Browsing by Author "Marsh, Elizabeth J"

In the face of declines in memory related to specific events, people maintain intact general knowledge into very old age. Older adults often use this knowledge to support their remembering. Semantic illusions involve situations in which presented information contradicts correct knowledge; the illusion occurs when people fail to notice a contradiction with what they know. Compared to younger adults, older adults' later memories are surprisingly less affected by semantic illusions. That is, they use fewer errors seen in the semantic illusions as answers when later asked related general knowledge questions. Why do older adults show this reduced suggestibility, and what role does their intact knowledge play? In 5 experiments, I explored these questions. Older adults' reduced suggestibility was not due to an age difference in error detection: older adults were no better than younger adults at detecting the errors that contradicted their stored knowledge. In addition, episodic memory failures were not a major factor either; the evidence for their direct involvement was mixed. Instead, prior knowledge seems to have been particularly protective for older adults. They demonstrated more knowledge to begin with but also gained access to even more of their stored knowledge across the duration of experiments, leading them to be less suggestible following semantic illusions. There was also an indication that when knowledge was stably accessible, older adults had a tendency to rely on it more than did younger adults. Broadly, these findings indicate that older adults' intact prior knowledge provides important benefits to their remembering and can even protect them against acquiring erroneous information about the world.

People expend a great deal of time and energy telling each other stories of events that are known to be invented. These fictional narratives—emerging from novels, films, television shows, radio dramas, and other media—can nevertheless leave an impact once a book’s cover is closed or the theater lights toggle on. This dissertation characterizes memories of fiction, a phenomenon both commonplace and understudied within empirical psychology. Not only is characterizing this behavior valuable in its own right, understanding how people remember and recruit memories of fiction also holds theoretical implications: any theory of memory which does not allow or account for how and why people recollect and use memories of events they know to be fiction is incomplete.

In Chapter 1, I knit together the theoretical precedent from prior work in autobiographical memory, mental models, and more, for considering memories of fiction as part of the “autobiographical record.” In subsequent chapters and across six studies, I examine the assumptions of this claim empirically. In Chapters 2 through 4, I characterize the subjective experience and function of memories of fiction by adapting established measures of autobiographical remembering across four studies, such as the Autobiographical Memory Questionnaire (AMQ), Centrality of Event (CES) scale, and Talking About Life Experiences (TALE) questionnaire (Berntsen & Rubin, 2006; Bluck et al., 2005; Rubin et al., 2003). I find that people readily ascribe phenomenological vivacity and functional significance to memories of fiction, and that these reports follow the same patterns as reports of memories of lived experience. On average, memories of fiction are less vivid and significant than personal memories, but not as a hard-and-fast-rule. Thus, these first four chapters provide evidence for claiming that the differences between memories of fiction and memories of lived experience are of degree, rather than kind. Chapter 5 (Studies 5 and 6) explore the extent to which memories from works of fiction are recruited to fulfill similar directive functions as autobiographical memories, especially in the absence of lived experience. Chapter 6 concludes by summarizing this body of work and a discussion of notable differences between memories of fiction and lived experience.

Psychological research on aging typically characterizes it as a period of decline. Numerous studies have reported age-related deficits in episodic memory, sensory perception, and fluid intelligence. These reports only add to society’s negative views of aging, which inevitably have a detrimental impact on older adults’ cognition, health, and general well-being. However, there are several other domains of cognition that remain stable or improve during healthy aging. For example, emotional functioning increases with age: older adults can better regulate their emotions and resist their desires compared to younger adults. Older adults are also more skilled at solving interpersonal problems and display intact implicit and procedural memory. This dissertation highlights two other areas that show improvement with age (i.e., decision making and knowledge) and considers how we can use these positive aspects to offset the negative aspects of aging. Chapter 2 investigates heuristic decision making. While some work suggests that older adults are more reliant on these shortcuts, there is little evidence to support this claim. To clarify this issue, participants from across the adult lifespan solved decision scenarios that tapped each of the following classic heuristics: anchoring, availability, recognition, representativeness, and sunk cost fallacy. Chapter 3 further explores knowledge. While the literature confirms that knowledge increases across the lifespan, it is unclear 1) if people are generally aware of this increase and 2) whether they hold expectations about the scope of younger vs. older adults' knowledge. To address these questions, younger and older participants predicted the knowledge of hypothetical younger and older adults. Chapter 4 focuses on application. While many studies have demonstrated that negative aging stereotypes negatively impact older adults’ memory performance, research on positive aging stereotypes’ influence is still inconclusive. In order to address this gap, older participants demonstrated their memory performance before and after viewing a neutral intervention or positive stereotype intervention about their knowledge advantage. Altogether, I find that older adults continue to use cognitively efficient decision strategies; they are not more reliant on classic heuristics and use them to the same degree as younger adults. Furthermore, I demonstrate that adults of all ages recognize that older individuals have a knowledge advantage over younger individuals, regardless of the difficulty of the information. Critically, if older adults are reminded of this advantage, they remember more words during a memory test. Taken together, this body of work sheds light on the cognitive improvements that accompany healthy aging and considers ways to leverage these positive aspects, with the goal of offsetting age-related deficits and promoting positive self-perceptions of aging.

