Running Head: OLDER ADULTS AND MORALLY-CHARGED INFORMATION
Are Older Adults More Attuned to Morally-Charged Information?
Darcia Narvaez Gabriel A. Radvansky
Nicholas A. Lynchard
University of Notre Dame
David E. Copeland
University of Nevada, Las Vegas
Whereas older adults typically show declines in various cognitive processes they also typically demonstrate greater interest in social relationships. Part of this increased focus on interpersonal relations may extend to morality, which by its very nature is concerned with social contracts, obligations and the give-and-take among people. We tested whether in comparison to younger adults older adults show increased activation and memory for morally-charged information relative to nonmoral information. Three experiments examined older and younger adult comprehension and memory of moral content in stories. Participants read stories and were tested for surface form, textbase, and situation model recognition memory. In contrast to past studies that have not focused on moral content, older adults had textbase memory for moral information equal to that of young adults, suggesting an enhanced attention to morally-charged details. To examine online moral inference making, Experiment 2 used lexical decision probes. There was greater facilitation of moral inferences for older adults relative to younger adults, suggesting greater focus of processing on moral content. Experiment 3 explored methodological issues to resolve some discrepancies between the experiments, and replicated the basic findings. In general, older adults had enhanced memory for morally-charged story events and, relative to younger adults, were more likely to draw moral inferences during comprehension.
Are Older Adults More Attuned to Morally-Charged Information?
It is well established that cognitive aging involves at least two competing components. On the one hand, some cognitive abilities decline, such as speed of processing (Myerson, Hale, Wagstaff, Poon, & Smith, 1990; Salthouse, 1996), processing capacity (Craik & Byrd, 1982), inhibitory processes (Dempster, 1992; Hasher & Zacks, 1988), the ability to self-initiate processes (Craik & Jennings, 1992), and frontal lobe functioning more generally (Albert & Kaplan, 1980). On the other hand, some abilities remain stable or increase in effectiveness with age, including crystallized intelligence (Baltes, Staudinger & Lindenberger, 1999), semantic priming (e.g., Howard, McAndrews & Lasaga, 1981), and situation model use (Radvansky & Curiel, 1998; Radvansky & Dijkstra, 2007). For example, older adults tend to focus more on the gist meaning of events rather than on the details (Pratt & Norris, 1999). Although older adults are handicapped by declines in processing (e.g., Myerson, et al., 1990; Pratt, Boyes, Robins, & Manchester, 1989; Salthouse, 1996), they are also more likely to focus on the deeper meaning of stories, recalling them more succinctly in accordance with a focus on meaning and integration. In other words, the deeper symbolic and psychological meaning of narratives becomes more salient with age (e.g., Carstensen & Turk-Charles, 1994; Adams et al., 1997).
In addition to these cognitive changes, there is also a shift in interest such that, relative to younger adults, older adults become more focused on positive emotional experiences. Because emotional experiences are often socially embedded, this shift contributes to a greater emphasis on emotion regulation in social contexts (e.g., Carstensen, Isaacowitz, & Charles, 1999). That is, as one ages, positive relational behavior in terms of expressivity and negotiation is increasingly favored (e.g., Carstensen et al., 2003). For example, relative to younger adults, older adults report fewer and less intense negative emotions in relationships (e.g., Diener, Sandvik, & Larsen, 1985; Lawton et al., 1992). According to Socioemotional Selectivity Theory (SST) these findings make sense because older adults’ increasing awareness of lifespan limitations causes a motivational shift that directs their attention towards emotionally meaningful goals, especially in social situations (Carstensen, 1992; Suddendorf & Corballis, 1997). Such a focus facilitates the improvements seen in older adults regarding emotion regulation (Carstensen, Fung, & Charles, 2003) and improved memory for details of their positive emotional experiences (Labouvie-Vief, Hakim-Larson, et al., 1989).
