In the age of alternative facts and fake news, interpreting and challenging information is the responsibility of any informed news consumer. However, questioning the numbers and data of journalists, politicians, and advertisers may seem the purview of academics, unfeasible for those of us without a degree in statistics.
Brian Kernighan, professor of computer science at Princeton University, says many easy techniques — estimations based on simple multiplication — can allow the general public to determine the accuracy of every day numbers. Kernighan has written several books on math and computer science for a general audience. His latest book, “Millions, Billions, Zillions: Defending Yourself in a World of Too Many Numbers,” explores techniques and tricks that allow the non-math literate to question data and discern when numbers in articles, political speeches, and advertising may be misleading or just plain wrong.
A workshop with Kernighan titled “Too Many Numbers” and organized by the Princeton ACM/IEEE will explore tricks for double-checking the data of everyday life on Thursday, December 13, at 8 p.m. in Room C104 of the Princeton University computer science building at William and Olden streets.
According to Kernighan, the most common instances of misleading numbers fall into several categories. The first category is unintentional error, which often occurs when journalists are in a rush to publish material and are reporting on large, unruly numbers. “I don’t think any of us have an association with those numbers — the difference between a million and a billion is a factor of a thousand, but if you’re in a hurry, it’s easy to slip one of those numbers and make a mistake,” Kernighan says. Another frequent error is a simple confusion of units — substituting days for years, or feet for inches.
Kernighan also often sees journalists sacrifice accuracy for pretty graphics; what Kernighan calls “a bizarre triumph of form over function.” For example, in graphics using circles of different sizes to represent values, often the circle’s radius rather than its area corresponds to the value being represented. The result is that the ratio between numbers is squared, and thus, the difference between smaller and larger results seems far greater visually than is accurate.
There are, of course, many instances when numbers are intentionally manipulated to deceive viewers. As Kernighan observes, “there are news sources where the graph is clearly just wrong and intentionally wrong because it’s a politically motivated graph — the intent was to present stuff that isn’t correct at all.” In these cases data and graphics are deployed to make information look significant or unimportant.
Kernighan’s interest in misleading numbers began as a series of lectures for a class he teaches, “Computers in Our World,” in which he explores the fundamentals of computing for a non-scientist audience. For the class Kernighan explored how students might question numbers they were presented with in daily life and generate their own numbers — for example, how they might estimate the weight of the Princeton cannon or the number of leaves falling from an average tree in the autumn.
Kernighan grew up in Canada in the 1960s; his interest in computers grew alongside the technology for them. Kernighan, whose father was a chemical engineer, studied engineering physics at University of Toronto —there was no computer science major at the time — and studied electrical engineering at Princeton as a graduate student. Kernighan became fascinated by computers and began working over the summers at Bell Labs, the inventors of groundbreaking scientific developments ranging from radio astronomy to several programming languages. After graduation Kernighan went straight to Bell Labs and stayed there for the next 30 years. It was an exciting time: “We were able to do really interesting things and build new software that people would actually use as people began using computers more.” As part of Bell Labs, he contributed to the development of the Unix programming languages and coauthored programming languages AWK and AMPL.
Kernighan has also found his “ecological niche,” as he refers to it, in communicating complicated technical information to a non-technical public. Kernighan says he enjoys this work because, “I like to meet all kinds of people who I would not otherwise meet — and teach people how to make calculations relevant to their daily lives.”
In his talk, Kernigan will explore techniques that can allow the public to generate their own numbers and probe the accuracy of data. These techniques are of use not only to those interested in math, but to all who work with numbers in daily life. For projects that involve estimations — revenue, expenses, rent — it’s best not to have to take for granted the numbers that someone else generates.
“When you’re contracting a project and thinking of how long something should take and how much something should cost, you want to be able to generate your own numbers,” Kernighan says. “So when someone comes in and gives a number for building a parking lot, for example, you want to know whether the number is suspiciously big or small.”
These skills may resonate with the business community. “Thinking numerically, making accurate estimates for your own purposes, and being able to come up with pretty reasonable assessments of the numbers that other people are giving you is useful in a business setting,” he says.
Kernighan hopes that a broader public will see the relevance of numerical thinking to their daily lives. “The talk is for everyone; you don’t have to have a technical background to get something out of it.”