Collective memory and autobiographical memory: Similar but not the same

Abstract

Sociologists and psychologists have independently identified the same general period in individual development for the formation of many memories. Yet the cross-disciplinary similarity is rarely recognized, because most psychologists study autobiographical memories of personal events, while sociologists focus on collective memories of national and world events. We examine autobiographical and collective memories together within a large national cross-section survey of Americans. Both types of memory are located primarily in the same broad period, identified here as ages 5–30 years. However, within that period, autobiographical personal memories as measured by cue-word associations typically refer to earlier ages, while collective memories assessed with a standard open-ended question typically refer to somewhat later ages. Other types of questions yield still other memory content and ages. Reconceptualizing the reminiscence bump or critical period reconciles these diverse results.

Keywords

Autobiographical memory collective memory critical period critical years reminiscence bump

For the past quarter century, psychologists studying autobiographical memory of personal events and sociologists studying collective memory of national and world events have identified much the same general period of human development as important for the formation of long-term memories: 10–30 years are ages often given by psychologists (e.g. Rubin et al., 1986: 213); adolescence and the 20s is the range usually reported by sociologists (e.g. Schuman and Scott, 1989: 359). In this article, we compare the two types of memory when, for the first time, they are both obtained from the same large probability sample—in this case, of Americans aged 18 years and older. We show that both forms of memory come from the broad age period characterized by psychologists as showing a “reminiscence bump” for personal events and by sociologists as indicating a “critical period” for memories of important national and world events. However, within that period, there are meaningful differences in modal ages that point to developmental variations valuable to document and interpret. Furthermore, after considering evidence from other investigators who studied memories for musical events or for sports figures, we connect distinct types of inquiries to different memory content and both to differences in the interests and preoccupations that are thought to characterize various developmental stages. In addition, within a developmental stage, the concept of “primacy” can help explain which specific event is recalled.

Background findings on the reminiscence bump and critical period

A major source of evidence for the distribution of autobiographical memories has been personal memories elicited by requesting associations to simple cue words. Rubin et al. (1998) illustrated such findings by presenting a figure based on four earlier cue-word studies showing ages from which the memories were said to come. When describing these results, they and other memory psychologists speak of a reminiscence bump, because disproportionate memories from the early years of life reverse the usual finding that recall is greatest soon after learning something new, followed by a consistent decline in remembering (Bower, 2000). Most individuals are especially likely to remember recent events, but the subsequent steady decline is interrupted by the reminiscence bump. Such a “bump” has been reported many times in a wide range of research and is now considered one of the most consistent results from studies of autobiographical memory (Glück and Bluck, 2007).

Evidence on collective memories comes from recollections by respondents when they are asked to name “especially important” national and world events from, say, the past 50 years (Schuman and Scott, 1989), and again there have been multiple replications of such findings, both for the United States and cross-nationally (Schuman and Corning, 2012). The results from collective memory questions are treated as one form of autobiographical memory by Rubin et al. (1998) in their widely cited article, which includes several charts from Schuman and Scott (1989) showing what Rubin et al. regard as a “reminiscence bump” for different national and world events—World War II, the assassination of President John F. Kennedy, and the Vietnam War. Schuman and Corning (2012) used the term “critical years” or “critical period” to refer to peak years for such early memories, rather than the term “reminiscence bump,” but for simplicity we will mostly use “bump” throughout this article to cover both terms.

The wide-ranging review by Rubin et al. (1998) of various forms of autobiographical memory saw the collective memory results as one of a number of sources of support for the general proposition that “things learned early … are remembered best,” an interpretation repeated in later writing by Rubin and his colleagues (e.g. Janssen et al., 2012). However, this evidence necessarily came from different investigators in different disciplines, using different questions and different data collected at different points in time. They did not, and could not, look closely at the exact basis in age for both autobiographical and collective memories. A subsequent lack of cross-disciplinary investigation of the relation of memories to age has been reinforced by the separation of journals by discipline, so that readers of research on one type of memory are often unaware of research on the other type. In addition, the different terminology—“reminiscence bump” versus “critical period”—has created a further barrier.

