the Challenger charts: when the warning was there but the message did not land

The crew

On 28 January 1986, at 11:38 in the morning Eastern time, the space shuttle Challenger lifted off from the Kennedy Space Center in Florida. Seventy-three seconds later, a booster joint failed and it broke apart. On board were the commander, Francis R. Scobee; the pilot, Michael J. Smith; the mission specialists Ellison S. Onizuka, Judith A. Resnik and Ronald E. McNair; and the payload specialists Gregory B. Jarvis and Christa McAuliffe, chosen as NASA's first Teacher in Space.

The presidential commission that investigated the accident, chaired by William P. Rogers, found that the cause was a failure of the seals in a joint between two lower segments of the right solid rocket motor, which allowed hot gas to escape. It also found that the temperature at launch, 36 degrees Fahrenheit, was 15 degrees lower than on any previous launch.

The story of how that decision was made is a story about communication.

The night before

On the evening of 27 January, the weather forecast for the launch site was for temperatures in the low twenties Fahrenheit overnight. Engineers at Morton Thiokol, the company that built the solid rocket motors, were worried. The rubber O-rings that sealed the joints in the motor had shown damage on earlier flights, and they feared that in the cold the seals would be slower to do their job.

A teleconference began in the late afternoon and resumed at about 8.45 in the evening Eastern time, linking Thiokol's Wasatch plant in Utah, NASA's Marshall Space Flight Center in Alabama and the Kennedy Space Center in Florida. Thiokol sent its charts by fax. A NASA manager later testified that about 14 charts were presented. One of the engineers, Roger Boisjoly, presented charts on the joint's behaviour.

Thiokol's engineers made a clear recommendation: do not launch if the O-ring temperature is below 53 degrees Fahrenheit, the lowest temperature of any previous flight. Joe Kilminster, a Thiokol vice president, said he could not recommend launch.

NASA managers pushed back. Lawrence Mulloy, who managed the solid rocket booster project at Marshall, questioned whether the data justified a new temperature limit. George Hardy, a deputy director at Marshall, said he was appalled by the recommendation, though he also said he would not recommend launch over Thiokol's objection.

Thiokol then asked for a break and went off the line for about half an hour. During that caucus, according to the commission, Jerald Mason, a senior vice president at Thiokol, said a management decision was needed, and asked Robert Lund, Thiokol's vice president for engineering, to take off his engineering hat and put on his management hat. Boisjoly and his colleague Arnold Thompson objected strongly. In the end, four managers decided.

At about 11 in the evening, Thiokol came back on the line and recommended launch. Lund later testified that the usual position had been reversed: instead of having to prove it was safe to fly, the engineers had felt they had to prove the motor would fail.

The commission concluded that the decision-making process was seriously flawed, and that Thiokol's managers had reversed their position, against the views of their engineers, at the urging of NASA's Marshall center and to accommodate a major customer.

Tufte's charge

In 1997, the statistician and design writer Edward Tufte published Visual Explanations. One chapter, later reprinted as a booklet, compared the Challenger charts with a famous map of a London cholera outbreak in 1854. His argument was that the Thiokol charts failed to show the one relationship that mattered: how O-ring damage changed with temperature. He presented his own version, a simple scatter plot of damage against temperature, in which the danger of a launch far colder than any before is easy to see. In his view there are right and wrong ways to show data, and the wrong way had been used when it counted most.

The chapter became one of the most cited lessons in data presentation.

What is disputed

Tufte's reading has been strongly challenged, including by one of the engineers on that call.

In 2002, the philosopher Wade Robison, Roger Boisjoly himself, and the engineers David Hoeker and Stefan Young published a detailed reply in the journal Science and Engineering Ethics. As summarised by the visualisation researcher Robert Kosara, they argued that Tufte's chart mixed O-ring temperatures with air temperatures, which can differ a great deal; that the real failure mechanism was hot gas blowing past the seals, not simply O-ring damage; that Tufte left out an earlier Thiokol test in which the seals failed to work properly even at 75 degrees; and, above all, that the launch was not decided on the facts. NASA pressed Thiokol's managers, and they overruled their engineers.

