This Article Abstract Full Text (PDF) Services Download to citation manager PubMed PubMed Citation

Word Frequency Effects in Priming Performance in Young and Older Adults

^a Department of Psychology, Washington University, St. Louis, Missouri

Rowena Gomez, Department of Psychology, Box 1125, Washington University, 1 Brookings Drive, St. Louis, MO 63130 E-mail: rggomez{at}artsci.wustl.edu.

Decision Editor: Toni C. Antonucci, PhD

	Abstract

TOP Abstract Experiment 1 Results and Discussion Experiment 2 Methods Results and Discussion Experiment 3 Methods Results and Discussion General Discussion References

 
The present study examined word frequency effects on implicit priming in older adults compared with younger adults. In Experiment 1 participants performed a spelling test consisting of primed and unprimed homophones (e.g., mourning) and nonhomophones (e.g., militant). Older adults spelled more unprimed, low-frequency homophones than did younger adults, suggesting that there are age-related differences in base-rate spelling of lower frequency homophones. Experiment 2 involved a word-fragment completion task that primed both high- and low-frequency words. Young adults showed larger priming effects for low-frequency words, whereas older adults showed smaller and similar priming effects for high- and low-frequency words. Experiment 3 replicated the finding that word frequency has no effect on priming performance in older adults on a word-fragment completion task. These studies found differential word frequency effects on priming performance between young and older adults.

Although imperfections in human memory are apparent at all ages, many researchers agree that these flaws increase in late adulthood (i.e., Poon 1985). For example, explicit memory is known to decline with age (i.e., Craik and McDowd 1987). Explicit memory requires the conscious and intentional recollection of recently presented information, such as in recall and recognition tasks. This can be contrasted with implicit memory, which occurs when previous experiences facilitate or influence performance without conscious and intentional recollections of those experiences on current performance (Schacter 1987). The results of research on the effects of age on implicit memory are not as consistent as research on explicit memory. Many studies have found a relative preservation of implicit memory with age (e.g., Howard and Howard 1989; Java 1992; Java and Gardiner 1991; Light and Albertson 1989; Light and Singh 1987; Park and Shaw 1992), whereas others have concluded that true age differences exist (e.g., Chiarello and Hoyer 1988; Hultsch, Masson, and Small 1991; McEvoy, Nelson, Holley, and Stienicki 1992; Rose, Yesavage, Hill, and Bower 1986).

Although this study does not aim to resolve the controversy of the preservation or decrement of implicit memory with age, it does examine the effect of word frequency on implicit-memory performance between young and older adults. Word frequency refers to the rate of occurrence for words in language. As the frequency of the usage of the words in the language increases, the degree of automaticity and speed of processing also increase (Macleod and Kampe 1996). Thus high-frequency words require less attention and are processed at a faster rate than low-frequency words. High-frequency words are also more familiar and more accessible than low-frequency words.

Because low-frequency words have a lower baseline of familiarity, these words are thought to have a greater increase in familiarity compared with high-frequency words with repeated presentation. Thus, low-frequency words are associated with a larger priming effect than are high-frequency words in the lexical decision task, the masked-word identification task, and the word-stem completion task (e.g., Duchek and Neely 1989; Forster and Davis 1984; Kirsner, Milech, and Standen 1983; Macleod and Kampe 1996).

Most priming studies involving word frequency have examined only young adults. Only one study directly investigated the effect of word frequency on priming and aging (Balota and Ferraro 1996). They conducted a rhyme-judgment task (priming task) followed by a lexical-decision task (implicit-memory task). On the basis of the response latency data, both healthy young adults and healthy older adults benefited from repetition for words (repetition priming) in the lexical decision task and produced similar priming effects for combined high- and low-frequency words. When high- and low-frequency words were separated, however, the priming effect was larger for low-frequency words compared with high-frequency words. This word frequency effect on priming was slightly larger for older adults compared with younger adults.