These tendencies among older adults led us to postulate that the processing of morally-charged events, which are often both emotionally meaningful and socially embedded, would be especially salient to older adults. Morality is defined here as cooperation, sharing the benefits and burdens of the necessity to live in social groups (Piaget, 1932/1965; Rest, 1986). When elements of a situation point to the need for cooperative response, or to social responsibility, we call that morally-charged information. Of particular concern here is the degree to which older adults process and remember morally-charged content in stories. On the one hand, because there are general cognitive declines and decreased abilities for information processing, one might expect that older adults would be less effective at processing morally-charged content. On the other hand, because moral situations focus more on emotional aspects of interpersonal relations among people, we expected older adults to show preserved or superior processing of morally-charged information. This idea is supported by evidence showing that moral judgments can use some of the same neurological structures and emotion processing (Greene & Haidt, 2002; Greene, Sommerville, Nystrom, Darley, & Cohen, 2001), particularly when the moral judgments are more personal than impersonal.
To date, no one has tested whether moral information is particularly salient for older adults. There has been some work on the probability of drawing morals from fables presented to younger and older adults, but not on the on-line activation of moral information, nor the ability to retain morally charged information later (Arbuckle & Harsany, 1985). The current project explored the influence of aging on the processing of morally-charged content and the drawing of inferences related to that content.
Two areas of research are brought together here: (1) age-related changes in event comprehension and memory (e.g., Radvansky, 1999; Radvansky & Dijkstra, 2007) and (2) moral information processing, in particular the examination of individual and developmental differences in the processing of moral discourse (Narvaez, 2002). Specifically, we studied the processing of narratives that had moral events and themes. As in normal text processing, the processing of moral texts also requires drawing inferences and constructing a situation model (Narvaez, 1998).
Reading Comprehension and Aging
Three levels of mental representation have been identified in text comprehension: the surface form, the propositional textbase, and the situation model (van Dijk & Kintsch, 1983). The surface form refers to the verbatim words and syntactic structures that were originally encountered. The propositional textbase level captures the underlying ideas conveyed by the text apart from their specific wording. Both the surface form and textbase representations are mental representations of the text itself. In contrast to this, the situation model captures the gist of the events in the text and is functionally isomorphic to the events to which the text refers. As such, it serves as a mental simulation of what the text is about. This involves integrating text elements with inferences drawn from prior knowledge (van den Broek, 1994).
In general, research on cognitive aging has shown that older adults do much more poorly than younger adults on tests that examine surface form and textbase memory. However, processing at the situation model level is relatively well-preserved (see Radvansky & Dijkstra, 2007, for a review). An aim of the current research is to assess the degree to which this pattern of deficit and preservation extends to the processing of morally-charged content.
In addition to these memory representations, readers may also draw meta-level inferences, such as evaluations of the characters and the actions that occur (e.g., “That was a stupid thing to do”). Although these evaluations may not be directly incorporated into a mental representation of the text, such thoughts often guide processing. For example, skilled readers generate more explanatory inferences while thinking aloud during reading than less skilled readers (e.g., Chi, de Leeuw, Chiu, & La Vancher, 1994; Graesser, Singer & Trabasso, 1994; Trabasso & Magliano, 1996; van den Broek & Lorch, 1993; Zwaan & Brown, 1996); moreover, readers with expert background knowledge do more explaining and evaluation of events in the text than non-experts (e.g., Chiesi, Spilich, & Voss, 1979; Lundeborg, 1987; Wineberg, 1991;Wyatt, Pressley, El-Dinary, Stein, Evans & Brown, 1993). Readers also draw meta-level inferences when reading morally charged content (Narvaez, Lapsley, Hagele & Lasky, 2006).
