Predictors of time-based prospective memory in children

https://doi.org/10.1016/j.jecp.2008.08.006Get rights and content

Abstract

This study identified age differences in time-based prospective memory performance in school-aged children and explored possible cognitive correlates of age-related performance. A total of 56 7- to 12-year-olds performed a prospective memory task in which prospective memory accuracy, ongoing task performance, and time monitoring were assessed. Additional tests of time estimation, working memory, task switching, and planning were performed. Results showed a robust relationship between age and prospective memory performance even after controlling for ongoing task performance. Developmental differences in time monitoring were also observed, with older children generally adopting a more strategic monitoring strategy than younger children. The majority of age-related variance in prospective memory task performance could be explained by cognitive resources, in particular planning and task switching. In contrast, no further independent contribution of time estimation was observed. Findings are in line with the development of strategic behavior, as well as executive functioning, in school-aged children.

Introduction

Being able to remember to do something in the future (e.g., remembering to take medication in time) is a vital skill for day-to-day functioning as autonomous individuals (Shallice & Burgess, 1991). This everyday memory task has been named prospective memory (PM) (Ellis, 1996). Specifically, the task of remembering to do something at some time point in the future is referred to as time-based prospective memory (TBPM) (Einstein & McDaniel, 1996). From a conceptual perspective, besides retrospective memory for the content of the delayed action, PM involves executive functions for the appropriate and self-initiated execution of the delayed action, for example, monitoring for the target time (Kliegel, Jäger, Altgassen, & Shum, 2008). A particular challenge for succeeding in this type of memory task is that the prospective task typically must be carried out in the midst of performing an attentionally demanding ongoing task (Ellis & Kvavilashvili, 2000).

From a developmental perspective, PM skills appear to emerge relatively early in childhood, with even preschoolers demonstrating the ability to remember to do something in the future under some conditions (Guajardo and Best, 2000, Kliegel and Jäger, 2007, Somerville et al., 1983). It has been suggested that the early development of PM has a particular functional value for the development of goal-directed behaviors in general (Winograd, 1988). Moreover, TBPM skills are increasingly called on as children enter the school environment, where they are expected to complete nonroutine tasks such as being in a particular location at a particular time, for example, going to the school hall at 12 o’clock to collect a letter for their parents. Therefore, it is surprising that, despite its clear importance for everyday functioning, age-related TBPM performance in school-aged children has not received much research attention in the developmental domain. Thus, the current study aimed to examine possible age differences in TBPM performance across school-aged children and to explore cognitive resources potentially underlying age-related TBPM performance. In the few existing studies on this topic, TBPM performance has been evaluated in terms of task accuracy (whether the delayed intention is carried out on one’s own initiative) and time monitoring behavior (how time is monitored to enable action at the appropriate point).

Studies investigating the development of TBPM accuracy generally show an inconsistent pattern of results and are likely to have been hampered by ceiling effects, inadequate measuring/reporting of ongoing task performance, and a lack of testing retrospective memory for task instructions (Kvavilashvili, Kyle, & Messer, 2008). Ceci and Bronfenbrenner (1985) found that the majority of their 10- and 14-year-old participants could remember to remove cupcakes from the oven after a 30-min delay (TBPM task) during which they played a computer game (ongoing task). Children could monitor the time by checking a clock that was mounted on the wall behind them. Unfortunately, performance of this single prospective act was largely at ceiling, potentially masking any age-related differences in TBPM performance, and ongoing task performance was not reported.1 A similar critique also holds for a study by Nigro, Senese, Natullo, and Sergi (2002), who reported superior performance of 7- to 11-year-olds on one PM task over another but did not report task performance in relation to age and retention of task instructions. In contrast, Kerns (2000) observed age effects in TBPM accuracy of 7- to 12-year-olds who performed a computer driving game (ongoing task) in which they were required to monitor the level of gas in the tank to select the right time to refuel (embedded TBPM task). Younger children ran out of gas more frequently than older children, and no ceiling effects were observed. However, because ongoing task performance was not measured, it is not clear whether the age effects reported resulted from true differences in TBPM itself or from differences in the requirements of the ongoing task. In addition, children were not assessed for recall of the TBPM task instruction at the end of the game. Mäntylä, Carelli, and Forman (2007) asked 8- to 12-year-olds and young adults to press a button every 5 min (TBPM task) while watching a video (ongoing task). Participants could monitor time by pressing a different button to make a clock appear on the screen. In the absence of ceiling effects, no age effects in TBPM performance were reported, with all participants achieving approximately 80% of correct responses. Although absorption in the ongoing task (video watching) could not be quantified, participants’ understanding of, and memory for, task instructions was verified.

In sum, possibly due to procedural limitations, the results reviewed above provide inconclusive evidence with regard to age differences in TBPM accuracy. Thus, the first aim of the current study was to test for possible age differences in TBPM accuracy while controlling for ongoing task performance.