Other priming studies involving homophones have indirectly examined word frequency effects on priming with age. Homophones are words that sound alike but have different spellings, meanings, and word frequencies. In studies by Davis and colleagues 1990 and Rose and colleagues 1986, the lower frequency homophones were inserted in the questions of the priming task—a general knowledge questionnaire. Afterward, a homophone spelling task was administered. Without priming, the higher frequency homophones were expected to be spelled more often than the lower frequency counterparts because the higher frequency homophones are more consciously accessible. Priming performance was found to be significantly worse for older adults compared with young adults. In both studies (Davis et al. 1990; Rose et al. 1986), however, they reported an unusually high baseline for older adults spelling out the low-frequency homophones without priming (i.e., their baseline was considerably higher than was that of the younger adults). The lower frequency homophones appear to have been more available and familiar to the older adults than to the younger adults in a natural or unprimed condition. The higher usage of unprimed, low-frequency words may have obscured the magnitude of priming effects for older adults. These homophone-spelling studies suggest that the word frequency pattern on priming performance may be different for older adults compared with younger adults (e.g., Davis et al. 1990; Howard 1988; Rose et al. 1986). The relatively higher usage of lower frequency homophone spellings by older adults in the unprimed condition also indicates that older adults may have different frequencies for words compared with young adults.

In sum, two types of studies—the lexical decision task study (Balota and Ferraro 1996) and the homophone-spelling studies (e.g., Davis et al. 1990, Rose et al. 1986)—have examined word frequency and repetition priming in older adults. The lexical decision task study indicated greater priming effects for lower frequency words compared with high-frequency words for both young and older adults, but this low-frequency word advantage on priming was found to be larger for the older age group than the young age group. In contrast, the homophone spelling studies that primed the lower frequency words (homophones) indicated smaller priming effects for older adults compared with young adults. The present study further investigated the pattern of word frequency effects on priming performance for older and young adults by using two types of priming paradigms, one involving homophone spelling and one involving word-fragment completion.

	Experiment 1

TOP Abstract Experiment 1 Results and Discussion Experiment 2 Methods Results and Discussion Experiment 3 Methods Results and Discussion General Discussion References

 
Methods
Participants
A total of 48 participants (24 young and 24 older adults) participated in this study. Nine male and 15 female young adults were undergraduates randomly selected from a volunteer pool maintained by the psychology department. The mean age of these participants was 20 years old (SD = 1.56; range = 18–23) with a mean of 14.5 years of education (SD = 1.25; range = 13–16). The older adult group included 7 men and 17 women. They were recruited from a separate volunteer pool of people from the community maintained in the Aging and Development Program in the psychology department. The older adults had a mean age of 73 years (SD = 5.65; range = 63–85) and an identical mean of 14.5 years of education (SD = 1.25; range = 13–16). The older adults were administered the Short Blessed Dementia Scale (Katzman et al. 1983) to screen for major cognitive or memory problems. The mean score for this cognitive functioning task was 2.16 (SD = 2.17). Both young and older age groups were administered the vocabulary subtest of the Wechsler Adult Intelligence Scale–Revised (WAIS–R; Wechsler 1981) to assess verbal ability. The mean raw score of the young adult group was 53.79 (SD = 8.36). The mean raw score of the older age group was 56.25 (SD = 5.38). The performances of the two groups on this vocabulary task did not significantly differ, F(1,46) = 1.46, p > .10.

Materials
Homophone and nonhomophone word sets.
The homophone words were selected from Galbraith and Taschman 1969 list of 88 homophone units and Terrell and Meadows 1985 homophone list. Using norms for word frequency prepared by Kucera and Francis 1967, we considered the lower frequency words with raw frequencies above 45 per million to be within the acceptable range.

The nonhomophone words were selected so as to have similar word frequency, be of the same number of letters, and begin with the same letter as the lower frequency homophone. A total of 216 words (72 homophone pairs and 72 matched nonhomophones) were selected for this study. The 72 lower frequency homophones and their 72 matched nonhomophones were randomly assigned to one of four word sets. Each of these word sets, therefore, contained 18 lower frequency homophones and 18 matched nonhomophones. Each set was then paired with every other set to produce a total of six combinations of word sets; each combination, therefore, contained 36 lower frequency homophones and 36 matched nonhomophones. Each homophone occurred in both primed and unprimed conditions across participants.

Trivia questionnaire.
The participants were given a 72-item trivia questionnaire about various topics (e.g., history, literature, entertainment). Each question in this task contained either a low-frequency homophone or a nonhomophone. One example of a question is "In Hamlet, who was the true heir to the throne?" In this question, the lower frequency homophone is heir and its higher frequency counterpart is air.

