A pervasive real-world memory task is remembering to carry out intended activities at appropriate moments in the future, such as remembering to call one’s mother after returning from work or to hand a message to a colleague when seeing him in the office on the next morning. Such types of tasks are termed prospective memory (PM) tasks (Einstein & McDaniel, 1996). PM has been identified as one of the most frequent everyday memory challenges (e.g., Maylor, 1990), particularly in old age (McDaniel, Einstein, & Rendell, 2008) and an intact PM is considered to be crucial for the maintenance of independent living (Kliegel & Martin, 2003). Therefore, many researchers have focused on the exploration of possible age differences in PM. While age-related deficits were found in standard lab-based PM tasks, age-related benefits occured in naturalistic tasks that are carried out in participants’ everyday lives. This surprising pattern has been called the age-PM-paradox (Rendell & Craik, 2000). It has been supported by a meta-analysis comparing PM age effects found in studies that focused either on lab-based or on naturalistic PM tasks (Henry, MacLeod, Phillips, & Crawford, 2004). However, the mechanisms which are critical in determining the direction of age effects remain poorly delineated. Thus, the overall aim of the research programme presented in the present thesis was to investigate the age-PM-paradox as well as potential cognitive, motivational and emotional mechanisms and processes associated with age-related PM performance. For that purpose, three experimental studies were conducted testing adult age effects in different PM task settings with different task material. Furthermore, several possible underlying mechanisms suggested by the literature on age effects in PM were measured and/ or varied experimentally. The first aim of Study 1 was to cross-validate the age-PM-paradox within a single sample. The second aim was to empirically explore the relative importance of four recently proposed factors (motivation, metacognitive awareness, activity absorption, and control over the task) that may be associated with the direction of age effects inside and outside of the laboratory. For that purpose, 20 young and 20 older adults performed a lab-based and a naturalistic PM task, which were similar in structure and demand. The level of control was experimentally manipulated in both task settings. The remaining possibly influencing factors (motivation, metacognitive awareness, and activity absorption) were assessed via questionnaires in the laboratory and with a daily diary in the field. First, analysing mean level age differences, the paradox was confirmed. Second, exploring possible correlates of the paradox revealed that the level of daily activity absorption (i.e., everyday stress) was the most important mechanism in naturalistic PM performance. Further, high motivation and good metacognitive awareness were associated with age benefits in PM performance in the naturalistic task, while high ongoing activity absorption and low control over the PM cue were related to deficits in lab-based tasks. Thus, Study 1 confirmed the age-PM-paradox within one sample and with carefully matched lab-based and naturalistic tasks. In addition, the results indicate that the relative importance of the suggested factors may vary as a function of setting. While cognitive factors were most influential in the laboratory, motivational and knowledge-based factors were associated with high PM performance in the naturalistic task. The strong association between PM performance in the field and everyday stress highlights the need for future studies exploring the mechanisms underlying this effect. Results from Study 1 suggest that cognitive resources are most influential for PM age effects in the laboratory. Yet, it is not clear, which specific cognitive resources are needed for successful PM performance and if these processes differ between young and older adults. Thus, Study 2 explored the role of executive functions (i.e. shifting, updating and inhibition) as possible developmental mechanisms associated with PM age effects. 170 young and 110 older adults performed a battery of cognitive tests including measures of PM, shifting, updating, inhibition, working memory and speed. A comprehensive set of statistical approaches (e.g. median analyses, structural equation modelling) was used to analyze the possible cognitive correlates in predicting PM performance. First, age effects were confirmed in PM and also obtained in measures of executive control. Moreover, the facets of executive control differently predicted PM performance. Specifically, shifting was the strongest predictor of PM performance in young and older adults as well as for explaining age differences in PM. Thus, Study 2 clarified the role of different facets of controlled attention in age effects in PM and bears important conceptual implications: The results suggest that executive functions are important developmental mechanisms of PM across adulthood beyond working memory and speed. Specifically, shifting appeared to be an essential aspect of cognitive control involved in age-related PM performance. Moreover, examining PM as a latent construct confirmed the convergent and discriminant validity of PM. This demonstrates PM as a separate cognitive construct and suggests that PM is related to, but not identical with, executive control. Study 3 was set out to explore if the amount of cognitive resources needed to successfully perform a PM task in the laboratory can be influenced by the emotionality of the task material. First studies suggested that emotional task material may enhance PM performance in young and older adults by heightening the salience of the task and thereby reducing the need for controlled attention. However, the extent and mechanisms of this effect are still under debate. Therefore, Study 3 explored possible differential effects of PM target cue valence on PM age effects. For that purpose, 45 young and 41 older adults performed a PM task in which emotional valence of the PM cue was manipulated (positive, negative, neutral). Results revealed an interaction indicating that age differences were smaller in both emotional valence conditions compared to the neutral condition. This finding supports an emotionally enhanced memory effect in PM, but only for the older adults as PM performance in young adults was not affected by cue valence. From a conceptual perspective, the results from Study 3 may also contribute to the explanation of the age-PM-paradox, as they suggest that the neutral material usually applied in lab-based studies might overestimate PM age effects. In summary, the present thesis makes an important contribution to the ongoing conceptual debate concerning adult age effects in PM performance assessed in the laboratory versus participants’ everyday lives. Results strongly suggest that mostly different variables may be crucial for understanding PM age deficits in the laboratory and age benefits in naturalistic PM tasks. Successful PM performance in the laboratory seems to require high levels of cognitive resources. The present results suggest that shifting ability is especially relevant in this respect. On a task level the emotionality of the material seems to influence the required amount of cognitive resources as it reduced PM age effects. Everyday stress seems to be particularly important for successful PM performance in the field. Thus, possible future studies should specify the relation between stress and PM as outlined in the general discussion.

