| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
School of Psychology, Georgia Institute of Technology, Atlanta.
Address correspondence to Deborah K. Eakin, Mississippi State University, Department of Psychology, PO Box 6161, Mississippi State, MS 39762. E-mail: deakin{at}psychology.msstate.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
). The present study examined age-related effects on memory under conditions of implicit interference. Recall and recognition performance of both younger and older adults was evaluated for small- versus large-set-size cues under two contexts. Comparable cue-set-size effects were obtained for both age groups under extralist cueing, but they were eliminated only for younger adults under intralist cueing. Older adults were not able to use the context to effectively eliminate implicit interference from associates of the cue as did younger adults, perhaps because of an inhibition deficit. Both groups had equivalent metamemory accuracy and sensitivity, indicating that the monitoring of learning prior to a test reflected the effects of implicit interference and is not impaired by aging.
CUED recall is influenced by the number of semantic associates of the cue. Targets linked to cues with a small number of associates, or small set size, are recalled better than targets linked to cues with a large set size (e.g., D. L. Nelson & McEvoy, 1979; Schreiber & Nelson, 1998). This finding is known as the cue-set-size effect. Cue-set-size effects can be understood as a consequence of implicit interference created by semantic activation. According to the Processing Implicit and Explicit Representations model (PIER2; D. L. Nelson, McKinney, Gee, & Janczura, 1998), lower recall occurs for items with high numbers of associates as a result of implicit interference created by competing semantic activations when the cue is presented. The PIER2 model states that processing a cue implicitly activates its semantic associates. During recall, a target can be retrieved by means of an explicit association created by intentional learning or by the activation of associates of the cue. In the latter case, the probability of sampling the correct target is a function of the number of implicitly activated associates of the cue. The probability of sampling a target linked to a small-set-size cue is greater than one linked to a large one, because small-set-size cues have fewer associates from which to sample (see Figure 1). In effect, the implicit interference from competing associates in the sampled set is greater when the set is large, resulting in lower recall under conditions of greater implicit interference.
|
found that older adults produce greater cue-set-size effects in memory than younger adults do in an extralist cueing procedure (i.e., when a studied list of targets are later cued by a semantic associate that was never studied with the target). The finding is consistent with an inhibition-deficit account (e.g., Hasher & Zacks, 1988; Zacks & Hasher, 1994; see also Holley & McEvoy, 1996), which argues that older adults have more difficulty recalling information under conditions of interference because of an age-related decline in inhibitory processes in the cognitive system.
The magnitude of cue-set-size effects is not the only way to examine the impact of implicit interference. Another method is to examine how cue-set-size effects can be eliminated. This approach has already been used to evaluate target-set-size effects. According to PIER2, during intralist cueing (when the cue and target are studied together; the technique commonly used in paired-associate learning), only associates related to both the cue and target are sampled. Hence, activation of semantic associates not shared between the cue and target falls away. In effect, the intralist sampling set is the intersection of both sampling sets (see the bottom of Figure 1), resulting in equal probability of recall for small- and large-set-size targets. The elimination of set-size effects has been obtained only when the set size of the target has been varied (D. L. Nelson & McEvoy, 1979; D. L. Nelson, McEvoy & Schreiber, 1990). McEvoy, Nelson, Holley, and Stelnicki (1992) reported that younger and older adults showed equivalent elimination of target-set-size effects, but Holley-Edwards (1999) found that this was true only for older adults with higher comprehension scores.
It is not clear whether cue-set-size effects will behave similarly. Whereas target-set-size effects are measures of indirectly activated associates of the target, which is not present at test, cue-set-size effects demonstrate the impact of items directly activated by the cue, which is present at test. This difference may have implications for observing age differences in interference effects. Older adults' inhibitory deficit may make it difficult for them to eliminate irrelevant associates of the cue from the sampling set when the cue is explicitly present at test. In the present study we sought to determine whether intralist cueing eliminates cue-set-size effects for younger but not for older adults, a finding that would provide support for the implicit interference account.
We also examined the sensitivity and accuracy of metacognitive judgments when cue set size was manipulated. Schreiber (1998) and Schreiber and Nelson (1998) showed that predictions of knowing (POKs) are more positive for small- than for large-set-size cues and targets, similar to the effects of set size on recall.1 In this experiment we evaluated whether POKs of older adults follow the same pattern as that of younger adults, suggesting an equivalent impact of implicit interference. We also wanted to determine whether cue-set-size effects on POKs were eliminated by the intralist cueing procedure for both age groups.
