Special Coverage on the 2020 Invited Talk

Everyday Amnesia: High Confidence Misses in Recognition Memory (217)

Henry L. Roediger, III and Eylul Tekin (Washington University in St. Louis)

Summary by Laura Mickes, Digital Content Editor

This recap is part of a special series of session summaries from the Psychonomic Society's 61st Annual Meeting. To read the rest of the series, click here.

Henry L. Roediger , the 2009 Keynote Speaker , is one of this year's Clifford T. Morgan Distinguished Leadership Award Recipients. As a recipient of this prestigious award, he gave an invited talk at this year's Annual Meeting.

Roediger's talks always draw big crowds, and his trend-setting research inspires much follow-up research. Consider the Deese Roediger-McDermott paradigm. Roediger and Kathleen McDermott  (one of this year's mid-career award recipients) dusted off a procedure conceptualized by Deese in the late 1950s. Their paper  on the paradigm has a whopping 4556 citations. He also popularized work on collective memory, the testing effect, and with this talk (and associated article), he is likely going to generate a series of investigations on "everyday amnesia."

"Everyday amnesia" occurs in everyone and describes the phenomenon of forgetting experienced events. To test the notion of "everyday amnesia," Roediger shifted the focus away from the most commonly considered hits and false alarms and onto misses, specifically high confidence misses.

Typically, when we express high confidence, we're highly accurate. Though high confidence false alarms can happen. But do we ever express high confidence in not seeing an item that we actually saw? That's the question that Roediger and Eylul Tekin asked in 5 experiments to investigate "everyday amnesia."

Across three experiments, between 16% and 20% of the misses were made with high confidence. As shown in the screenshot below, people know how to use rating scales.

What else can explain it? They investigated two possible explanations:

To test this first possibility, in one experiment, participants read 100 words aloud and 100 silently. This manipulation ensures that each word read aloud was encoded. Of the words read aloud, there were 25% of high confidence misses, ruling out the first possibility. 

To test the second possibility, in another experiment, participants deeply encoded items by categorized each one. This manipulation ensured that deep encoding. Again, the high confidence miss rate is 20%, ruling out the second possibility. 

Roediger offered the Atkinson & Shiffrin multi-store memory model  (diagram pictured in screenshot below) as one model that can account for these findings. 

I can't wait to see the follow-up research this work inspires!