He cited four core concepts that their approach considers central:
- Arithmetic comparisons ("% more than")
- Ratios (percentages, rates, probability, risk/chance)
- Comparisons of ratios ("more likely", "more prevalent")
- Standardizing (comparing apples to apples)
Questions that come up that students should be able to cope with include:
- What is the metric?
- Difference between "percentage change" and "percentage point difference"
- Students should recognize the difference between "the percent of men who are runners" and "the percent of men among runners" -- they should understand which is the pie and which is the slice.
- Students should recognize when a statement containing a statistic or percentage is a prediction and when it is a confirmation of something already known.
- Students whould understand ways in which confounding factors help to create misimpressions. They should learn to spot confounders.
The idea is always to focus on numbers in conext, asking "what are the vulnerabilities of a given statistic?" Sheild suggested that students "take CARE" -- that is -- methodically inspect for Confounding factors, ask how the nyumber was Assembled (what's the data, what's the calculation, how were cases selected), and should understand that they can guard against randomness with a large sample. Finally, students should be alert to measurement Error and/or Bias.
I also learned about something called Simpson's Paradox -- an arithmetic situation in which a comparison can favor one side consistently in "partial" explorations of data, but the other other side can have the best record "overall" -- e.g., two batters -- one has a higher batting average in each of five games, while the other has a higher batting average overall. It happens.
Schield says he thinks the two most important groups to touch with this stuff are people who will go on to touch others -- he focuses on journalists and teachers. Journalism students, and education students or teachers in training.
By September 15 I plan to: develop a module for COM314 that teaches critical use of numbers in news stories. Moreover, I will engineer it so my third year review sessions have this going on, and in that way I can unobstrusively demonstrate both the value of doing this, and a way to do it, to my colleagues in Humanities.
Related Links to check out:
