Effect of Aging on Ability to Cope With Uncertainty in Learning Task
Studies of the effects of aging on performance in cognitive tasks have shown decreasing performance with age. However, nearly all previous studies of learning in the elderly employed tasks which are deterministic in nature (e.g., learning paired associations of symbols or words).
In contrast, learning probabilistic judgment is critical throughout life. Consequently, elderly people may be required to cope just as well as middle-aged adults, young adults, adolescents, or children with new environments that contain new probabilistic relations.
Learning these relations can be a difficult task for them because of the uncertainty inherent in the environment. Uncertainty is typical of situations requiring Multiple Cue Probability Learning (MCPL). The MCPL paradigm captures the essential qualities of many common life learning situations, notably the uncertainty as to the strength of cue-criterion relationships, and the probabilistic relations.
Most of the MCPL experiments to date have been conducted with young adults, usually students. The general finding is that uncertainty affects learning performance (e.g., Naylor & Schenck, 1968).
In the first MCPL study conducted on elderly people (Chasseigne, Mullet, & Stewart, 1997), we studied the relation between age and the ability to learn direct and inverse probabilistic relationships between cues and criterion.
In one condition, we showed that when all cue-criterion relationships were direct, elderly participants were able to learn nearly as well as young participants. The level of task predictability (R2) was held constant and equal to 0.93.
The aim of my latest study (Chasseigne, Grau, Mullet, & Cama, 1999) was to investigate the effect of aging on the ability to cope with uncertainty in a learning task, keeping constant the direction of cue-criterion relationships.
A total of 220 individuals, 55 in each of four age groups (18-25, 40-50, 65-74, 75-90 years old) participated in this experiment. The materials consisted of four sets of 30 cards each, showing three cue values in the form of vertical colored bars (pink) whose heights varied from card to card. Five levels of task predictability were used (0.96, 0.80, 0.64, 0.48, and 0.32). The ecological validities of the three cues in each of the five uncertainty conditions were equal (0.56, 0.52, 0.46, 0.40, and 0.33).
The participants were told that the task which faced them was a weather forecasting task. They were asked to learn the relationships between the levels of the three indicators and the pleasantness of the next day's weather. They were presented with different scenarios, each of which was characterized by sets of three cue values displayed on the front of different cards and the actual value of the criterion displayed on the back of the same cards (outcome feedback, OFB). The subjects were also told that an exact weather forecast was nearly impossible because of a myriad of other factors acting independently.
Our hypothesis was that the relationship between uncertainty in the task and knowledge and control in the participants would not be strongly dependent upon age. In other words, we expected few differences between elderly people and young people in the way they cope with uncertainty in a learning task when all relationships are direct ones. We also expected no Uncertainty x Age interactions. This hypothesis was based on two considerations.
1. In contrast to a deterministic task, where elderly people would be less apt to keep information in memory than young people (Salthouse, 1994), and thus would show lower performance, our task was a probabilistic one. MCPL does not imply learning through associations. It implies the learning of the relationship between the cues and the criterion. This relationship is independent of predictability conditions. As a consequence, elderly people could perform well.
2. As suggested by Brehmer (1974), direct relationships are the relationships assumed by default. If nothing suggests alternative relationships at work in the task, there is no reason for participants, whatever their age, to change their mind about the nature of the relationships.
Our hypothesis of few age differences was supported by data. As in our preceding study (Direct Relationships condition), the present study found a slight effect of age on control and knowledge. Moreover, no Uncertainty x Age interaction was significant. As in several previous studies a strong effect of uncertainty on control was found. Uncertainty also had a slight effect on knowledge. Finally, in MCPL tasks with outcome feedback, the performance of elderly people is nearly equivalent to that of young people when relationships are direct, regardless of the amount of uncertainty in the task.