References to sociologists apply also to political scientists, who in writing about political socialization use still another term, “impressionable years” (Sears and Brown, 2013) as essentially synonymous with “critical years.” Neither “impressionable years” nor “critical period” is meant to reflect an instinctual process as in some animal studies, but instead they designate a period of increased exposure and openness to change that children experience as they enter adolescence and early adulthood.

We wish to address the issue of whether the similarity between the reminiscence bump and the critical period continues to appear when observed more closely and with the same individuals responding to questions calling for both personal and collective memories. Our initial goal is to determine how the different forms of memory are similar and how different in the developmental stage they tend to draw from, and we then proceed to consider what larger interpretation such a comparison stimulates. We will also show how the present terminology unintentionally leads to confusion over what may in fact be legitimately different findings.

Our main sources of evidence

In a national sample survey of US adults in June 2011, we included both psychologists’ inquiries about autobiographical events, using word cues to elicit personal memories, and sociologists’ inquiries about memories of important national and world events, using the question noted earlier and reproduced in full below—thus repeating the original Rubin et al. (1998) comparison, but doing so with the same individuals. In order to obtain adequate precision for the age distributions of both kinds of memories, we obtained a large probability sample, N = 2085, which has the additional advantage of including variations by age, education, gender, and other social background characteristics. Probability sampling of such a clearly defined population facilitates later replication in the same population, whether to assess reliability or to experiment with changes in wording—both of which we carried out, as we report at a later point in this article.

Posing questions about both collective and autobiographical memories to the same individuals might create an interaction with question order, and we therefore randomized the two sequences of questions and examined the possible effect of order. We found that question order did not change the basic patterns reported in this article for either type of memory. We also included on a random basis an unrelated question on respondent’s place of birth as a “buffer” between the two types of memory questions, but it had no significant effect (p > .10) either alone or in combination with question order and need not be dealt with further here. (See Appendix for our design of the research.)

Assessing autobiographical memories of personal events

In our survey, respondents were given eight simple words one at a time (flower, horse, fire, bird, lake, window, book, and friend) and asked to respond by indicating

a specific event from your own life that each word brings to mind. It may or may not be an event you think of as important, but it should have happened at a specific point in your life, whether recent, long ago, or at any other time.

After describing each memory, respondents indicated how old they recalled having been at the time of the personal event they remembered. Examples of answers to the cue words “fire,” “bird,” and “window” were “Roasting marshmallows over a camp fire,” “An owl I found that was sick,” and “Chipping ice from windows in winter,” respectively.

Figure 1 presents the associations with age that we obtained using each of the eight cue words in our present research. A reminiscence bump or critical period is apparent for each word, and since the distributions are quite similar, we averaged them and will refer mainly to that summary, shown in Figure 2, although the evidence in Figure 1 is valuable as a set of replications. Studies of the reminiscence bump in memories of personal events ordinarily limit their examination to respondents beyond the bump age itself, because for people in their teens and 20s, the period of the reminiscence bump coincides with the period of recency, so that the two effects are confounded. In both Figures 1 and 2, we follow Conway and Williams (2008) in restricting our sample to those aged 35 years and older, although, in fact, neither using the full sample nor limiting our focus to people over 65 years of age leads to changes in our figures that would affect our conclusions.¹ The bump in Figure 2 is in the range covered by ages 5–20, with the peak at ages 9–12. (Although the figures show ages 1–4, there are very few claims of recall from those ages, which fits the assumption that this is the period of infantile amnesia; such responses may be either erroneous in dating or otherwise very unusual.) The range is similar to that reported by Jansari and Parkin (1996), using a sample of students and employees at a British university, and also by Janssen et al. (2011), using a large Internet sample. Indeed, Rubin and Schulkind (1997) suggest that the bump may well go back to age 5—“that is, back to the childhood amnesia component” (p. 529). Our evidence supports such an extension to earlier than the 10–30 range often given for the reminiscence bump.²

Figure 1.

