The Challenger Study That Broke the Theory of Perfect Memory

Space Shuttle Challenger breaking apart in the sky shortly after launch on January 28, 1986, the event studied in Ulric Neisser's flashbulb memory research

Psychologists once believed certain memories worked like a camera flash: a shocking event burns the moment permanently into memory, complete and unchanging, ready to be replayed decades later exactly as it happened. The Space Shuttle Challenger disaster gave one researcher an unplanned opportunity to test that claim against real data collected the morning after. What he found didn't just complicate the theory — it caught his own research subjects insisting, with total confidence, on memories that their own handwriting proved were wrong. Readers of CurioLink's Elevator Game coverage will recognize the underlying lesson: confidence in what someone remembers, or thinks they perceived, is not evidence that it happened that way.

The Theory Being Tested

In 1977, psychologists Roger Brown and James Kulik coined the term "flashbulb memory" to describe recollections of the circumstances in which someone learned of a surprising, emotionally significant event — where they were, what they were doing, who told them — rather than memory of the event itself. Brown and Kulik's original paper, built around recollections of the John F. Kennedy assassination, proposed that a special neural mechanism might preserve these memories with unusual permanence and accuracy — a "photographic" model of memory that treated the moment of hearing shocking news as functionally different from ordinary remembering.

What Cornell Psychologist Ulric Neisser Actually Did

On January 28, 1986, the Space Shuttle Challenger broke apart 73 seconds after launch, killing all seven crew members on live television. The morning after, Cornell psychologist Ulric Neisser and his collaborator Nicole Harsch gave a written questionnaire to 106 students in an introductory psychology course at Emory University, asking them to describe exactly how they had heard the news: where they were, what they were doing, who told them, and how they felt. Two and a half to three years later, the researchers tracked down 44 of the same students and asked them to fill out the identical questionnaire from memory, then rated each new answer against the original for consistency across seven scorable content questions.

The Numbers Behind the Distortion

Measure	Result
Mean consistency score (out of 7 possible points)	2.95
Participants scoring 0 out of 7	11 of 44
Participants scoring 2 or fewer	22 of 44
Participants recalling that they had filled out the original questionnaire at all	11 of 44
Average self-rated confidence in the follow-up memory's accuracy	4.17 out of a possible 5
Reported hearing the news via television, Day 1 vs. Year 2.5	21% vs. 45%

The television statistic is particularly telling: nearly half of participants came to remember first hearing the news from a TV broadcast, even though most of them hadn't actually seen it that way — a pattern researchers call the "wrong time slice" effect, in which repeated later viewings of replayed disaster footage get folded into the memory of the original, unwitnessed moment.

"That's My Handwriting, But That's Not What Happened"

Neisser and Harsch didn't just score written answers; they sat down with participants, showed them their own original, signed questionnaires, and asked them to reconcile the two accounts. Participants were, without exception, unable to explain the discrepancies — and several insisted their newer, incorrect memory was the accurate one even while looking directly at their own earlier handwriting. One participant's response to being shown the contradiction, repeated across many secondary accounts of the study, was blunt: "That's my handwriting, but that's not what happened." The confidence participants expressed in their later, inaccurate memories was not measurably related to how wrong those memories actually were.

The Debate the Study Started

Neisser and Harsch's findings didn't end the argument over flashbulb memory — they redirected it. A separate 1988 study of the same Challenger disaster by researcher Michael McCloskey and colleagues, using a different consistency-scoring method, reported 60% of participants giving consistent accounts across two time points — a sharply higher figure than Neisser and Harsch's 2.95-out-of-7 average, illustrating how much a flashbulb memory study's conclusions can depend on how "consistency" gets measured and scored. The dispute over method, not just result, has continued since. A related 1988 study by John Bohannon found that participants who had discussed the Challenger disaster more frequently with others in the years afterward reported more vivid, detailed memories of hearing the news — evidence that rehearsal through retelling, not any special one-time encoding process, may be what makes these memories feel so solid, whether or not they are accurate. In 2003, Duke University researchers Jennifer Talarico and David Rubin tested the theory again using the September 11 attacks, tracking 54 students' memories against an ordinary recent event over up to 32 weeks. Both flashbulb and everyday memories decayed at statistically indistinguishable rates — but only the flashbulb memories kept their high vividness and confidence ratings even as their accuracy fell. Talarico and Rubin's conclusion reframes rather than reverses Neisser's: flashbulb memories aren't unusually accurate, only unusually convincing.

What's Confirmed, and What Isn't

Documented in the original study data and its direct replications: that Challenger-memory accuracy declined sharply over roughly three years despite high reported confidence; that confidence and accuracy showed no reliable correlation; that a meaningful share of participants shifted their reported news source toward television over time; and that the general pattern — high confidence paired with declining accuracy in emotionally significant memories — has been independently replicated using the September 11 attacks.

Genuinely disputed among memory researchers: whether flashbulb memories involve any special encoding mechanism at all, or are simply ordinary memories preserved through unusually frequent rehearsal and retelling (the position Neisser and later researchers favor); and how much a study's specific consistency-scoring method shapes its reported results, given that McCloskey's 1988 team and Neisser's team reached different accuracy figures from broadly similar events.

FAQ

What is a flashbulb memory?
A term coined by Roger Brown and James Kulik in 1977 for a vivid, confident memory of the personal circumstances surrounding one's first hearing of a shocking or emotionally significant public event — not a memory of the event itself.

Did Neisser prove flashbulb memories are always wrong?
No. His study showed they can be substantially distorted despite high confidence, but some participants did recall accurately, and later research (Talarico & Rubin, 2003) found flashbulb memories decay at rates similar to ordinary memories rather than being uniquely unreliable.

Why did so many students misremember hearing the news from TV?
Researchers attribute this to the "wrong time slice" effect: repeated exposure to replayed news footage after the event gets misattributed to the moment of first hearing about it.