Each participant was administered one of six forms of the trivia task. These six forms corresponded to the six combinations of the homophone and nonhomophone word sets. The word sets presented to the participant in this task were considered the primed word sets. The word sets not included in the trivia questionnaire were the unprimed words.

The spelling task.
In the test phase all participants performed a 72-item spelling test. Homophones and nonhomophones were randomly assigned to one of two spelling lists. Alternate forms of these spelling lists were constructed by reversing the order of item presentation to produce a total of four spelling lists. Each spelling list included 36 words from the primed word set (18 homophones and 18 nonhomophones) and 36 from the unprimed word set (18 homophones and 18 nonhomophones).

Participant awareness questionnaire.
After the end of the spelling task the participants rated their level of awareness about the relatedness of the trivia task and the spelling on a 5-point Likert scale ranging from completely unaware (1) to completely aware (5). The young adults were significantly more aware of the relation between the trivia and spelling tasks (M = 2.83, SD = 1.49) than the older adults were (M = 1.25, SD = 0.90), F(1,46) = 19.82, p < .001. Notably, no significant correlation was found between awareness and priming effect, r = .15, p > .10.

Procedure
The participants came to the laboratory and were told that the study examined general cognitive functioning. They were told that the study involved various tasks such as a vocabulary task, a trivia task, and a spelling task. The older adults were also told about the Short Blessed Dementia Scale (Katzman et al. 1983). The participants were given the vocabulary task, the trivia task, the spelling task, and the awareness questionnaire, in that order. Afterward, the participants were debriefed about the study's purpose and given $7.00 for their time. The entire experiment took approximately 45 min.

In the trivia task each question was presented on a large index card for 15 s. The order of questions was varied randomly for each person. Participants were instructed to read the question out loud and give an answer within the allotted time. They were encouraged to guess if they did not know the answer.

In the spelling task the experimenter read aloud the spelling words one at a time. Although the task was not timed, participants were instructed to spell each word as quickly as they could. Participants were also told that some words may have more than one acceptable spelling and to spell the first word that came to mind.

	Results and Discussion

TOP Abstract Experiment 1 Results and Discussion Experiment 2 Methods Results and Discussion Experiment 3 Methods Results and Discussion General Discussion References

 
Table 1 presents the data for the homophone spelling task. Planned age comparisons using a Bonferroni-corrected alpha level of .025 (.05/2) were conducted for the primed and unprimed words. The comparison for the unprimed words was one tailed; as discussed earlier it was hypothesized that older adults would spell lower frequency homophones more often than would young adults in this unprimed condition.

In accordance with the two prior homophone studies (Davis et al. 1990; Rose et al. 1986), the older age group spelled more of the unprimed lower frequency homophones compared with the younger age group. The higher baseline of spelling lower frequency homophones resulted in a smaller priming effect for the older adults, albeit nonsignificantly different from that of the younger adults. An explanation for these results is that the smaller priming effect may not be caused by an age decline in implicit memory; the problem may lie in word frequencies that change because of experience with increased age. The lower frequency homophones may not be as low for older adults as they are for young adults. These results prompted a second study to determine if these differential word frequency effects on priming would generalize to a different priming task.

	Experiment 2

TOP Abstract Experiment 1 Results and Discussion Experiment 2 Methods Results and Discussion Experiment 3 Methods Results and Discussion General Discussion References

 
A word-fragment completion task was used in the second experiment. No previous study has investigated word frequency effects on word-fragment completion performance by older adults. One major addition to this experiment was the inclusion of high-frequency words. Consistent with past studies (Duchek and Neely 1989; Forster and Davis 1984; Kirsner et al. 1983; Macleod and Kampe 1996), younger adults were predicted to show a low-frequency word advantage. In other words, younger adults were expected to show larger priming effects for lower frequency words compared with higher frequency words. Given the homophone studies, older adults were predicted to have slightly reduced priming effects and differential word frequency effects than young adults on priming performance and completion of items.

This study was modeled after Experiment 1 and the study by Macleod and Kampe 1996. In their priming task a list of words was presented serially on a computer screen; then the word-fragment completion task was administered. They found that priming effects were larger for low-frequency words than for high-frequency words in young adults.