It has been suggested that to successfully perform TBPM tasks, we adopt a strategy to monitor the passing time. This strategy is indicated by the pattern of time checks observed. In adults, this involves checking the time relatively often at the beginning of a delay period, less frequently in the middle, and then very often toward the end—resulting in a U- or J-shaped distribution of time checking (Harris & Wilkins, 1982). Results so far indicate that children appear to adopt a similar time monitoring strategy when performing TBPM tasks, although the exact shape of the monitoring pattern is still under debate. Ceci and Bronfenbrenner (1985) reported a U-shaped monitoring function, Kerns (2000) reported a J-shaped function, and Mäntylä and colleagues (2007) observed accelerated monitoring patterns in their participants with no initial peak. Moreover, there is some evidence supporting age differences in time monitoring; in general, younger children checked the clock (or fuel gauge) more frequently than older children or adults (Ceci and Bronfenbrenner, 1985, Kerns, 2000, Mäntylä et al., 2007). In addition, Ceci and Bronfenbrenner (1985) found that children monitored the clock more strategically in their own homes than in an unfamiliar laboratory setting. This led them to differentiate between U-shaped and linearly increasing patterns of clock monitoring. They further found evidence to support a relationship between monitoring type and TBPM accuracy; children who engaged in linearly increasing monitoring tended to respond late to the TBPM task.

In sum, there is some evidence that a relationship may exist among children’s age, time monitoring, and TBPM accuracy. However, evidence regarding an age-related pattern of time monitoring in TBPM tasks is inconclusive. Therefore, the second aim of this study was to measure participants’ time monitoring patterns, explore possible age differences in time monitoring, and compare them with any age differences observed in TBPM accuracy.

Conceptually, the studies reviewed above mainly explored age differences in TBPM performance on a rather descriptive level. Extending this approach, the current study rests on the conceptual rationale that age differences in TBPM are to be expected in school-aged children because of the involvement of specific cognitive resources (i.e., memory and executive functions) in TBPM that show well-documented age effects in this age range. Thus, besides examining age differences in TBPM performance in children per se, the current study set out to explore developmental cognitive mechanisms that may be associated with those age differences. Current conceptual frameworks of PM suggest that TBPM is dependent on specific cognitive resources such as working memory and may also draw on the more specific cognitive skill of time estimation (Glicksohn & Myslobodsky, 2006). According to a general three-phase model of TBPM consisting of (a) forming an intention, (b) maintaining the intention in mind while monitoring the time, and (c) interrupting the ongoing activity to initiate and execute the intention at the appropriate time (see, e.g., Ellis, 1996), at least three basic cognitive resources can be assumed to be associated with TBPM: planning (Phase A), working memory (Phase B), and task switching (Phase C) (for similar proposals for PM in general, see Kliegel et al., 2002, McDaniel et al., 1999).

Importantly, these processes and the development of these processes have been studied extensively in the child development literature. Basic cognitive resources such as working memory have long been associated with the temporal organization of behavior and are known to develop throughout childhood (see, e.g., Anderson, 1998, Anderson et al., 1996, Gathercole et al., 2004, Levin et al., 1991, Schneider and Pressley, 1997, Welsh et al., 1991). Specifically, all three cognitive resources addressed in the current study are well known to develop across childhood, with working memory span increasing (e.g., Schneider & Björklund, 1998) and both planning and task switching becoming more efficient (for an overview, see Goswami, 2008). Thus, although conceptual models of PM and the literature on general cognitive development both strongly suggest a relationship between developments in TBPM and executive functioning, this possibility was explored in only two of the developmental studies discussed above. Kerns (2000) observed correlations between TBPM accuracy and performance on measures of inhibition and visuospatial working memory. She did not, however, find any relationship between those measures and monitoring behavior. Mäntylä and colleagues (2007) assessed performance on several tests of executive functioning corresponding to the three components—inhibition, updating, and shifting—suggested by Miyake and colleagues (2000). Exploratory factor analyses indicated that inhibition and updating were particularly important for time monitoring; this relationship was not observed for the shifting component. Mäntylä and colleagues (2007) suggested that updating and retaining dynamic event information in working memory contributes to a sense of temporal continuity, so that individuals with efficient updating and inhibition functions are able to rely longer on this temporal information when monitoring deadlines. In contrast, individuals with difficulties in temporary maintenance and elaboration of working memory content may experience discontinuities in their sense of time, leading to an earlier and more frequent reliance on external time keeping.