Misleading claims surround us – we encounter them in news stories, advertising campaigns, and political propaganda. How do people separate facts from fiction? Decades of work implicate fluency, or subjective ease. Repeated statements feel easier to process, and thus more truthful, than new ones (i.e., illusory truth). This dissertation identifies additional cues for truth and takes a lifespan perspective. Older adults accumulate impressive amounts of knowledge (which may protect them), but also unduly attend to positive information (which may leave them vulnerable to emotional appeals). In two experiments, older adults exhibited illusory truth only when they lacked knowledge about claims, unlike young adults. Three additional experiments encouraged young adults to “stick with” what they knew. Evaluating truth at exposure prompted young adults to use their knowledge later, wiping out the illusion. Three final experiments disproved the idea that positivity “feels like” truth. Young adults exhibited a negativity bias, where negative faces made claims seem less true than neutral ones. Neither positive nor negative faces swayed older adults’ judgments. These results inform many theoretical perspectives – from fluency and referential theories of truth, to dual-process and socioemotional selectivity theories of aging. They also have important practical implications for preventing and correcting misconceptions in a “post-truth world,” where falsehoods travel farther and faster than the truth.

The focus on how one is behaving, feeling, and thinking, provides a powerful source of self-knowledge. How is this self-knowledge utilized in the dynamic reconstruction of autobiographical memories? How, in turn, might autobiographical memories support identity and the self-system? I address these questions through a critical review of the literature on autobiographical memory and the self-system, with a special focus on the self-concept, self-knowledge, and identity. I then outline the methods and results of a prospective longitudinal study examining the effects of an identity change on memory for events related to that identity. Participant-rated memory characteristics, computer-generated ratings of narrative content and structure, and neutral-observer ratings of coherence were examined for changes over time related to an identity-change, as well as for their ability to predict an identity-change. The conclusions from this study are threefold: (1) when the rated centrality of an event decreases, the reported instances of retrieval, as well as the phenomenology associated with retrieval and the number of words used to describe the memory, also decrease; (2) memory accuracy (here, estimating past behaviors) was not influenced by an identity change; and (3) remembering is not unidirectional – characteristics of identity-relevant memories and the life story predict and may help support persistence with an identity (here, an academic trajectory).

People are always at risk of making errors when they attempt to retrieve information from memory. An important question is how to create the optimal learning conditions so that, over time, the correct information is learned and the number of mistakes declines. Feedback is a powerful tool, both for reinforcing new learning and correcting memory errors. In 5 experiments, I sought to understand the best procedures for administering feedback during learning. First, I evaluated the popular recommendation that feedback is most effective when given immediately, and I showed that this recommendation does not always hold when correcting errors made with educational materials in the classroom. Second, I asked whether immediate feedback is more effective in a particular case—when correcting false memories, or strongly-held errors that may be difficult to notice even when the learner is confronted with the feedback message. Third, I examined whether varying levels of learner motivation might help to explain cross-experimental variability in feedback timing effects: Are unmotivated learners less likely to benefit from corrective feedback, especially when it is administered at a delay? Overall, the results revealed that there is no best “one-size-fits-all” recommendation for administering feedback; the optimal procedure depends on various characteristics of learners and their errors. As a package, the data are consistent with the spacing hypothesis of feedback timing, although this theoretical account does not successfully explain all of the data in the larger literature.