	Methods

TOP Abstract Experiment 1 Results and Discussion Experiment 2 Methods Results and Discussion Experiment 3 Methods Results and Discussion General Discussion References

 
Participants
A total of 96 participants (48 young and 48 older adults) were included in the study. None were participants from Experiment 1. Twenty-seven male and 21 female young adults were undergraduates randomly selected from a volunteer pool maintained by the psychology department. The mean age of these participants was 19 years (SD = 1.08; range = 18–22) with a mean of 13.73 years of education (SD = 1.03; range = 13–16). The older adult age group included 16 men and 32 women. They were randomly selected from a separate volunteer pool maintained by the Aging and Development Program in the psychology department. This group had a mean age of 71 years (SD = 5.58; range = 56–84). Their mean education was 14.67 years (SD = 2.44; range = 12–22). The older adults were administered the Short Blessed Dementia Scale (Katzman et al. 1983) to screen for major memory impairments. The mean score for this brief cognitive measure was 1.83 (SD = 2.42).

Both young and older age groups were given the vocabulary subtest of the WAIS–R so we could assess verbal ability (Wechsler 1981). Older adults were found to have significantly higher vocabulary scores compared with young adults, F(1,94) = 11.80, p = .001. The mean score of the young adult group was 52.04 (SD = 5.48), whereas the mean of the older age group was 56.50 (SD = 7.13). The correlation between vocabulary and priming effect, however, was not significant, r = .12, p = .24. Thus, verbal ability as measured by the WAIS–R was not included as a covariate in later analyses; analysis of covariance in which the covariate is uncorrelated with the dependent variable is ineffectual.

Materials
Word lists.
The high-frequency, low-frequency, and filler words were taken from Macleod and Kampe 1996. The entire list consisted of 144 items: 48 high-, 48 medium- (filler), and 48 low-frequency words. The high- and low-frequency words were selected from the Dahl 1979 norms. High-frequency words had frequencies of 20 or greater per million, whereas the low-frequency words had frequencies of 1 or fewer per million. The 48 high- and 48 low-frequency words were randomly assigned to one of four word sets. Each of these word sets, therefore, contained 12 high- and 12 low-frequency words. Each set was paired with every other set to produce a total of six combinations of word sets with 48 words in each combination. Each high- and low-frequency item occurred in both primed and unprimed conditions across participants.

Trivia task.
The participants were given a 48-item trivia questionnaire about various topics (e.g., history, art, entertainment). Although the questions were different from the questions in Experiment 1, the procedure was identical to that of the previous experiment. The six types of trivia questionnaires corresponded to the six combinations of word sets. The word sets presented in the trivia task were labeled the primed words. The word sets not presented during this task were labeled the unprimed words.

Digit Symbol subtest.
All participants were given the Digit Symbol subtest from the WAIS–R (Wechsler 1981). This task served two purposes. First, it served as a brief, nonverbal filler task between the priming phase (trivia task) and the implicit-memory phase (word-fragment completion task). The use of a filler task should decrease the level of awareness for the relatedness between the trivia and word-fragment completion tasks. It also provided a measure of psychomotor speed. Psychomotor speed was measured because the word-fragment completion task is a difficult, timed task. As expected, younger adults performed significantly better on Digit Symbol than older adults (M = 72.29, SD = 9.84, and M = 50.57, SD = 10.68, respectively), F(1,94) = 106.32, p < .001. The slower speed of older adults was a potential disadvantage in performing the word-fragment completion task, but it was not correlated with priming performance, r = .16, p > .10.

Word-fragment completion task.
In the test phase all participants were administered the word-fragment completion task. The 144 fragments were taken from Macleod and Kampe 1996. The word fragments were randomly assigned to one of two test forms. I constructed alternate forms of these test forms by reversing the item presentation to produce a total of four test forms. Each form consisted of 12 primed and 12 unprimed low-frequency items, 12 primed and 12 unprimed high-frequency items, and 48 filler items. All fragments were presented in lower case with an underscore representing each missing letter.