Another one of our goals was to evaluate implicit interference effects on the resolution, or relative accuracy, of POKs. Resolution is defined as accuracy in discriminating between items that will be or will not be recalled. We computed intraindividual Goodman–Kruskal gamma (
) correlations between each person's POKs and recall across all items (T. O. Nelson, 1984). Older and younger adults typically show equivalent resolution of metacognitive judgments, including POKs and delayed judgments of learning (JOLs; see, e.g., Connor, Dunlosky, & Hertzog, 1997; Eakin, 2005b; Hertzog, Kidder, Powell-Moman, & Dunlosky, 2002; but see Souchay, Isingrini, & Espagnet, 2000). Given an accessibility view of metacognitive judgments (e.g., Koriat, 1997), implicit interference should change accessibility to target information both when the POK is made and at recall, leaving resolution unaffected. However, age differences in implicit interference effects under cue-set-size manipulations could produce differential resolution for POKs of younger and older adults if the interference effects are both greater and more variable for older adults, reducing the similarity of interference effects occurring at the time of the POK and at the time of the memory test.
Finally, we evaluated cue-set-size effects on recognition memory by administering a five-alternative forced-choice recognition test. We did this because recognition tests can be conceptualized as a source of retrieval support (e.g., Bäckman, 1989; Craik & Jennings, 1992). As such, cue-set-size effects in recognition memory may differ from the effects in cued recall because presentation of the target as an option on the recognition test circumvents implicit interference operating during cued retrieval search. In effect, presentation of the cue with a set of five associates as alternatives, one of which is the target, may effectively reduce the associative set of all items to a small set of five. If this is the case, the biggest benefit to memory performance between recall and recognition should be observed in the large-set-size condition.
We also examined metamemory accuracy by correlating POKs with recognition memory performance. We were interested in determining whether POKs were more weakly correlated with recognition than recall. Weaver and Kelemen (2003) found that metamemory accuracy was worse for recognition than for recall, even though cued recognition performance exceeded cued recall. They attributed this effect to the particular diagnosticity of delayed JOLs (in which cues are presented to elicit the judgments) for successful cued retrieval as opposed to a recognition test. If POKs are more correlated with recall than with recognition, the findings could be taken as evidence that POKs and recall are similarly affected by interference effects, and that those effects are removed from the recognition test by retrieval support.
|
METHODS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Materials
Stimulus materials
We used the University of South Florida Word Association Norms (D. L. Nelson, McEvoy, & Schreiber, 1990) to create two separate lists of 44 related cue and target word pairs.2 Half of the word pairs contained small-set-size cues (5 to 9 semantic associates, M = 6.79, SD =.16), and half had large-set-size cues (16 to 24, M = 19.75, SD =.10). We equated forward association strength (M =.12, SD =.03) and backward association strength (M =.03, SD =.01) across cue set size and list. We also equated target set size, printed word frequency (Ku
era & Francis, 1967), concreteness, and connectivity (D. L. Nelson et al., 1998). In addition, we used the ListChecker Pro 1.2 program (Eakin, Schreiber, & Nelson, 2005)3 to ensure that each cue was semantically related only to its intended target and not to any other target or cue on the list.
A five-alternative forced-choice recognition test presented the 44 cues followed by their intended target and four semantically associated foils. For example, the cue KITE was followed in a random order by the target (STRING) and four associates of the cue (WIND, FLY, SKY, and TAIL). The intended target had the strongest forward strength association with the cue only about 15% of the time.
Additional materials
We used the Mini-Mental State Exam (Cockrell & Folstein, 1988) to screen for participants with impaired cognitive functioning; we excluded 8 older adults as a result of their low performance (Mini-Mental State Exam score < 27). We administered a demographic questionnaire, including questions about age, education, overall health, and ethnicity. Older adults reported significantly better overall health, F(1, 154) = 6.91, p =.009, 
p2 =.04, than the younger adults did, and they outperformed younger adults on the Shipley (1940) vocabulary test, F(1, 154) = 62.15, p <.001, p2 =.29. The two age groups did not differ significantly on any other demographic dimension.
Procedure
We programmed the experimental tasks by using E-prime version 1.1 (Psychology Software Tools, Inc.) and executed them on standard PCs. The computerized experiment consisted of three distinct phases: (a) the encoding phase, (b) the POK prediction phase, and (c) the test phase.4 A six-item practice phase preceded the encoding and POK phases.
Intralist cueing procedure
During the encoding phase, instructors presented participants with cue–target word pairs (e.g., KITE–STRING) for 8 seconds, during which time they told the participants to attempt to form an interactive image of the two words. Participants pressed the space bar to indicate when an image was formed. After studying each pair, participants were instructed to rate the image they had formed as vivid (clear with lots of detail) or neutral (unclear and vague), or to press the "enter" key if they were unable to form an image. The 44 list pairs were presented in random order.