Relative to younger adults, text processing studies show that older adults tend to focus on, recall, and prefer general information over details, (Labouvie-Vief, 1982; Spillich, 1983). Some argue that this synthesizing style may be a compensatory mechanism for processing limitations (Labouvie-Vief, 1982). Others suggest that the aging mind is evolutionarily conditioned to interpret new instances as familiar and to rely on existing schemas rather than form new ones (Lachman, Lachman, & Taylor, 1983). This tendency is particularly apparent in the changing saliency of social events during aging. Gould (2004) tested recall and elaboration of younger and older adults within discussion groups of different sizes. Along with recall they measured denotative elaborations (making inferences based on story events) and annotative elaborations (making evaluative comments on events and characters or associating events to their own experience). Denotative elaborations were significantly related to recall, both of which were significantly greater among younger adults. However, annotative comments increased with age and the size of the group. Although Gould (2004) interprets this as a compensatory mechanism for declining memory skills, it could also be viewed as an increase in emphasis on the social world on the part of older adults (e.g., Carstensen, Hansen, & Freund, 1995). In the current experiments moral information processing is expected to draw on the increased emphasis on social relations.
Aging and Moral Development
Few research studies have been done with older adults in the moral domain and those that have been done typically focus on moral reasoning. Cross-sectional studies suggest there is no loss of moral reasoning function with age when educational attainment is controlled (e.g., Chap, 1986). Other studies find differences between older and younger adults. For example, in comparison to younger adults, older adults show greater reflectiveness and breadth in judgment about hypothetical moral dilemmas and tend to focus more on the meaning of personal moral dilemmas (Pratt, Golding and Kerig, 1987). Further, whereas younger adults focused on the external constraints (e.g., duty) on their behavior in personal moral dilemmas, older adults focused on personal virtues (e.g., kindness), similar to those of any age with higher moral judgment scores. Pratt, Golding, Hunter and Norris (1988) studied the moral reasoning of 242 respondents, ages 14-92, on both objective and open-ended measures. Older respondents focused on general, not specific, information and were more likely than younger respondents to assimilate moral dilemma information to their own general cognitive frameworks, consistent with a preference for annotative elaboration, as noted earlier, and a hypothesis of greater synthesis in judgment among the elderly.
Longitudinal studies paint a different picture. Armon and Dawson (1997) present data from a longitudinal study over 13 years during which 23 participants were tested four times with moral judgment interviews. Moral reasoning stage increased sequentially through the lifespan but decreased with advancing age in a slightly curvilinear fashion. Although moral judgment was strongly correlated with education in younger participants, in old age there was only a moderate correlation. Pratt, Diessner, Pratt, Hunsberger, and Pancer (1996) tested a group of adults twice in a 4-year period for moral reasoning stage, complexity of social reasoning, and perspective taking. There were 23 middle aged adults (ages 35-54) and 27 older adults (ages 64-80). Moral reasoning did not decline for either age group, although social reasoning complexity degraded in both groups. Only in the older group did perspective taking decline but was mitigated by education level, health and social-cognitive support. That is, those with more education, better health, and supportive relationships were less likely to decline in perspective taking skills. However, Ligneaur-Herve & Mullet (2005) found that older adults generally were less able to discount information irrelevant from the perspective of another person.
In summary, moral reasoning research maps development through the lifespan finding mixed results for the stability of moral judgment. The present studies focused on a different aspect of moral functioning, moral information processing.
Moral Information Processing
In studies of moral development, there has been a shift away from studying how people reason about hypothetical dilemmas towards a broader study of moral information processing. Like tests of moral reasoning that use dilemmas as stimuli, the ability to process moral information is also marked by developmental differences. Early studies in moral comprehension gave participants moral reasoning statements to paraphrase. Individuals successfully paraphrased reasoning at their own moral stage or below, demonstrating developmental differences in processing the discourse (Rest, 1973; Rest, Turiel, & Kohlberg, 1969; Walker, deVries, & Bichard, 1984). Studies of moral narrative recall indicates that recall of moral reasoning embedded in stories increases with age until adulthood (Narvaez, 1998). Those with moral identities or moral character orientation (i.e., attending more to other people’s character rather than their attractiveness) are more likely to spontaneously infer moral traits and make evaluative judgments of others during reading (Narvaez, Lapsley, Hagele & Lasky, 2006). We expected that because older adults are more focused on interpersonal relations, it would be evident in how they process morally-charged narratives, specifically in having greater memory for moral detail.