Participant awareness questionnaire.
Similar to the procedure for the questionnaire from Experiment 1, the participants rated their level of awareness about the relatedness between the trivia task and the word-fragment completion task on a 5-point scale. The young adults were significantly more aware of the relationship between these two tasks (M = 3.06, SD = 1.21) compared with the older adults (M = 1.52, SD = 1.07), F(1,94) = 43.67, p < .001. Young adults' awareness was significantly correlated with priming effect of low-frequency words (r = .43, p < .01) but not of high-frequency words (r = .04, p > .10). Older adults' awareness did not correlate with the priming effect for both low- and high-frequency words (r = .11, p > .10, and r = .07, p > .10, respectively). Because awareness did not correlate in all Age Group x Word Frequency conditions, awareness cannot be covaried out of the analyses because of parallelism problems.

Procedure
The participants came to the laboratory and were told that the study examined general cognitive functioning. They were told that the study involved various tasks such as a vocabulary task, a trivia task, and a word-fragment task. The participants were given the vocabulary task, the trivia task, the Digit Symbol task, the word-fragment completion task, and the awareness questionnaire, in that order. Afterward, the participants were debriefed about the purpose of the study and were given $7.00. The entire experiment took approximately 1 hr.

The procedure for administering the trivia task was identical to that for Experiment 1. In the Digit Symbol subtest the participants were told to use the key at the top of the page to write the matching symbol under each printed number. They were given 90 s to complete as many items as they could without skipping. In the word-fragment completion task, participants attempted to complete 96 word fragments that were presented on paper. The participants were instructed to complete each fragment item with letters to form a real word. They were told that they had no more than 10 s to complete each item. They were also given a sample item to ensure that they understood the task. The sample item was "ps__ch__l__gy." The correct answer for this item is "psychology." The participants were timed by the experimenter.

	Results and Discussion

TOP Abstract Experiment 1 Results and Discussion Experiment 2 Methods Results and Discussion Experiment 3 Methods Results and Discussion General Discussion References

 
Table 2 displays the means and standard deviations of the proportion of fragments completed in each priming and word-frequency condition. A 2 (age) x 2 (word frequency) x 2 (priming) mixed ANOVA was conducted on these proportions. Age was a between-subject independent variable; word frequency and priming were within-subject variables. As hypothesized, the Age x Word Frequency x Prime interaction was significant, F(1,94) = 16.46, p < .001. As expected, there was a significant interaction between priming and word frequency, F(1,94) = 20.97, p < .001, for the young adults. The priming effect was significant for low-frequency words, F(1,94) = 59.70, p < .001, but not for high-frequency words, F(1,94) = 1.90, p > .10. For older adults the interaction between word frequency and priming condition was not significant, F(1,94) = 1.34, p = .25. The older adults had a significant and similar priming effect for high- and low-frequency words, F(1,94) = 12.82, p = .001. To look at it another way, there was no difference in priming effect for high-frequency words between age groups, F(1,93) = 2.38, p > .10, but older adults had a significantly smaller priming effect for low-frequency words compared with young adults, F(1,94) = 28.38, p < .001. Similar to Experiment 1, priming performance for low-frequency words appeared to be different for older adults compared with young adults.

	Experiment 3

TOP Abstract Experiment 1 Results and Discussion Experiment 2 Methods Results and Discussion Experiment 3 Methods Results and Discussion General Discussion References

 
Although the older adults' priming effects for high- and low-frequency words were not statistically different in Experiment 2, the high-frequency priming effect was higher (M = 0.07) than the low-frequency priming effect (M = 0.04). Experiment 3 reexamined the effects of word frequency on priming performance in older adults. This study consisted of a new set of high-, middle-, and low-frequency items, a new set of questions for the trivia questionnaire, and a new set of word-fragment items. The Digit Symbol subtest and the awareness questionnaire were taken from Experiment 2. All the procedures were identical to Experiment 2.

	Methods

TOP Abstract Experiment 1 Results and Discussion Experiment 2 Methods Results and Discussion Experiment 3 Methods Results and Discussion General Discussion References

 
Participants
Forty-eight older adults (15 men, 33 women) were randomly selected from a participant pool at Washington University. No participants from Experiment 1 or Experiment 2 were included in this study. Their mean age was 72 years (SD = 4.74, range = 61–83) and their mean education was 15 years (SD = 2.23, range = 12–20). They were administered the Short Blessed Dementia Scale (Katzman et al. 1983) to screen for major memory impairments. The mean score for this brief cognitive measure was 1.98 (SD = 2.51). Their mean score on the WAIS–R vocabulary subtest (Wechsler 1981) was 54.06 (SD = 9.34).