Besides executive functioning, it has also been proposed that TBPM may rely partly on time estimation abilities (Goldstein, 2005, Mäntylä and Carelli, 2005; see also Carelli, Forman, & Mäntylä, 2008). The rationale for expecting this influence rests on task-specific demands typically associated with TBPM paradigms. Many TBPM tasks require the representation of a discrete temporal unit (e.g., press a key every 5 min); hence, the ability to estimate such durations may itself be a predictor of time monitoring and/or TBPM accuracy. Several different types of time estimation task can be differentiated in which participants are asked to estimate the length of a given duration (e.g., the time elapsed between two tones), produce a target duration themselves, reproduce an experienced duration themselves, or retrospectively make a duration judgment after the time period has passed, thereby being unaware of the requirement to monitor time (Block, Zakay, & Hancock, 1999). Results of a recent meta-analysis showed that, compared with adolescents and young adults, children make larger time estimations of experienced durations, with the precision of these estimations increasing from 7 to 10 years of age, and they tend to make shorter time reproductions but are generally able to make comparable productions of time durations (Block et al., 1999). The empirical evidence for time estimation as a potential factor underlying TBPM is both scarce and contradictory. Mäntylä and Carelli (2005) investigated time estimation as a possible predictor of performance on the TBPM task discussed above (Mäntylä et al., 2007). No age differences were observed between school-aged children and young adults who reproduced durations of between 4 and 36 s, and time reproduction did not correlate with time monitoring performance on the TBPM task. In contrast, Goldstein (2005) reported a significant relationship between time estimation and both TBPM accuracy and clock monitoring in a sample of young adults. Participants who were accurately able to produce 1-min intervals made significantly more timely responses on a TBPM task (to press a button every 7 min) and checked the clock more frequently than did inaccurate estimators.

Taken together, previous studies provide mixed evidence with regard to age differences in TBPM. Thus, the first aim of this study was to examine individual and developmental differences in TBPM performance while controlling for individual and developmental differences in ongoing task performance. The second aim was to investigate the possible relation among age, time monitoring, and TBPM accuracy. The third aim was assess whether age-related variance in children’s TBPM performance is related to development in executive functions, including planning, working memory, task switching, and time estimation. To address these aims, 7- to 12-year-olds performed a TBPM task in which task accuracy, ongoing task performance, and time monitoring were assessed. Their performance on standardized tests of planning, working memory, and task switching was also measured. In addition, we assessed children’s time estimation skills; for the first time, children were asked to estimate the same time span as they had to observe for the TBPM task. The 7- to 12-year age range chosen is consistent with the age span examined in most previous studies of children’s PM and is an age span for which developmental differences in the cognitive resources targeted have been reported.

Section snippets

Participants

A total of 56 7- to 12-year-olds participated in the study. The mean age was 10 years 1 month (SD = 20.18 months, range = 7 years 2 months to 12 years 7 months), with 26 boys and 30 girls being divided evenly across the age range. An additional 4 children were tested but excluded from the final sample because 3 were unable to tell the time and 1 failed to answer comprehension questions at the end of the TBPM test. All participants were White and from German-speaking middle-class families. Children

Results

To address the main research questions of this study, key variables were analyzed as follows. Three aspects of TBPM task performance, their interrelationships, and their relation to age were considered: (a) accuracy in implementation of the TBPM task (correct pressing of the yellow key), (b) ongoing task accuracy (correct responses on the one-back task), and (c) time monitoring (frequency and distribution of clock checks). To address the question of possible predictors of age-related TBPM

Discussion

The current study revealed four major findings. First, results showed a robust relationship between age and TBPM accuracy even after controlling for ongoing task performance. Second, developmental differences in time monitoring were observed, with older children generally showing a more accelerated monitoring pattern than younger children. Third, hierarchical regression analysis revealed that the majority of age-related variance in TBPM performance could be explained by cognitive resources,

Acknowledgments

We acknowledge Aline German, Martina Hürschler, and Elsa Sägesser for help with data collection as well as all schools and children who participated.

References (49)

  • S.J. Ceci et al.

    “Don’t forget to take the cupcakes out of the oven”: Prospective memory, strategic time-monitoring, and context

    Child Development

    (1985)
  • G.O. Einstein et al.

    Retrieval processes in prospective memory: Theoretical approaches and some new findings

  • G.O. Einstein et al.

    Aging and prospective memory: Examining the influences of self-initiated retrieval

    Journal of Experimental Psychology: Learning, Memory, and Cognition

    (1995)
  • J. Ellis

    Prospective memory or the realization of delayed intentions: A conceptual framework for research

  • J. Ellis et al.

    Prospective memory in 2000: Past, present, and future directions

    Applied Cognitive Psychology

    (2000)
  • H. Emslie et al.

    Behavioral Assessment of the Dysexecutive Syndrome for Children. Bury St

    (2003)
  • S.E. Gathercole et al.

    The structure of working memory from 4 to 15 years of age

    Developmental Psychology

    (2004)
  • J. Glicksohn et al.

    Timing the future

    (2006)
  • A. Goldstein

    July). Time-based prospective memory: Effects of cognitive load, time estimation

    (2005)
  • U. Goswami

    Cognitive development: The learning brain

    (2008)
  • J.E. Harris et al.

    Remembering to do things: A theoretical framework and an illustrative experiment

    Human Learning

    (1982)
  • K.A. Kerns

    The CyberCruiser: An investigation of development of prospective memory in children

    Journal of the International Neuropsychological Society

    (2000)
  • Kliegel, M., Brandenberger, M., & Aberle, I. (in press). Effect of motivational incentives on prospective memory...
  • M. Kliegel et al.

    Clinical neuropsychology of prospective memory

  • Cited by (0)

    View full text