In summary, moral information processing research has been conducted primarily with participants in graduate school or younger. Studies of reasoning about moral dilemmas show that older respondents are more likely than younger respondents to focus on general rather than specific information and are more likely to assimilate moral dilemma information to their own general cognitive frameworks, consistent with a hypothesis of greater annotative processing and cognitive synthesis among the elderly (e.g., Pratt, Golding, Hunter & Norris, 1988). The purpose of this study is to examine moral information processing among older adults.
Social cognitive theory (see Lapsley & Narvaez, 2004, for a review) offers a framework for the research questions of this study. Social cognitive theory emphasizes the central importance of cognitive features such as mental representations or schemas and social information processing but also emphasizes the centrality of motivation in the processing of events. Moreover, it considers affect and cognition to be co-regulating interwoven processes. Affect influences memory, salience and attention (Bugental & Goodnow, 1998). Finally, the cognitive-affective social information processing system is dynamic, interacting with the situation, personal meaning, and goals.
Hess (2006) has applied a social cognitive theory to age differences in the processing of social information. If one considers social information processing to be part of adaptive functioning, one can examine how life contexts can change what is adaptive. For example with young adults, it has often been assumed that a person processes information based on current goals (Chen & Chaiken, 1999). More recently, personal meaning has been shown to influence information processing. Meaningful information is more likely to draw attention and increase performance regardless of age (Hess & Auman, 2001). However, for older adults, social information is particularly meaningful (e.g., Carstensen, 1992). We anticipate that moral information will also be salient for older adults.
The current study examined moral text processing and memory, using recognition tests and online inference assessment methods. We hypothesized that, in comparison to younger adults, older adults would be more attuned to moral information because they are more socially oriented and more socially experienced, thus facilitating social information processing (Hess, Osowski, & Leclere, 2005), and in turn, moral information processing. Moral information in the stories would be especially meaningful, in comparison to neutral or nonmoral information, for older adults, resulting in facilitation of processing for moral information.
In Experiment 1, we tested recognition memory using moral stories by examining the three levels of representation: the surface form, textbase, and situation model. For nonmoral information we expected to replicate previously observed patterns (e.g., Radvansky, Zwaan, Curiel, & Copeland, 2001). Namely, both groups should have relatively poor surface form memory, but if there is a difference it was expected that the younger adults would have better memory at this level. Also, it was expected that the young adults would have superior textbase memory, and the older adults would have superior situation model memory. In addition, we expected older adults to have better memory for moral information, relative to nonmoral information. However, at the outset we were unsure whether this would manifest itself at the textbase level, situation model level, or both.
Specifically, as noted earlier, we suspect that older adults place a greater emphasis on social information. Thus, we might expect that this would increase attention to information in the text that more directly conveys social / moral information. As such, memory for that textbase information would be improved for the older adults for moral relative to nonmoral information. Also, because of their increased focus on social-moral information it is possible that older adults would be more likely to draw morality-related inferences, and incorporate them into the situation model. If these inferences were properly probed for in the recognition test, then the strength of the situation model level would also increase for the older adults.
Participants. Thirty-six people were tested in each of the two age groups. The younger adults ranged from 18 to 30 (M = 21.5, SD = 3.1, 78% female), were recruited from the University of Notre Dame, and received partial course credit. The older adults ranged from 60 to 82 (M = 71.8, SD = 6.1, 56% female), were recruited via a newspaper ad, and were paid for their time. The younger adults had fewer years of education (M = 13.8, SD = 1.2) than the older adults (M = 15.8, SD = 2.6), t(70) = 4.25, p < .001, and scored lower on the Shipley vocabulary test (M = 29.4, SD = 3.3) than the older adults (M = 34.1, SD = 3.7), t(70) = 5.66, p < .001. However, the younger adults scored higher on the Salthouse and Babcock (1991) test of processing speed (M = 18.9, SD = 6.2) than the older adults (M = 10.2, SD = 3.4), t(70) = 7.31, p < .001, and on a working memory span test (i.e., comprehension span test; Waters & Caplan, 1996) (M = 19.2, SD = 13.5) than the older adults (M = 4.2, SD = 3.5), t(70) = 6.17, p < .001. In summary, consistent with most research in cognitive aging, the older adults had a higher level of education and vocabulary knowledge, as well as the usual decrements for the processing measures.
