Younger and older adults read short expository passages across 2 times of m
easurement for subsequent comprehension or recall. Regression analysis was
used to decompose word-by-word reading times into resources allocated to wo
rd- and textbase-level processes. Readers were more sensitive to these dema
nds when reading for recall than when reading for comprehension. Patterns o
f resource allocation showed good test-retest reliabilities and were predic
tive of memory performance. Within age group, resource allocation parameter
s were not systematically correlated with other individual-difference measu
res, suggesting that strategies of on-line resource allocation may be a uni
que source of individual differences in determining comprehension of and me
mory for text. Age differences in allocation patterns appeared to reflect g
eneral slowing among the older adults. Because older adults showed equivale
nt memory performance to that of younger readers, the reading time data may
represent the on-line resource allocation needed for comparable outcomes a
mong older and younger readers.