Autobiographical memories of adults aged 35+ years: Personal events recalled in response to individual word cues (percentage of memories recalled from each age).

Figure 2.

Autobiographical memories of adults aged 35+ years: Personal events (average percentage across eight word cues of memories recalled from each age).

Assessing collective memories

We adapted the question introduced by Schuman and Scott (1989) and used in a number of later studies (e.g. Jennings and Zhang, 2005; Schuman and Corning, 2012):

There have been a lot of national and world events and changes over the past 50 or so years—say, from about 1930 right up until today [1985 in the original research]. Would you mention one or two such events or changes that seem to you to have been especially important?” [IF ONLY ONE MENTION: “Is there any other national or world event or change over the past 50 years that you feel was especially important?”]

The total period of time given to respondents changed depending on the date of the survey, but 1930 was always the starting point. In 2011, the total period indicated was “80 or so years.”

Responses to the collective memory question can be used to create dichotomous variables showing mention of a particular event as especially important versus no mention of the event.³ Specific events given by Americans in 2011 as important included World War II (1941–1945), the assassination of President Kennedy (1963), the Vietnam War (1965–1973), the Fall of the Berlin Wall (1989), and the 9/11/2001 terrorist attack on the World Trade Center. Respondents’ ages at the time of each event named can be readily calculated, and when mentions of the event are distributed by age at the time of the event, the results have the shape of a reminiscence bump, as shown for three events in Figure 3, with peaks in the late teens and early 20s similar to those reported earlier by Schuman and Scott (1989) and in two instances reproduced by Rubin et al. (1998). (In Figures 3 and 4, data points are plotted only for base Ns of at least 30.) Despite the different dates of the three events themselves, the ages are largely within the often cited reminiscence bump or critical period of ages 10–30 (or 5–30 in our amended version). But it is important to note that collective memories peak in the later years of that range, between 17 and 24 for each event.

Figure 3.

Collective memories: Percentages mentioning three national and world events as important, by age at time of the event.

Figure 4.

Collective memories: Percentage mentioning World War II and the 9/11 attacks as important, by age at time of the event.

Although we show here only the data on national and world events from our 2011 experiment, which included cue-word associations for personal memories, a great deal of supporting evidence on the relation of mentions of each national and world event to age is available in Schuman and Corning (2012), to which we refer at points below. Readers can consult the figures in that earlier article to gain a clear sense of the relations, based on multiple replications. (Birth cohort is used rather than age in that earlier article in order to indicate location in time, but birth cohort and age are interchangeable: subtracting a birth cohort year from the actual date of an event gives the respondent’s age when the event occurred.)

It is necessary to keep in mind that the questions typically used to obtain autobiographical and collective memories are framed differently. The x-axis in Figures 1 and 2 shows each respondent’s remembered age at the time of the personal event he or she had reported in response to a cue word, and percentages are based on the total sample drawn for the survey. The distribution by ages at the time of each national and world event shown in Figure 3 reflects the proportions who mentioned each event within each age category. This difference precludes a straightforward analysis using cross-classification, and our approach therefore compares the ages of respondents at the time of the personal events they remember and their ages at the dates of the national and world events they recall as especially important.⁴

Very early and very recent events

Two frequently mentioned events plotted separately in Figure 4—World War II and the 9/11 terrorist attack—display age relations that are consistent with the reminiscence bump, but are not entirely adequate for the present purpose. In the case of World War II, which for Americans took place from the end of 1941 to late 1945, there were simply too few people (only 141) still alive in 2011 who had been in the reminiscence bump ages during that war. Schuman and Corning’s (2012; Figure 2) evidence for World War II showed that the earliest of eight large national samples gathered between 1985 and 2009–2010 yielded a curve that reflects a reminiscence bump for the expected ages (approximately 18–23; p. 11). However, it is not possible to rule out completely an interpretation based on a “period” or lifetime effect, since that would require having in the sample a sufficient number of still older individuals who had been past the bump closing age of 30 during the war to ascertain whether they mentioned the war; but few, if any, such individuals were still alive even at the time of the 1985 survey. (See Glenn, 2005 on the problems of separating cohort and period effects.)