Materials. Eight stories were used. There were four experimental stories (78, 89, 95 and 97 sentences long) and four filler stories. Each story was presented one sentence at a time on a computer screen. The experimental and filler stories were randomly ordered. The experimental stories were shortened and slightly revised versions of moral stories used in previous research (e.g., Narvaez, Gleason, Mitchell & Bentley, 1999). The stories contain elements of moral sensitivity, moral reasoning, moral focus, and moral action (Narvaez & Rest, 1995). The experimental stories were the following. (a) “California and the Cattle:” On the way home from helping an elderly neighbor, Cal, from a small Western town, must decide whether she should save and guard the town’s cattle through the night of thunderstorms. This story has the same structure as “The Boy and the Dike” and involves self-sacrifice for the welfare of the community (see Appendix A for the full text of this story). (b) “Home Alone with Jed.” Jed is supposed to stay home one afternoon to take care of his baby brother, but neglects his responsibilities to play football in the park with the other neighborhood boys. (c) “Move to a New City.” The Perez family is moving to Minneapolis in search of employment, and on the way, Mr. and Mrs. Perez’s daughter, Kim, is given too much change at a convenient store. (d) “Malcolm’s Neighborhood.” Malcolm and Tyrone are best friends, and, when Tyrone decides to frame a special needs child for his mistakes, Malcolm wrestles with telling the truth and implicating Tyrone. The filler stories were fictional tales about a farmer rebellion, collecting beanie babies, abstaining from chocolate, and starting a rock band. No analyses were done on the filler stories as they did not contain the manipulations of interest. They were included to obscure the primary purpose of the study.
Measures. After reading all of the stories, participants had several memory and comprehension tasks. First, they were given a recognition test consisting of eight probes for each story. An excerpt from “Move to a New City” is provided below. In this excerpt there are two sentences that were selected to assess memory for nonmoral (general actions or characteristics) and moral information (actions and characteristics about relating to others) respectively.
… They pulled into the gas station. Everybody out for a stretch! Mr. Perez didn't have to convince anyone to get out of the car. They all jumped right out. As her dad filled the gas tank, Kim leaned against the car… She thought about the money. Then she heard her father's voice inside her head. His boss had given him too much money in his paycheck. "If you want to be a good person, you should always try to be honest. And you must always be honest because you are a Perez. We, Perez, are all honest, good people. Everybody knows that." Was she being dishonest by keeping money put in her hand by someone she didn't even know?
Using the Schmalhofer and Glavanov (1986) procedure, four versions of a recognition probe were generated for each of these sentences. For example, in the moral information condition, the sentence “His boss had given him too much money in his paycheck.” was a verbatim sentence. The paraphrase version was “Her father's boss had given him too much money in his check.” The inference version was “Her dad gave back the extra money from the paycheck.” Based on the context of the story, this is an inference that could plausibly be made by readers. Finally, the incorrect version was “Her dad's boss had not given him enough money in his paycheck.” This is thematically consistent, but at odds with the information conveyed by the story. As an example for the non-moral information condition, the sentence “Mr. Perez didn't have to convince anyone to get out of the car.” was a verbatim sentence. The paraphrase version was “Mr. Perez didn't have to tell anyone to step out of the car.” The inference version was “Everyone wanted to get out of the car and stretch.” Finally, the incorrect probe version was “Everyone stayed in the car while Mr. Perez filled the tank.” The appendix shows the moral and neutral probes for the “California and the Cattle” story as well as the full text for the story.