Students recalled the early presidents very well—George Washington was remembered with virtual perfection—and they recalled the currently-serving president and his immediate predecessors with great accuracy. Irrespective of the time of test, memory for presidents became fuzzier further into the past, and the shape of this forgetting curve did not differ across time. Thus, whereas everyone remembered Richard Nixon in 1974, 17 years later he was remembered only by roughly 60% of college students, and by 2009 only 25% remembered Nixon, Watergate notwithstanding
Notably, recall of presidents before the point of the performance advantage arising from historical recency (roughly the last 9 presidents) did not differ much between the groups, even though more than 35 years elapsed between the first and the last group being tested—it appears that once a historical fact is in the distant past, it no longer matters much exactly how distant that past is.
This invariance of historical memories across time is confirmed by another aspect of the data: Again regardless of the time of test, people remembered Abraham Lincoln and the two presidents who immediately succeeded him in office (Johnson and Grant) far better than any of the 7 or so before or after.
In another experiment, Roediger and DeSoto tested three groups of adults in 2014: The three groups were chosen on the basis of age, such that they would have been college students at the three times of test in the first study. So one group of “baby boomers” would have been in their late teens or early 20s in 1974; a group of “generation X” participants would have been of college age in 1991; and the “millennials” were a college age in 2009. This study yielded remarkably consistent results with the first one, and is shown in the bottom panel of the figure above. Again, irrespective of a person’s age, the first presidents were remembered well, as was Abraham Lincoln and his successors, and as were the most recent presidents. There were some small differences between groups concerning the most recent presidents: The older people were, the further back their historical memories extended, presumably reflecting their earlier personal experiences—there is something that makes historical events more vivid if one experiences them than if one learns about them from parents or teachers.
Roediger and DeSoto conclude by noting the “great stability in how the presidents are remembered across generations—a seemingly permanent form of collective memory.” The way in which those memories are forgotten appears to be equally stable: The regularity in their data was sufficiently strong for Roediger and DeSoto to predict when a president will have been forgotten. For example, they expect that President Truman will be forgotten by three-quarters of college students by 2040, 87 years after leaving office.
Astronomers can predict a solar eclipse to the millisecond a few centuries in advance: Predicting collective memories “only” a few decades hence may appear to pale in comparison—but given how capricious human behavior can be, such predictions are no mean feat. All we need to do now is to wait for another 26 years to see if Roediger and DeSoto got it right. My money is on them being correct within a small margin of error.