The 9/11 attack presents the opposite problem: Figure 4 shows that, as expected, those in the bump years at the time do mention the attack most frequently, but for them, the attack was fairly recent, and we cannot be certain that the apparent reminiscence bump does not also reflect a recency effect. There is further support in Schuman and Corning (2012; Figure 3) for the high frequency of mentions occurring in the bump period, but future surveys are required to be certain that individuals still younger than age 5 or not even born in 2001 tend not to mention the 2001 attack, or do so only because of later learning (p. 12).

Because of the complications inherent in the World War II and 9/11 data, we prefer a conservative approach that focuses on events not so susceptible to alternative interpretations. For this purpose, the Kennedy assassination, the Vietnam War, and the Fall of the Berlin Wall are particularly valuable because each shows the predicted effect as part of a curvilinear distribution that precludes competing period or recency interpretations.

The reliability of results for personal and collective memories

Our most important finding is that the reminiscence bump for personal memories elicited by cue words tends to be located at younger ages (especially 5–16 or a little older) than the bump for memories of important national and world events (approximately 17–24), even though both kinds of memories fall within the broad age span of childhood, adolescence, and early adulthood that has often been called the reminiscence bump period—in our figures, approximately ages 5–30.

We can confirm the reliability of the curvilinear distributions of memories in two ways. First, using logistic regression with both a linear and a quadratic term (the latter a test for nonlinearity), collective memories show for the quadratic term an odds ratio = .93, p < .01 for JFK assassination; odds ratio = .91, p < .01 for Vietnam War, and odds ratio = .96, p < .04 for the Fall of the Berlin Wall.⁵ The difference between the personal memories (Figure 2) and the national and world event memories (Figure 3) is more difficult to test because of the skew of the distributions. (For example, the bump for personal memories in Figure 2 is mainly in the categories of 5–16 years, but the mean of the distribution reflects a much later age (29 years) because of the cumulation of cases toward older ages.) More generally, problems with traditional significance testing in social science have been increasingly recognized (see the entire issue of Perspectives on Psychological Science, November 2012), and we believe that independent replication provides greater confidence in the reliability of results such as those in Figures 2 and 3, which show the difference between each type of bump. As Bonett (2012) notes, “Replication evidence is the gold standard by which scientific claims are evaluated …” (p. 410).

With regard to the shape and location of the bump for national and world events, multiple replications are reported by Schuman and Corning (2012; Figures 4, 5 and 7 for JFK assassination, Vietnam War, and the Fall of the Berlin Wall, respectively). Each distribution is based on a large national sample (the sizes range from 744–3884), and some are reanalyses of data from national surveys gathered for purposes unrelated to the present article or authors (e.g. data gathered by the University of Chicago’s General Social Survey). In addition to the eight separate surveys from 1985 to 2009–2010, the figures in Schuman and Corning (2012) also show a weighted average for each birth cohort (the heavy black line) which summarizes the shape of the relation over all the separate surveys: approximately ages 13–22 for the peak points for JFK assassination; 18–22 for Vietnam War; 19–28 for the Fall of the Berlin Wall (with one small later peak possibly due to sampling error). Our new 2011 data points in Figure 3 are not exact duplications of these earlier findings (and we must allow for both sampling error and small shifts in memory over time), but the weighted averages in Schuman and Corning’s figures confirm the location of the reminiscence bumps shown in Figure 3.

Figure 5.

Autobiographical memories of adults aged 35+ years: Personal events when “important” is included and when omitted (average percentage of memories recalled from each age in response to “book” and “fire” word cues).

