The above studies indicate that demonstrated skills do not necessarily correspond between everyday tasks (e.g., price-comparison shopping) and traditional academic tasks (e.g., math achievement tests). In other words, some people are able to solve concrete, ill-defined problems better than well-defined, abstract problems that have little relevance to their personal lives, and vice versa. Few of these researchers would claim, however, that academic skills are totally irrelevant to performance in these various contexts. There is evidence that conventional tests of cognition predict both school performance and job performance (Barrett and Depinet, 1991; Schmidt and Hunter, 1998; Wigdor and Garner, 1982). What these studies do suggest is that there are other aspects of cognition that may be independent of academic cognition and that are important to performance, but that largely have been neglected in the measurement of cognition. We also observe this incongruity between conventional notions of real-world skills in research on age-related changes in cognitive skill.

1.1.2 The fun of growing older: Do age-related patterns in practical cognition resemble those in conventional cognition?

Throughout the century of existence of cognitive psychology, many cognitive variables (mostly those contributing to the g-factor—for review, see Berg, in press; Sternberg and Berg, 1992) have been found to be associated with age across the life-span. Most of these associations are rather complex and of curvilinear nature, reflecting rapid growth during the years of formal schooling and slow decline thereafter (Salthouse, 1998). However, the results of research also suggest somewhat different developmental functions for changes in performance on various kinds of cognitive tasks across the adult life span. In particular, data show that older adults commonly report growth in practical skills over the years, even though their academic skills decline (Williams, Denney, and Schadler, 1983).

As for specific cognitive functions, cognition during adulthood is characterized, on one hand, by losses in the speed of mental processes, abstract reasoning, and specific characteristics of memory performance (see Salthouse, 1991, for a review) and, on the other hand, by gains in the metacognitive skill to integrate cognitive, interpersonal, and emotional thinking in a synthetic understanding of the world, self, and others (Labouvie- Vief, 1992, for a review).

The most commonly used theoretical framework adapted for the interpretation of findings on age-related changes in cognitive performance is that of fluid and crystallized cognitive skills (Horn, 1994; Horn and Cattell, 1966). Fluid skills are those required to deal with novelty, such as in the immediate testing situation (e.g., discovering the pattern in a figure sequence). Crystallized skills are represented by accumulated knowledge (e.g., finding a synonym of a low-frequency word). Utilizing this distinction, many studies have demonstrated that fluid skills are relatively susceptible to age-related decline, whereas crystallized skills are relatively resistant to aging (Dixon and Baltes, 1986; Horn, 1982; Labouvie-Vief, 1982; Schaie, 1977/1978), except near the end of one's life.

In addition, Willis and Schaie (1986) studied the relationships between fluid and crystallized skills and everyday cognition (the latter being defined as the skill to perform core activities of independent life—e.g., cooking, managing finances, or using the telephone and measured by a variant of the ETS Basic Skills Test) in the elderly. The researchers reported substantial correlations between performance on the Basic Skills Test and a measure of fluid (r = .83) and crystallized (r = .78) skills.

The majority of these findings, however, were obtained in the framework of cross-sectional methodologies, that is, by comparing different groups of individuals of various ages. When the same individuals are followed across time in the framework of longitudinal design, the findings show that, with respect to fluid cognition, decline does not generally begin until the sixties and loss of crystallized cognition occurs almost a decade later, in the seventies (Schaie, 1996).>