Why Traumatic Events Should Be Differentially Remembered

One of the bases for the expectation of differences in memories of stressful or traumatic experiences and nontraumatic experiences is the classic

Yerkes-Dodson Law of optimal arousal (Yerkes & Dodson, 1908). Briefly, the law predicts a systematic relation between level of arousal and performance, such that when arousal is either too low or too high, performance will be impaired relative to when arousal level is optimal. Stressful or traumatic experiences are expected to increase arousal relative to more benign events. Assuming that the level of arousal associated with a traumatic experience is high, but not too high, application of the Yerkes-Dodson law leads to the prediction that the experience will be well remembered. If the level of arousal associated with the traumatic event exceeds the optimal level, then memory for the experience will be impaired (Easter-brook, 1959).

In addition to differences in how much may be remembered about traumatic relative to nontraumatic events, memories of the different event types may differ qualitatively (see Goodman & Quas, 1997, for discussion). On the one hand, memories of traumatic experiences may have a vivid "flashbulb" quality (e.g., Brown & Kulik, 1977; Winograd & Neisser, 1992). Others suggest that in traumatic circumstances, attention is focused on the central features of the event, with correspondingly poorer memory for more peripheral features (e.g., Christianson, 1992).

The question of whether there are differences in children's memories for traumatic and nontraumatic events is not easy to address. For obvious practical and ethical reasons, researchers cannot intentionally subject children to traumatic events in order to study their memories for them. Instead, we investigate the question by examining the retrospective reports of children who have experienced varying degrees of trauma during stressful medical procedures (e.g., Brown et al., 1999; Goodman, Hirschman, Hepps, & Rudy, 1991; Vandermaas, Hess, & Baker-Ward, 1993) or other negative events, such as hurricanes (e.g., Bahrick, Parker, Merritt, & Fivush, 1998). Most often, the comparisons are across studies or between groups of children. For example, Brown et al. (1999) investigated the influence of trauma on remembering by comparing the reports of children (3- to 5-year-olds) who experienced a painful and embarrassing catheterization procedure (a voiding cystourethrogram—VCUG) to children who experienced a substantially less stressful pediatric exam. Bahrick et al. (1998; see also Parker, Bahrick, Lundy, Fivush, & Levitt, 1998) examined Florida preschool children's recall of the events surrounding Hurricane Andrew, in 1992, as a function of the damage their family sustained during the storm. Overall, stressful events are recalled at least as well as, and in some cases better than, more benign experiences (e.g., Ornstein, 1995; see Fivush, 1998, 2002, for reviews). However, in some studies, children who experienced high levels of stress produced more memory errors relative to children who were less stressed (e.g., Goodman & Quas, 1997).

These investigations inform the question of whether memories for traumatic and nontraumatic events differ quantitatively, qualitatively, or both. However, because the studies were between, rather than within, subjects, they cannot address the question of whether apparent differences between types of events actually are differences between groups. The few studies that have featured within-subjects comparisons of reports of traumatic and nontraumatic events suggest both similarities and differences. For instance, Fivush, Hazzard, Sales, Sarfati, and Brown (2003) interviewed 5- to 12-year-old children about nontraumatic events, such as vacations, parties, and family outings, and stressful or traumatic events, such as parental separation, violent or minor interpersonal altercations, and serious illness or death. Overall, children's narratives about the stressful or traumatic and the nontraumatic events were comparable in length. However, the children mentioned more people and objects and included more descriptions (i.e., use of adjectives, adverbs, possessives, and modifiers) in their accounts of nontraumatic experiences than in their accounts of stressful or traumatic experiences. On the other hand, the children provided more information about internal states (i.e., emotional, cognitive, and volitional states of self or other) when recounting traumatic relative to nontraumatic experiences. In addition, researchers rated narratives about traumatic experiences as more narratively coherent than narratives about nontraumatic experiences. The results of this study suggest that memory narratives about traumatic and nontraumatic events differ, and that narratives about traumatic events are more complete and better integrated, and more internally focused, compared with narratives about nontraumatic events.

A somewhat different perspective on the question is provided by Sales, Fivush, and Peterson (2003). Sales and colleagues examined conversations between parents (primarily mothers) and their 3- to 5-year-old children about a positive event, such as a family vacation, and the stressful or traumatic experience of a medical emergency sufficient to require a trip to the emergency room for the child (e.g., lacerations requiring stitches, broken bones). The conversations about the traumatic experiences were longer than the conversations about the more positive experiences (in Fivush et al., 2003, conversational length did not differ across event types). In addition, both parents and children spent proportionally more time talking about emotions in their conversations about positive events relative to traumatic events (the opposite was observed in Fivush et al., 2003). Conversely, both parents and their children spent proportionally more time discussing the causes of behavior (e.g., "What did you do to get hurt?" p. 192) in their conversations about the traumatic events relative to positive events. The reasons for the somewhat different patterns of findings in Sales et al. and in Fivush et al. are unknown, but may relate to the ages of the children (3- to 5-year-olds and 5- to 12-year-olds, respectively) or the basis for recruitment into the study (in Sales et al., families were recruited precisely because their child had experienced a traumatic event).

Address of the question of whether memories of traumatic and non-traumatic events differ is limited in another way as well: few studies have compared children's memory reports of traumatic and nontraumatic experiences over time. Rather, the majority of research has employed one-time examinations of children's memory reports. Those studies that have included follow-up examinations of children's memory for traumatic events (e.g., Fivush, Sales, Goldberg, Bahrick, & Parker, 2004; Howe, Courage, & Peterson, 1994; Peterson & Whalen, 2001; Quas et al., 1999) have not compared children's reports of traumatic and nontraumatic events. As a result, it is difficult to know whether changes in children's reports over time are due to the traumatic nature of the event or are more event-general. Examination of children's reports of a tornado that devastated the town of St. Peter, Minnesota, permitted us to address both of these limitations, as well as initiate address of the question of whether children's reports were affected by the narrative style of their conversational partners, namely their mothers.

0 0

Post a comment