A further replication

One other separate replication seemed to us desirable with regard to the word-cued associations, all of which came from our single 2011 survey. We believe it is the entire orientation of the cue-word task presented to respondents that accounts for the difference between the age location of the word-cued associations and the collective memories, and not simply inclusion of the word “important” in the latter question and not the former. However, to test this hypothesis, in a second survey (N = 1406), in September 2011, we repeated two of the original word cues, but added the adjective “important” to the word-cue memory instructions, with the rest of the instructions essentially unchanged:⁶

On the next screens, you will see two different words. Please briefly describe a specific important event from your own life that each word brings to mind. It should have happened at a specific point in your life, whether recent, long ago, or at any other time.

As Figure 5 shows, the autobiographical word associations do not move closer in age to the collective memories when the word “important” is added as an adjective within the instruction to give personal word associations. The location of autobiographical memories is replicated almost perfectly despite this change; if anything, the replication points to an earlier age for the reminiscence bump than the original cue-word inquiry.

These various replications lead us to treat as reliable the conclusion that autobiographical and collective memories tend to be located at different ages: purely personal memories in the earlier years, mainly ages 5 to about 20 in terms of our categories, and national and world event memories in the later years, mainly ages 17–24. The figures do show some overlap, with some young children mentioning larger national and world events and some respondents locating personal memories in later years. There are no doubt individual differences in development for both kinds of memories and also situational differences in exposure to external events that can affect collective memories (e.g. quite young children being affected by a parent drafted into an ongoing war). But superimposing the personal events in Figure 2 on the national and world events in Figure 3 presents a clear difference in ages for each type of memory, whether using modes or other ways of judging the degree of separation.

We did not attempt in our design to discover explanatory correlates (such as emotional content, reaction time, etc.) of word-cued bump memories, but note that most such systematic attempts have not produced clear positive results (notably Janssen and Murre, 2008; Rubin and Schulkind, 1997). We also did not discover differences in the location of the bump by education, gender, or race, and, consistent with most, though not all, past reports on word-cued memories (e.g. Jansari and Parkin, 1996; Rubin and Schulkind, 1997), memories with both positive and negative content show similar curves in our data.

Evidence on age for other types of memories

Autobiographical personal memories and national and world memories are obviously different in content. It is useful for interpretive purposes to consider other efforts involving memory from early years, drawing on previously published research. Evidence from two more specialized memory inquiries points to other content differences, as well as what seems to be an intermediate age between that for cue-word associations and that for collective memories of national and world events.

Music

Schuman et al. (1997) asked respondents in a 1993 national survey to nominate “an event, a performer, or a style having to do with music over the past half century or so that you especially liked,” and followed with “How old were you when [that] made an impression on you?” The findings were clear-cut: the median age of reported impression, irrespective of the kind of music named, was 16, and although the age of impression increased slightly with the present age of the respondent, the more striking result was that almost all the median ages were in the teens. Smith’s (1994) study of musical tastes based on a national sample revealed similar results, and Janssen et al.’s (2007) broader study that included music along with books and movies arrived at similar critical years for music (see especially their Figure 1). The latter study used a large sample (N = 2161) recruited in diverse ways. Holbrook and Schindler (1989) also reported a curvilinear relation in their examination of musical tastes, although with a slightly older peak—which might result from their use of a convenience sample from specialized groups (church, parent–teachers meetings, and a rugby club); as the authors of the study acknowledge, it “would be dangerous to generalize … to the rest of the American … population” (p. 120). We take the mid-teens as the best estimate of recall of preferred music.

Sports memories

Janssen et al. (2012) asked Dutch respondents to give the names of the five football (soccer) players they thought were the best of all time. (Respondents were first shown a list of 190 players known to have been judged excellent, and they were also allowed to add names to the list.) The authors write,

The results were very strong. The mode of the age of the participants at the midpoint of the nominated players’ careers was 17. The mode did not change much when the different selection criteria were used. It remained at approximately the same value with players who were still active or with players who had retired, with the most popular or with less popular players, with Dutch or with foreign players, and with male or with female participants. (p. 174)

The modal age of 17 for remembered sports figures is very close to that for memories in the realm of music.