Materials
Word lists.
The high-frequency, low-frequency, and filler words were selected from the Kucera and Francis norms (1967). The entire list consisted of 144 items: 48 high-, 48 medium- (filler), and 48 low-frequency words. High-frequency words had log frequencies of 45 to 53 per million, whereas the low-frequency words had log frequencies of 8 to 10 per million. The high- and low-frequency words were randomly assigned to one of four word sets. Each of these word sets, therefore, contained 12 high- and 12 low-frequency words. Each set was then paired with every other set to produce a total of six combinations of word sets. These six combinations corresponded to the six forms of the trivia task.

Trivia task.
The participants were given one of the six forms of the 48-item trivia task about various topics (e.g., history, art, entertainment). Although the questions were different from the questions in Experiment 2, the development of the questions and the procedure were identical to those in the previous experiment.

Digit Symbol subtest.
The Digit Symbol subtest from the WAIS–R (Wechsler 1981) was given to all participants as a filler task. The older adults' mean score was 48.79 (SD = 9.90).

Word-fragment completion task.
The generation of the word-fragment forms and the procedures of this task were identical to those in Experiment 2.

Participant awareness questionnaire.
Similar to the questionnaire from Experiment 2, the older adults' rating of their awareness about the relatedness between the trivia task and the word-fragment completion task indicated that they were unaware of the connection (M = 1.27, SD = 0.71).

	Results and Discussion

TOP Abstract Experiment 1 Results and Discussion Experiment 2 Methods Results and Discussion Experiment 3 Methods Results and Discussion General Discussion References

 
Table 3 displays the means and standard deviations of the proportion of fragments completed in each priming and word-frequency condition in the word-fragment task. A 2 (word frequency) x 2 (priming) within-subject ANOVA was conducted on those proportions. No interaction was found, F(1,47) = 0.03, p = .86. The effect of priming condition was significant, F(1,47) = 7.40, p < .01. The effect of word frequency was also significant, F(1,47) = 4.10, p < .05. The priming effects for high-frequency words and low-frequency words were identical and small (Ms = .05). Fig. 1 shows that these results replicated the findings from Experiment 2.

View larger version (16K): [in this window] [in a new window]   Figure 1. Mean priming effect of low and high frequency words for young and older adults. Exp = Experiment.

	General Discussion

TOP Abstract Experiment 1 Results and Discussion Experiment 2 Methods Results and Discussion Experiment 3 Methods Results and Discussion General Discussion References

Individual differences in verbal ability provide one possible explanation for the differential word-frequency patterns on priming between the age groups. Verbal ability as measured by vocabulary tasks is known to influence word frequency. A person with greater verbal ability (greater vocabulary performance) would have different frequencies for words than would someone with lesser verbal ability. In Experiment 1 older adults and younger adults had similar vocabulary performance as indicated by the WAIS–R vocabulary subtest (Wechsler 1981). Older adults' increased spelling of lower frequency homophones without priming, therefore, cannot be fully explained by verbal ability. Although the older age group had slightly greater vocabulary performance than the young age group in Experiment 2, vocabulary performance did not significantly correlate with priming performance. Individual differences in verbal ability as measured by a commonly used indicator, therefore, cannot account for the differential word frequency effects on priming between age groups.

An alternative explanation for these results is that the lower frequency words may be relatively higher in frequency for older adults than for young adults because of age and experience. With increased exposure to lower frequency homophones, the distinctions between lower and the higher frequency homophones would become less prominent. This would explain why older adults in Experiment 1 consistently spelled more lower frequency homophones than did young adults without priming. In Experiment 2, the increased experience with the lower frequency words may have resulted in reduced benefits of priming for older adults on low-frequency words. This would help explain why older adults had similar priming effects for low- and high-frequency words in Experiment 2 and 3. This would also explain why these priming effects are similar to the young adults' priming effects for the high-frequency words in Experiment 2.

