the mode

was 138. These values and the histogram chart are also positively skewed as more people are

overweight than underweight. Furthermore, when skewness is a positive number such as this

case at .555 we can also tell in advance the histogram will illustrate a longer tail to the right.

Thus, the histogram provides a visual of this skewness and also depict differences in the average

weights between men and women. As to be expected, men had higher weight average than the

women did. Therefore, we could expect their other descriptive values to also be slightly

different. However, both histograms for men and women illustrated positive skewness with

longer tails to the right. This proves both genders contribute to the positive skewness of the

overall weight distribution noted earlier for the heart data sample. However, inferences would be

more accurate after descriptive statistics were pulled for each gender.

Both variables are important to analyze as they have been shown to contribute to chronic

diseases such as obesity, cardiovascular disease, and diabetes. If individuals in the sample are

truly representative of our total population then this heart data set illustrates why we have seen

many more cases of obesity, diabetes, and cardiovascular disease in the U.S.