Both sports and music may be spheres where interests and preferences are usually established during the teenage years, and rehearsed frequently thereafter, and thus where such bump or critical period effects are readily found. Interests that are strong in these years may be linked to what is most apt to be recalled from that same period.

Life scripts

Although most of the early studies that gave rise to the term “reminiscence bump” used cue-word associations, over the past decade, research on the bump has increasingly focused on the concept of “life scripts” (e.g. Berntsen and Rubin, 2002, 2004; Dickson et al., 2011; Rubin and Berntsen, 2003; Thomsen et al., 2011). Life scripts are thought to be “culturally shared representations of the timing of major transitional life events” (Berntsen and Rubin, 2004: 427) that structure and help explain the bump, which in the life script research is seen as occurring mainly at ages 15–30 and to consist largely of positive but not negative events.

Furthermore, some studies that do not stress the specific concept of life script ask somewhat related direct questions. For example, Holmes and Conway (1999) instructed participants recruited near Bristol University “to list private events from their own lives which they considered to be important,” and found a reminiscence bump for ages 20–29 years that referred to such major areas of life as marriage, divorce, work, and education. It does not seem surprising that a focus on “importance” of events from such a limited sample led to mentions of life changes from a later period than cue-word associations.⁷

Integrating the evidence on the reminiscence bump

The term “reminiscence bump” is used widely, well beyond the studies that can be described here, and the age range identified for the bump extends from early childhood, as young as age 5, to at least the late 20s. We can make sense of such a range of usage by recognizing that different types of questioning about memory can legitimately produce different memory content and that the different content tends to be connected to different ages. This conclusion is consistent with some 75 years of split-sample questionnaire experiments showing that relatively small changes in question form, wording, and context can have important effects on answers (Cantril, 1944; Schuman and Presser, 1996; Sudman et al., 1996). Such experimental evidence on the question–answer process has been neglected in some of the psychological research on autobiographical memory.

In sum, almost all of the research on the reminiscence bump points to the broad age range of approximately 5–30 years as the span that interrupts the classic forgetting curve described by Bower (2000). However, different questions or procedures lead to different “bumps” within that span, with no one bump necessarily being more legitimate than another.

Memories within a particular domain tend to come from the point in the lifecourse during which events occur that are especially interesting and meaningful to many individuals. Simple early memories can be sparked by cue-word associations; memories for sports and music may be especially elicited when activities important to teenagers are asked about; and the late teens and 20s seem to be the source of many memories having to do, on the one hand, with serious personal relationships, work, and similar early adult concerns, and on the other hand, with larger national and world events encountered via the mass media. If there is no single reminiscence bump, there is no need to decide which is “the” bump.

Neither we nor previous writers have provided a still further theoretical explanation for which specific event is remembered within a particular reminiscence bump. In the case of collective memories of national and world events, it seems likely that the first event that most people would consider important has a particularly strong impact on those who are in adolescence or early adulthood. Thus a major war, an economic crash, or a singular event like the 9/11 attacks should be most easily remembered from this stage of life because there is nothing earlier meaningful for comparison. The event itself becomes a kind of baseline.

At the purely personal level of early childhood, with memories elicited by cue words, even a simple first experience, such as “roasting marshmallows over a camp fire,” can turn out to be memorable. The event may not be judged important later in life by the same individual (Rubin and Schulkind, 1997), but it may well have seemed important at the time of its first occurrence in childhood, leading to its memorable quality at the time, its rehearsal with friends and family (Rundus, 1971), and its long-term recall in the form of the top-of-the-head associations we found.⁸ The same subjective importance may occur for experiences connected with both sports and music during the teenage years. In still later years, it makes sense that what seem important to most individuals are unusual, often milestone experiences involving their education, work, and relationships.