Another explanation is that the differences in word frequencies between age groups may represent a cohort effect rather than an age effect. With time and each generation, the language changes. New words are added and the usage of old words may increase or decrease. This explanation is slightly different than the previous explanation based on increased exposure to lower frequency words with increased age. Because the effects of age and cohort cannot be distinguished in these studies, perhaps a more accurate approach is to combine the explanations as an age-cohort effect.

Another explanation focuses on the priming performance of the young adults. The awareness questionnaire indicated that the young adults were more aware than the older adults were of the relations between the priming task and the implicit-memory task. This indicates a possible explicit-memory contamination. Their much greater priming effect for low-frequency words compared with the high-frequency words is consistent with the expected memory performance in an explicit-memory recognition task (Macleod and Kampe 1996). In this task low-frequency words are recognized more often than high-frequency words. This would explain the difference in word frequency effects on priming performance between the two age groups. This, however, does not explain the lack of word frequency effect on the priming performance of older adults.

This study found both similar priming effects between age groups and age-related deficits in priming. Experiment 1 found similar priming effects for both age groups, whereas Experiment 2 indicated smaller but significant priming effects for older adults compared with those of younger adults. In comparison to the words used in Experiment 2 and 3, Experiment 1 involved relatively high-frequency words. From this perspective, the results in Experiment 1 correspond to those in Experiment 2 and 3, in which older adults had similar priming effects for high-frequency words compared with the young adults. Furthermore, in Experiment 3 the completed word-fragment proportions for primed and unprimed high- and low-frequency words by older adults are similar to young adults' completion of primed and unprimed high-frequency words in Experiment 2.

The word-frequency pattern for priming for older adults in Experiments 2 and 3 conflicts with the results of Balota and Ferraro 1996. As reported earlier, these researchers found that the low-frequency word advantage on priming performance was greater for older adults than for young adults, whereas in this study older adults did not show this advantage. Perhaps the different tasks and measurements of priming are responsible for the different word-frequency priming patterns of older adults between their study and this study. Their study measured reaction times to indicate priming effects, whereas the proportion of completed word fragments was measured in Experiments 2 and 3. Differential tasks could also explain the conflicting priming results in which the Balota and Ferraro study found no differences in priming effect between the young and older age groups, whereas the word-fragment completion task in Experiment 2 showed older adults to have smaller priming effects compared with the young adults. This explanation, however, is not convincing because Allen and colleagues (Allen, Madden, and Crozier 1991; Allen, Madden, Weber, and Groth 1993) have used reaction times as the dependent variable and found equivalent word frequency effects between age groups or even smaller word frequency effects for older adults. Thus, a more reasonable explanation is yet to be found.

This study found differential word frequency effects between young and older adults on priming performance. With higher frequency words, the young and older adults have similar priming effects. With lower frequency words, the age-related deficit in priming becomes greater. One approach to help control for age-cohort effects on word frequency would be to have both young and older adults learn a large number of nonwords. By controlling the number of exposures of these nonwords, researchers could ensure that young and older adults would have similar frequencies that could then be used to more accurately examine the effects of priming for high- and low-frequency "words" between age groups.

	Acknowledgments

 
This study was supported by Training Grant AG0030 from the National Institute on Aging. Special thanks go to Martha Storandt and David Balota for guidance and support in the conduct of this study and the preparation of this article.

Received for publication July 12, 1999. Accepted for publication March 8, 2001.

	References

TOP Abstract Experiment 1 Results and Discussion Experiment 2 Methods Results and Discussion Experiment 3 Methods Results and Discussion General Discussion References

 

