Abstract
In this study, the authors show that Hannon and Daneman's (Citation2001, Journal of Educational Psychology, 93, 103–128) component processes task can be used to investigate individual differences in older readers' comprehension performance, and to determine which components of comprehension are most susceptible to declines with normal aging. Results revealed that the ability to remember new text information, to make inferences about new text information, to access prior knowledge in long-term memory, and to integrate prior knowledge with new text information all accounted for a substantial proportion of variance in older adults' reading comprehension performance. Although there were age-related declines in all of these component processes, the components associated with new learning were more susceptible to age-related declines than were the components associated with accessing what already is known. The findings suggest that age-related declines in reading comprehension might be a consequence of declines in a number of component processes rather than one specific process.
Notes
1These knowledge integration statements are called high-knowledge integration statements in Hannon and Daneman (Citation2001). Hannon and Daneman's component processes task also included two other types of knowledge integration statements (low-knowledge integration and medium-knowledge integration) that required less complex knowledge access and integration processes. The high-knowledge integration statements, which were more strongly correlated with reading comprehension, are the ones used in the current study. For simplicity, we refer to them simply as knowledge integration statements.
a With the exception of the speed component, means, standard deviations, and ranges for the components in the component processes task are reported as percentages.
∗p < .05; ∗∗p < .01; ∗∗∗p < .001.
2Positive correlations reflect speed-accuracy trade-offs, whereas negative correlations (like the one we observed here) indicate that higher performers on the accuracy measures tend to be fast responders and low performers on the accuracy measures tend to be slow responders.
∗p < .05; ∗∗p < .01; ∗∗∗p < .001.
3Again, this negative correlation indicated that higher performers on the accuracy measure also tended to be faster responders.