Thus it is possible that there is no general explanation within a particular bump beyond the occurrence of the first event that seemed important to individuals at that stage of life and in that domain. Tulving (2008) offers a general proposal in the form of a “law of primacy,” writing that,

Most traditional explanations of primacy do not explain primacy as much as they explain primacy away. After they are finished there is no real phenomenon left to explain. (p. 38)

He maintains instead “that primacy is primary” and that “it reflects a basic property of the brain.” Thus, within a stage of life that makes most individuals open to a particular type of event, we can speculate that it is the first such event—whether a childhood experience, sports activity, form of music, life transition, or national or world event—that appears important and thus registers most strongly for later recall.

Finally, based on our main results, it seems best to call the 5–30 age span the “critical period” or “critical years” for early memories and to reserve the term “reminiscence bump” for particular peaks that have substantive significance within that larger time span.

Footnotes

Appendix 1 Acknowledgements

We are grateful to David B. Pillemer, David C. Rubin, and Marc R. Schuman for the suggestions each made on an earlier version of the present article.

Funding

This research was made possible by use of data collected by means of Time-sharing Experiments for the Social Sciences, NSF Grant 0818839, Penny Visser and Jeremy Freese, Principal Investigators.

Notes

Author biographies

Howard Schuman is a Professor of Sociology and Research Scientist Emeritus at the University of Michigan. He is the author of Methods and Meaning in Polls & Surveys (2008). Together with Amy Corning, he is writing a book on Generations and Collective Memory, and they published an earlier article in Memory Studies on Thomas Jefferson and Sally Hemings.

Amy Corning is a Research Investigator at the Institute for Social Research at the University of Michigan, USA. With Howard Schuman and Vladas Gaidys, she is the author of “Transformative Events and Generational Memory: A Case Study over Time in Lithuania,” in Sociological Forum 28: 373–394, and with Howard Schuman, “Commemoration Matters: The Anniversaries of 9/11 and Woodstock,” in Public Opinion Quarterly 77: 433–454.

References

Berntsen

Rubin

(2002) Emotionally charged autobiographical memories across the life span: the recall of happy, sad, traumatic, and involuntary memories. Psychology and Aging 17: 636–652.

Berntsen

Rubin

(2004) Cultural life scripts structure recall from autobiographical memory. Memory and Cognition 32: 427–442.

Bonett

(2012) Replication-extension studies. Current Directions in Psychological Science 21: 409–412.

Bower

(2000) A brief history of memory research. In: Tulving

Craik

FIM

(eds) The Oxford Handbook of Memory. Oxford: Oxford University Press, pp. 3–32.

Cantril

(1944) Gauging Public Opinion. Princeton, NJ: Princeton University Press.

Conway

(1990) Autobiographical Memory: An Introduction. Philadelphia, PA: Open University Press.

Conway

Williams

(2008) Autobiographical memory. In: Roediger

(ed.) Learning and Memory: A Comprehensive Reference. Volume 2: The Cognitive Psychology of Memory. New York: Elsevier, pp. 893–909.

Dickson

Pillemer

Bruehl

(2011) The reminiscence bump for salient personal memories: is a cultural life script required? Memory and Cognition 39: 977–991.

Easton

Dennis

(1969) Children in the Political System: Origins of Political Legitimacy. New York: McGraw-Hill.

10.

Glenn

(2005) Cohort Analysis Quantitative Applications in the Social Sciences. 2nd ed. Thousand Oaks, CA: SAGE.

11.

Glück

Bluck

(2007) Looking back across the life span: a life story account of the reminiscence bump. Memory and Cognition 35: 1928–1939.

12.

Groves

Peytcheva

(2008) The impact of nonresponse rates on nonresponse bias: a meta-analysis. Public Opinion Quarterly 72: 167–189.

13.

Holbrook

Schindler

(1989) Some exploratory findings on the development of musical tastes. Journal of Consumer Research 16: 119–124.

14.

Holmes

Conway

(1999) Generation identity and the reminiscence bump: memory for public and private events. Journal of Adult Development 6: 621–634.

15.

Jansari

Parkin

(1996) Things that go bump in your life: explaining the reminiscence bump in autobiographical memory. Psychology and Aging 11: 85–91.

16.