• Allen P. A., Madden D. J., Crozier L. C., 1991. Adult age differences in letter-level and word-level processing. Psychology and Aging 6:261-271. [Medline]
• Allen P. A., Madden D. J., Weber T. A., Groth K. E., 1993. Influence of age and processing stage in visual word recognition. Psychology of Aging 8:274-282.
• Balota D., Ferraro F. R., 1996. Lexical, sublexical, and implicit memory processes in healthy young and healthy older adults and in individuals with dementia of the Alzheimer's type. Neuropsychology 10:82-95.
• Chiarello C., Hoyer W. J., 1988. Adult differences in implicit and explicit memory: Time course and encoding effects. Psychology and Aging 3:358-366. [Medline]
• Craik F. I. M., McDowd J. M., 1987. Age differences in recall and recognition. Journal of Experimental Psychology: Learning, Memory, and Cognition 13:474-479.
• Dahl H., 1979. Word frequency effects in associative and item recognition. Memory and Cognition 20:231-243.
• Davis H. P., Cohen A., Gandy M., Colombo P., VanDusseldorp G., Simolke N., et al. 1990. Lexical priming as a function of age. Behavioral Neuroscience 104:288-297. [Medline]
• Duchek J. M., Neely J. H., 1989. A dissociative word-frequency x levels-of-processing interaction in episodic recognition and lexical decision task. Memory and Cognition 17:148-162. [Medline]
• Forster K. I., Davis C., 1984. Repetition priming and frequency attenuation in lexical access. Journal of Experimental Psychology: Learning, Memory, and Cognition 10:680-698.
• Galbraith G. C., Taschman C. S., 1969. Homophone units: A normative and methodological investigation of the strength of components elements. Journal of Verbal Learning and Verbal Behavior 8:737-744.
• Howard D. V., 1988. Implicit and explicit assessment of cognitive aging. Howe M. L., Brainerd C. J., , ed.Cognitive development in adulthood: Progress in cognitive development research 3-37. Springer-Verlag, New York.
• Howard D. V., Howard J. H., Jr. 1989. Age differences in learning serial patterns: Direct versus indirect measures. Psychology and Aging 4:357-364. [Medline]
• Hultsch D. F., Masson M. E. J., Small B. J., 1991. Adult age differences in direct and indirect tests of memory. Journal of Gerontology: Psychological Sciences 46:P22-P30.
• Java R. I., 1992. Priming and aging: Evidence of preserved memory function in an anagram solution task. American Journal of Psychology 105:541-548.
• Java R. I., Gardiner J. M., 1991. Priming and aging: Further evidence of preserved memory function. American Journal of Psychology 104:89-100.
• Katzman R., Brown T., Fuld P., Peck A., Schechter R., Schimmel H., 1983. Validation of a short orientation-memory-concentration test of cognitive impairment. American Journal of Psychiatry 140:734-739. [Abstract/Free Full Text]
• Kirsner K., Milech D., Standen P., 1983. Common and modality-specific processes in the mental lexicon. Memory and Cognition 11:621-630. [Medline]
• Kucera H., Francis W. N., 1967. Computational analysis of present-day American English Brown University Press, Providence, RI.
• Light L. L., Albertson S. A., 1989. Direct and indirect tests of memory for category exemplars in young and old adults. Psychology and Aging 4:487-492. [Medline]
• Light L. L., Singh A., 1987. Implicit and explicit memory in young and older adults. Journal of Experimental Psychology: Learning, Memory, and Cognition 13:531-541. [Medline]
• Macleod C. M., Kampe K. E., 1996. Word frequency effects on recall, recognition and word fragment completion tests. Journal of Experimental Psychology: Learning, Memory, and Cognition 22:132-142. [Medline]
• McEvoy C. L., Nelson D. L., Holley P. E., Stienicki G. S., 1992. Implicit processing in the cued recall of young and old adults. Psychology and Aging 7:401-408. [Medline]
• Park D. C., Shaw R. J., 1992. Effect of environmental support on implicit and explicit memory in younger and older adults. Psychology and Aging 7:632-642. [Medline]
• Poon L. W., 1985. Differences in human memory with aging: Nature, causes, and clinical implications. Birren J. E., Schaie K. W., , ed.Handbook of the psychology of aging 2nd ed. 427-462. Academic Press, New York.
• Rose T. L., Yesavage J. A., Hill R. D., Bower G. H., 1986. Priming effects and recognition memory in young and elderly adults. Experimental Aging Research 12:31-37. [Medline]
• Schacter D. L., 1987. Implicit memory: History and current status. Journal of Experimental Psychology 13:501-518.
• Terrell C. D., Meadows B., 1985. A list of English homophones. The Quarterly Journal of Experimental Psychology 37A:627-631.
• Wechsler D., 1981. Wechsler Adult Intelligence Scale–Revised Psychological Corp, San Antonio.

This Article Abstract Full Text (PDF) Services Download to citation manager PubMed PubMed Citation