Leonard W. Poon

The effect of age on speed of lexical access

The present study concerns the effect of aging on speed of lexical access. Normative word-frequency was manipulated in a lexical decision task with older and younger adults. Three methods of comparing processing time in the lexical access stage across age groups were evaluated: the subtraction method, the analysis of covariance, and the additive-factor method. The adequacy of the subtraction method was questioned because of the difficulty of finding a control task identical to the experimental task in all processing stages except lexical access. The analysis of covariance was considered to be inconclusive in this kind of application because of the correlation between the covariate and the independent variable, age. The additive-factor method appeared to be a viable alternative to these two analyses. Within this framework, the lack of an interaction between the two variables, age and word-frequency, indicates that the factors affected different stages of processing. Since word-frequency has been shown to affect the lexical access stage, the present result was taken to suggest that aging does not affect speed of lexical access.

Age, imagery, and practice in paired-associate learning

Elder and college-aged subjects performed a series of paired-associate tasks designed to investigate age differences in the use of mnemonic imagery as a strategy in verbal learning. Subjects were tested on three occasions at two-week intervals. Two ten-pair lists of concrete-noun paired associates were learned at each session. Standard (no-imagery), self-generated, and experimenter-provided imagery instructions were used. Results indicated that imagery instructions are quite beneficial for elders in the short run. However, elders tended not to use the strategy effectively after a two-week interval unless reminded to do so. Results also indicated that elders were capable of generating useful learning strategies on their own when given sufficient experience with the task.

Effects of priming in word retrieval.

Three studies were carried out to investigate orthographic and semantic priming effects in word retrieval. In this paradigm a prime is presented prior to the definition of a target word. The task is to produce the target word. We view the process of word retrieval as requiring access from a conceptually organized semantic network to an orthographically and phonemically organized lexical network. Primes that were orthographically (and phonemically) related to the target words were found to facilitate word retrieval. Both semantically related and unrelated primes inhibited word retrieval. Inhibition from both of these sources is attributed in part to the orthographic unrelatedness of these primes. This explanation is shown to be consistent with data from several other word retrieval studies (Brown, 1979; Roediger, Neely, & Blaxton, 1983). The word retrieval task is contrasted with the lexical decision task. It is suggested that the differential effects of orthographic and semantic priming in word retrieval and lexical decision are due to differences in retrieval processing between the two tasks.

On the use of word association norms in aging research

The word associations of young, middle-aged, and older adults were compared to cohort matched normative word associations. Cohort differences in most common response were observed for a subset of stimulus words which did not have strong primary responses in the norms. For those stimulus words assumed to be free of cohort differences, there were no significant differences between the three age groups in amount of agreement with the normative responses. It is concluded that standard word association norms are as reliable for older individuals as for the young provided that words with strong associates are selected.

Age-related differences in the time course of encoding

The speed with which normal elderly and young subjects encoded digits in a speeded same/different matching task was investigated. One condition required a judgment based on physical identity, while in a second condition judgments were based on an odd/even dichotomy. The pattern of results demonstrated a slowing in the speed of encoding in the older subjects. This effect, while statistically significant, was slight.