Chapter 2: Section 3
The accompanying pie chart reports the number of hours of sleep "last night" for 128 college students. What would be a better type of graph for displaying these data? Explain why this pie chart is hard to interpret.
1 a. Histogram and Dot plot b. There are so many possible numerical values causing the pie chart to have too many "slices", which makes it difficult to tell which is which.
Refer to the accompanying graph, which shows the time spent on a typical day talking on the cell phone for some men and women. Each person was asked to choose the one of four intervals that best fitted the amount of time they spent on the phone (for example, "0 to 4 hours" or "12 or more hours"). Complete parts (a) through (d) below.
1a. One variable is gender which is a categorical variable since its describes qualities. The other variable is time range which is a categorical variable since it describes qualities b. The graph is a bar chart. It is the better choice since both variables are categorical. c. Two histograms should be used since the time would be a numerical variable d. The mode for women is 4-8 hours because it is the tallest bar for women. The mode for men is 0-4 hours because it is the tallest bar for men. Therefore, the distributions show that women talk more
Suppose that you have data which indicates that 90% of adults in a nearby town have cell phones. Of those who have cell phones, 30% use Carrier A, 30% use Carrier B, 10% use Carrier C, 20% use Carrier D, 5% use Carrier E, and 5% use other carriers. Would a bar graph or pie chart be better if the goal is to compare Carrier B and Carrier C? Explain.
A bar graph would be better since you are trying to compare two parts, not a part to the whole. The angles might be difficult to judge on a pie chart, making it hard to directly compare two sectors.
What is the most common trick to mislead readers of bar graphs?
Change the scale of the vertical axis so that it does not start at 0.
What is the first step in almost every investigation of data?
Make an appropriate graph.
This histogram shows the salary distribution for 20 employees at a small firm. Explain what is wrong with this histogram.
The bin width is too large to display the data in a useful way.
Suppose you construct a graph to compare the student populations of the five largest high schools in your city and choose to depict the populations with school buildings of various sizes. If the school buildings are drawn so that the length and the width are each in proportion to the population of the corresponding schools, is the resulting graph misleading? Why or why not?
The graph will be misleading since the student populations are one-dimensional data, but the graph uses a two-dimensional school building to represent it.
Suppose you have the following bar graph showing foreclosure rates for a few select states. If you want to make the point that the rates are all very similar, how would you change the graph?
To deemphasize the differences in foreclosure rates, change the scale on the y-axis to cover a larger range, 0-20 percent, for example.