Others have pushed back on Tufte's method too. Reviewing the book in Eye magazine in 1997, Paul Stiff asked whether Tufte's redesigned graphics were actually better for real readers, and how he knew, since he had not tested them. A University of Wisconsin course on data visualisation uses the chapter as an example of how not to write a critique, pointing students to the Robison paper.

Defenders of Tufte reply that his central point still stands: the evidence for a link between cold and seal trouble could have been shown far more clearly. Both things can be true. A clearer chart might have helped. It would not, on its own, have fixed a process in which the people with the evidence had to prove the danger beyond doubt.

The lesson for the rest of us

Most of us will never face a decision with lives so directly at stake. But nurses raise concerns on ward rounds, engineers flag risks to clients, and junior analysts tell senior leaders that the numbers do not support the plan. In each case, someone has to make a warning land.

Say the conclusion first. Do not make your listeners assemble it from fourteen charts. Put the recommendation in the first sentence.

"My recommendation is that we do not go ahead tomorrow. Here is the one reason."

Make the one chart that answers the question. If the question is whether something gets worse as conditions change, plot exactly that relationship, and nothing else, on the first slide. Everything else is backup.

"This single chart shows every past case. Look at where tomorrow sits."

Name the burden of proof. Lund's testimony is the quietest and most important part of this story. If you are being asked to prove something will fail, say so plainly, and ask for the question to be the right way round.

"I can't prove it will fail. Can anyone here show it's safe at this temperature?"

And if you lead a team, remember the other half of the lesson. A perfect chart cannot help a person who is afraid to show it, or a room that has already decided.

Sources

1. Presidential Commission on the Space Shuttle Challenger Accident (Rogers Commission), Report, Volume 1, Chapter 2, Events Leading Up to the Challenger Mission (crew, launch time, 73 seconds), NASA, 1986. www.nasa.gov/history/rogersrep/v1ch2.htm
2. Rogers Commission, Report, Volume 1, Chapter 4, The Cause of the Accident (joint failure, 36 degree launch temperature), NASA, 1986. www.nasa.gov/history/rogersrep/v1ch4.htm
3. Rogers Commission, Report, Volume 1, Chapter 5, The Contributing Cause of the Accident (the 27 January teleconference and findings), NASA, 1986. www.nasa.gov/history/rogersrep/v1ch5.htm
4. NASA, Remembering Challenger and Her Crew. www.nasa.gov/image-article/remembering-challenger-her-crew
5. Edward R. Tufte, Visual and Statistical Thinking: Displays of Evidence for Making Decisions (reprint of chapter 2 of Visual Explanations, 1997), Graphics Press. www.edwardtufte.com/book/visual-and-statistical-thinking-displays-of-evidence-for-making-decisions/
6. Paul Stiff, Strategies for telling truths (review of Visual Explanations), Eye, no. 25, 1997. eyemagazine.com/review/article/strategies-for-telling-truths
7. Robert Kosara, Tufte and the Truth About the Challenger, eagereyes, 2 February 2011 (summary of Wade Robison, Roger Boisjoly, David Hoeker and Stefan Young, Representation and Misrepresentation: Tufte and the Morton Thiokol Engineers on the Challenger, Science and Engineering Ethics, 8(1), 59 to 81, 2002). eagereyes.org/criticism/tufte-and-the-truth-about-the-challenger
8. University of Wisconsin, Madison, CS 765 Data Visualization, Readings 02: Why Visualize?. pages.graphics.cs.wisc.edu/765-22/all-readings/readings02

Published by The Full Stop, the journal of Brightside English, in the Landmark Moments series. Every factual claim links to a source below; where accounts or findings are disputed, we say so. Sources checked on 26 Oct 2026. Reviewed by a Brightside senior coach. Corrections: see our editorial standard.