A pervasive real-world memory task is remembering to carry out intended activities at appropriate moments in the future, such as remembering to call one’s mother after returning from work or to hand a message to a colleague when seeing him in the office on the next morning. Such types of tasks are termed prospective memory (PM) tasks (Einstein & McDaniel, 1996). PM has been identified as one of the most frequent everyday memory challenges (e.g., Maylor, 1990), particularly in old age (McDaniel, Einstein, & Rendell, 2008) and an intact PM is considered to be crucial for the maintenance of independent living (Kliegel & Martin, 2003). Therefore, many researchers have focused on the exploration of possible age differences in PM. While age-related deficits were found in standard lab-based PM tasks, age-related benefits occured in naturalistic tasks that are carried out in participants’ everyday lives. This surprising pattern has been called the age-PM-paradox (Rendell & Craik, 2000). It has been supported by a meta-analysis comparing PM age effects found in studies that focused either on lab-based or on naturalistic PM tasks (Henry, MacLeod, Phillips, & Crawford, 2004). However, the mechanisms which are critical in determining the direction of age effects remain poorly delineated. Thus, the overall aim of the research programme presented in the present thesis was to investigate the age-PM-paradox as well as potential cognitive, motivational and emotional mechanisms and processes associated with age-related PM performance. For that purpose, three experimental studies were conducted testing adult age effects in different PM task settings with different task material. Furthermore, several possible underlying mechanisms suggested by the literature on age effects in PM were measured and/ or varied experimentally. The first aim of Study 1 was to cross-validate the age-PM-paradox within a single sample. The second aim was to empirically explore the relative importance of four recently proposed factors (motivation, metacognitive awareness, activity absorption, and control over the task) that may be associated with the direction of age effects inside and outside of the laboratory. For that purpose, 20 young and 20 older adults performed a lab-based and a naturalistic PM task, which were similar in structure and demand. The level of control was experimentally manipulated in both task settings. The remaining possibly influencing factors (motivation, metacognitive awareness, and activity absorption) were assessed via questionnaires in the laboratory and with a daily diary in the field. First, analysing mean level age differences, the paradox was confirmed. Second, exploring possible correlates of the paradox revealed that the level of daily activity absorption (i.e., everyday stress) was the most important mechanism in naturalistic PM performance. Further, high motivation and good metacognitive awareness were associated with age benefits in PM performance in the naturalistic task, while high ongoing activity absorption and low control over the PM cue were related to deficits in lab-based tasks. Thus, Study 1 confirmed the age-PM-paradox within one sample and with carefully matched lab-based and naturalistic tasks. In addition, the results indicate that the relative importance of the suggested factors may vary as a function of setting. While cognitive factors were most influential in the laboratory, motivational and knowledge-based factors were associated with high PM performance in the naturalistic task. The strong association between PM performance in the field and everyday stress highlights the need for future studies exploring the mechanisms underlying this effect. Results from Study 1 suggest that cognitive resources are most influential for PM age effects in the laboratory. Yet, it is not clear, which specific cognitive resources are needed for successful PM performance and if these processes differ between young and older adults. Thus, Study 2 explored the role of executive functions (i.e. shifting, updating and inhibition) as possible developmental mechanisms associated with PM age effects. 170 young and 110 older adults performed a battery of cognitive tests including measures of PM, shifting, updating, inhibition, working memory and speed. A comprehensive set of statistical approaches (e.g. median analyses, structural equation modelling) was used to analyze the possible cognitive correlates in predicting PM performance. First, age effects were confirmed in PM and also obtained in measures of executive control. Moreover, the facets of executive control differently predicted PM performance. Specifically, shifting was the strongest predictor of PM performance in young and older adults as well as for explaining age differences in PM. Thus, Study 2 clarified the role of different facets of controlled attention in age effects in PM and bears important conceptual implications: The results suggest that executive functions are important developmental mechanisms of PM across adulthood beyond working memory and speed. Specifically, shifting appeared to be an essential aspect of cognitive control involved in age-related PM performance. Moreover, examining PM as a latent construct confirmed the convergent and discriminant validity of PM. This demonstrates PM as a separate cognitive construct and suggests that PM is related to, but not identical with, executive control. Study 3 was set out to explore if the amount of cognitive resources needed to successfully perform a PM task in the laboratory can be influenced by the emotionality of the task material. First studies suggested that emotional task material may enhance PM performance in young and older adults by heightening the salience of the task and thereby reducing the need for controlled attention. However, the extent and mechanisms of this effect are still under debate. Therefore, Study 3 explored possible differential effects of PM target cue valence on PM age effects. For that purpose, 45 young and 41 older adults performed a PM task in which emotional valence of the PM cue was manipulated (positive, negative, neutral). Results revealed an interaction indicating that age differences were smaller in both emotional valence conditions compared to the neutral condition. This finding supports an emotionally enhanced memory effect in PM, but only for the older adults as PM performance in young adults was not affected by cue valence. From a conceptual perspective, the results from Study 3 may also contribute to the explanation of the age-PM-paradox, as they suggest that the neutral material usually applied in lab-based studies might overestimate PM age effects. In summary, the present thesis makes an important contribution to the ongoing conceptual debate concerning adult age effects in PM performance assessed in the laboratory versus participants’ everyday lives. Results strongly suggest that mostly different variables may be crucial for understanding PM age deficits in the laboratory and age benefits in naturalistic PM tasks. Successful PM performance in the laboratory seems to require high levels of cognitive resources. The present results suggest that shifting ability is especially relevant in this respect. On a task level the emotionality of the material seems to influence the required amount of cognitive resources as it reduced PM age effects. Everyday stress seems to be particularly important for successful PM performance in the field. Thus, possible future studies should specify the relation between stress and PM as outlined in the general discussion.