lecture 10

Question 1

how to compare distributions using boxplots

Answer 1

plot several boxplots for comparison of groups = may examine
central tendency
spread/variability
skewness
outliers

Question 2

explain comparing distributions - ebola outbreak - data

Answer 2

ebola cases and deaths by country over 3 weeks
mean larger than median = skewed
Different countries = can view as diff pops or subpop in a larger pop - depends

Question 3

explain comparing distributions - ebola outbreak - boxplots

Answer 3

3 diff pops - 3 diff samples
median below mean in every case = skewed, upper tail, right skew
conveys overall variability but ignores time - it was over 10 weeks

Question 4

explain comparing distributions - ebola outbreak - notes about time

Answer 4

variation across time not shown in boxplot
since infectious disease = expect systematic changes over time
week 1 is not comparable with obs of week 10
also common to report per1000 individuals to account for differences in sizes of populations in each country
boxplot cannot show this aspect of change overtime = no appropriate

Question 5

what is a better to present ebola data - explain

Answer 5

plot of time series of counts
for this type of data = better representation
usually standardize by pop — even better if use per something

graph of death rate - can see increases or decreases
easier to use box plot for this because rates stay relatively constant over time