Your Name

Your e-mail address

1. Compare the "display of evidence" at the bottom of page 29 (bar graphs) to that on page 30-31 (map): both concern the mysterious deaths from cholera in 1854 London. Why is the map on page 30-31 better for doing something about this crisis than the charts on 29?
deceptive scaling fails to reveal causality on 29
chartjunk on 29 is "ludicrous and corrupt"
maps are always better than bar graphs
selection and ordering of data

2. Compare the charts on pages 41-42 (the acutal documents presented by engineers from Morton Thiokol to NASA administrators) with the chart on 45 that Tufte prepared from available data. Both charts have the same purpose: to convince NASA administrators that a cold weather launch of the shuttle could be catastrophic.

Except for one, all of the following are Tufte's explanations for why the charts on 41-42 ineffective and the chart on 45 persuasive. Choose the answer below that one that is not Tufte's.

the engineers who created the charts on pages 41-42 didn't understand the threat to the shuttle, and so they couldn't create convincing graphs
the amount of damage is broken up into categories on pages 41-42, rather than aggregated into a single "index" of relevant damage
on pages 41-42, multiple names and numbers are used for the same rockets, making comparisons difficult
on page 45, appropriate and relevant comparisons are clearly and consistently enforced: air temperature and amount of damage

3. On page 35, three maps, covering the same area of London, show the concentration of deaths of cholera in 1854. One of the maps clearly points to the correct source of the epidemic (the Broad Street pump), while the other two don't. (Similarly, on page 36, the only one of the three charts show dip in a company's revenues in 1982—the other two charts don't.)

How is it possible, according to Tufte, that maps of the same area with the same death figures (and charts of the same time period with the same revenue figures) can present such different visual pattern and apparent conclusions?

different variables shown in the various maps—selection of data defines the terms of the decision
same data is "aggregated" (grouped) differently in the various maps
differences of scale in the maps can reveal or conceal.
displaying causality vs. descriptive narration

4 & 5. In his conclusion on page 53, Tufte lists six principles "both for reasoning about statistical data and for the design of statistical graphics." For the two principles listed below, choose a bad example from Tufte's chapter which demonstrates that principle very poorly.

"insistently enforcing appropriate comparisons" (not!)

pg. 47: the "little rocket" chart as comparing O-ring damage to temperature
pg. 30-31: the map as comparing numbers of cholera deaths to proximity to the Broad Street pump.
pg. 44: the table of figures as comparing O-ring damage to temperature

"demonstrating mechanisms of cause and effect" (not!)

pg. 45: scatterplot as illustrating the effect of temperature on O-ring damage
pg. 33: the bar graph as demonstrating the effectiveness of removing the handle from the Broad Street pump.
pg. 50-51: the ice-water demonstration in the investigative hearings as illustrating the effect of cold on O-ring resilience.