Janssen

Murre

JMJ

(2008) Reminiscence bump in autobiographical memory: unexplained by novelty, emotionality, valence, or importance of personal events. Quarterly Journal of Experimental Psychology 61: 1847–1860.

17.

Janssen

SMJ

Chessa

Murre

JMJ

(2005) The reminiscence bump in autobiographical memory: effects of age, gender, education, and culture. Memory 13: 658–668.

18.

Janssen

SMJ

Chessa

Murre

JMJ

(2007) Temporal distribution of favourite books, movies, and records: differential encoding and resampling. Memory 15: 755–767.

19.

Janssen

SMJ

Rubin

Conway

(2012) The reminiscence bump in the temporal distribution of the best football players of all time: Pelé, Cruijff or Maradona? Quarterly Journal of Experimental Psychology 65: 165–178.

20.

Janssen

SMJ

Rubin

St Jacques

(2011) The temporal distribution of autobiographical memory: changes in reliving and vividness over the life span do not explain the reminiscence bump. Memory and Cognition 39: 1–11.

21.

Jennings

Zhang

(2005) Generations, political status, and collective memories in the Chinese countryside. Journal of Politics 67: 1164–1189.

22.

Keeter

Kennedy

Dimock

. (2006) Gauging the impact of growing nonresponse on estimates from a national RDD telephone survey. Public Opinion Quarterly 70: 759–779.

23.

Rubin

Berntsen

(2003) Life scripts help to maintain autobiographical memories of highly positive, but not highly negative, events. Memory and Cognition 31: 1–14.

24.

Rubin

Schulkind

(1997) Distribution of important and word-cued autobiographical memories in 20-, 35-, and 70 year-old adults. Psychology and Aging 12: 524–535.

25.

Rubin

Rahhal

Poon

(1998) Things learned in early adulthood are remembered best. Memory and Cognition 26: 3–19.

26.

Rubin

Wetzler

Nebes

(1986) Autobiographical memory across the lifespan. In: Rubin

(ed.) Autobiographical Memory. Cambridge: Cambridge University Press, pp. 202–221.

27.

Rundus

(1971) Analysis of rehearsal processes in free recall. Journal of Experimental Psychology 89: 63–77.

28.

Schuman

Corning

(2012) The critical period and generational memory: evidence for national and world events. Public Opinion Quarterly 76: 1–31.

29.

Schuman

Presser

(1996) Questions and Answers in Attitude Surveys: Experiment on Question Form, Wording, and Context. Thousand Oaks, CA: SAGE.

30.

Schuman

Scott

(1989) Generations and collective memories. American Sociological Review 54: 351–381.

31.

Schuman

Belli

Bischoping

(1997) The generational basis of historical knowledge. In: Pennebaker

Paez

Rimé

(eds) Collective Memories of Political Events: Social Psychological Perspectives. Mahwah, NJ: Lawrence Erlbaum Associates, pp. 47–77.

32.

Sears

Brown

(2013) Childhood and adult political development. In: Huddy

Sears

Levy

(eds) Oxford Handbook of Political Psychology. 2nd ed. New York: Oxford University Press, pp. 60–145.

33.

Smith

(1994) Generational differences in musical preferences. Popular Music and Society 18: 43–59.

34.

Sudman

Bradburn

Schwarz

(1996) Thinking about Answers: The Application of Cognitive Processes to Survey Methodology. San Francisco, CA: Jossey-Bass.

35.

Thomsen

Pillemer

Ivcevic

(2011) Life story chapters, specific memories and the reminiscence bump. Memory 19: 267–279.

36.

Tulving

(2008) On the law of primacy. In: Gluck

Anderson

Kosslyn

(eds) Memory and Mind: A Festschrift for Gordon H. Bower. New York: Lawrence Erlbaum Associates, pp. 31–48.

37.

Yeager

Krosnick

Chang

. (2011) Comparing the accuracy of RDD telephone surveys and internet surveys conducted with probability and non-probability samples. Public Opinion Quarterly 75: 709–747.