In the current digital age, people are increasingly relying on the internet as their primary source for looking up and learning new information. In 9 experiments, this dissertation seeks to understand how searching for information affects people’s metacognitive judgments and learning outcomes. First, we investigate how searching the internet for explanations impacts people’s confidence in their explanatory ability and the accuracy of their subsequent explanations. Second, we examine how looking up translations online affects people’s judgments of learning and their performance on a learning test. Third, we test how solving word searches influences people’s estimates of their knowledge of definitions. As people use cues to infer what they know, the effect of searching on metacognitive judgments depends on the cues that are available during searching. Specifically, searching inflates confidence in one’s knowledge when features in the search environment increase feelings of fluency (Ch. 2), but reduces confidence in one’s knowledge when searching produces feelings of disfluency (Ch. 3 & 4). Although searching involves active engagement, our results suggest that searching does not benefit learning (Ch. 2) and can even impair learning when it disrupts the encoding of to-be-learned information (Ch. 3). Overall, our findings suggest that it is not the act of searching itself but rather the cues that are available during searching that influence how people assess their own knowledge and how well they learn new information.

Despite the ubiquity of wrongdoing in everyday life, the vast majority of people believe that they are truly morally good. Across 15 studies that employ a combination of correlational, experimental, and mediation designs, this dissertation first examines the role of memory in constructing, protecting, and maintaining a morally good self-concept, and then it investigates whether the ways in which our moral transgressions are remembered and mutated can play a role in learning from past mistakes. In two initial studies, autobiographical memories of people’s especially morally good past actions were particularly central to constructions of personal identity. In three subsequent studies, a “knew-it-all-along effect” after acting dishonestly offered a way for people to explain away those past improprieties that could have presented a threat to a morally good self-concept. Then, across seven additional studies, past wrongdoings were attributed to a distant, dissimilar past self who had changed considerably over time for the better. The results of three final studies indicate that remembering, reflecting on, and mutating past events can serve a directive function by strengthening intentions for future moral improvement.

Simply recalling a memory has very different consequences than retelling it. Accuracy is often emphasized when events are recalled; however, retelling an event in a conversational manner may compromise accuracy in order to make the story more entertaining (e.g., Dudukovic, Marsh, & Tversky, 2004), or to support a specific argument (e.g., Tversky & Marsh, 2000). I will focus on the memorial consequences and underlying mechanisms of retellings.

First, I will review research that shows the inaccuracies of retellings. In reviewing this research, I will identify possible underlying mechanisms that change memory through retellings, such as the following: schema-guided reconstruction, interference, transfer inappropriate processing, and retrieval-induced-forgetting.

Second, I will experimentally investigate possible cognitive mechanisms underlying these memorial changes. In the first experiment, I show that the elaborative nature of storytelling does not influence memory more so than simply selectively rehearsing that information. In the second and third experiments, I investigate how retelling autobiographical events influences qualitative aspects of memory. These two experiments suggest that retrieving autobiographical events influences memory quality, although the specific nature in which they are retold has no effect. I close by connecting these three experiments to the broader literature.

While an impressive amount of knowledge is stored in memory, individual items can fluctuate in accessibility: Two attempts to retrieve the same knowledge (e.g., the US States) often yield somewhat inconsistent results. In this way, knowledge is unstable. The first goal of my dissertation was to document these fluctuations and to examine whether delay and testing impact this instability. Second, since knowledge is often described as being more stable relative to event memory, I directly compared fluctuations in access for the two memory types. The substantial fluctuations in knowledge documented in my first two experiments should have key implications for education, where students are regularly expected to draw upon prior knowledge. Accordingly, my third goal was to examine the power of multiple-choice testing as a tool for reactivating knowledge that has become inaccessible. In the laboratory, I compared multiple-choice testing to studying the target knowledge. My final goal was to investigate these issues in a real classroom; I evaluated the effectiveness of multiple-choice testing intervention in reactivating background course knowledge and promoting the acquisition of new material. Overall, my results highlight the instability of the knowledge base, with individual pieces of information coming in and out of reach.

The hypercorrection effect refers to the finding that high-confidence errors are more likely to be corrected after feedback than are low-confidence errors (Butterfield & Metcalfe, 2001). In 5 experiments I examine the hypercorrection effect, offer possible explanations for why the effect occurs, and examine the durability of the effect. In Experiment 1, I replicate the hypercorrection effect and show that delaying the feedback does not reduce the effect. In a secondary item analysis I also show that the effect is not caused by "tricky" questions. In Experiments 2 and 3, I show that subjects are more likely to remember the source of the feedback after both high-confidence errors and low-confidence correct responses. This suggests that it is the discrepancy between the subject's expectation and the actual feedback that causes the hypercorrection effect. In Experiment 4 I show that the hypercorrection effect also occurs for episodic false memories showing the diversity of the effect. Finally, in Experiment 5 I examine the durability of the effect. Initial high-confidence errors that are corrected after feedback remain corrected one week later.