Next, participants made POKs based on their sense of whether they would be able to remember the target, given the cue, on a subsequent recall test. Time to respond was not limited, but participants were told to make their predictions as quickly as possible by using a percentage confidence rating scale of 0–100, with a rating of 0 indicating no chance that they would be able to recall the target and a rating of 100 indicating 100% certainty that they would be able to recall it. Instructors told participants to use the entire range of the scale for predictions falling in between 0 and 100. Cues were presented in random order until POKs had been collected for all 44 cues.
During the test phase, we again had cues presented in random order, and target recall was solicited. Participants typed their responses on the keyboard. Time to respond was not limited, and instructors encouraged participants to try hard to recall the target. If they were unable to do so, they were encouraged to guess or to enter "NEXT" if they were unable to guess.
Extralist cueing procedure
All instructions and procedures for the extralist cueing procedure were identical to the intralist cueing procedure with one exception. During the encoding phase, only the target from each cue–target pair was presented. Participants were instructed to use imagery to encode the target and to rate the vividness of that image following the same procedure described for the intralist cueing procedure. POKs and target recall were obtained on the basis of the same related cue used for intralist cueing, and participants were told to refer to the target they had studied that was related to the cue presented.
Following the computerized portion of the experiment, instructors gave the five-alternative forced-choice recognition test as a self-paced paper-and-pencil task.
|
RESULTS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
p2 = 2.54, cueing procedure, F(1, 156 ) = 176.50, p <.001, p2 =.53, and cue set size, F(1, 156 ) = 110.55, p <.001, p2 =.42. Cued recall was significantly higher for younger adults than for older adults (M = 0.66, SEM = 0.01 vs M = 0.51, SEM = 0.01), higher in the intralist condition than in the extralist condition (M = 0.72, SEM = 0.01 vs M = 0.45, SEM = 0.01), and higher for cues with a small set size (M = 0.63, SEM = 0.01) versus a large set size (M = 0.54, SEM = 0.01). The Age Group x Cue Set Size interaction was not significant, F(1, 156 ) = 1.36, p >.05, p2 =.01. The Cueing Procedure x Cue Set Size interaction was significant, F(1, 156) = 33.81, p <.001, p2 =.18, reflecting the fact that set-size effects in cued recall were obtained only in the extralist cueing procedure, as expected. Most importantly, the three-way Cue Set Size x Cueing Procedure x Age interaction was significant, F(1, 156) = 12.66, p <.001, p2 =.08. Cue-set-size effects were eliminated under intralist cueing for younger adults (recall difference of.20 for extralist cues vs.02 with intralist cues), but not for older adults (mean difference of.11 for extralist cues vs.07 for intralist cues). For younger adults, probability of recall for large-set-size cues increased by.41 from extralist to intralist cueing. For older adults, the recall increase was only.25.
|
p2 =.05. Older adults tended to give a vivid rating to proportionally fewer items in the intralist cueing procedure for large-set-size (M = 0.71, SEM = 0.03) than for small-set-size (M = 0.78, SEM = 0.29) cues. However, vividness ratings were not correlated with recall, as demonstrated by a Pearson product–moment correlation between vividness rating and probability of recall. Neither the overall correlation, r(78) = –.09, nor the correlations for either cueing procedure, extralist, r(78) =.03, intralist, r(78) = –.19, were significant. The correlations for the two age groups, r(78) = –.04 and r(78) =.03 for younger and older adults, respectively, were also not significant. Finally, controlling on vividness as a covariate did not eliminate the critical interaction tests for recall. The Cue Set Size x Covariate interaction was not significant (F < 1), and there was still a significant three-way Cue Set Size x Cueing Procedure x Age Group interaction, F(1, 155) = 12.90, p <.05, p2 =.08, indicating that cue-set-size effects were obtained for both age groups in extralist cueing and eliminated in intralist cueing only for younger adults.
Metamemory Sensitivity
Mean POKs for each of the experimental conditions are reported in Table 2. Overall, POKs tracked the experimental effects on cued recall. A main effect of cue set size, F(1, 156 ) = 92.25, p <.001, p2 =.37, showed that POKs were more positive for small-set-size (M = 74.38, SEM = 2.22) than for large-set-size (M = 67.06, SD = 2.67) cues. In addition, the reliable main effect of cueing procedure, F(1, 156 ) = 59.94, p <.001, p2 =.27, indicated that POKs were more positive for intralist (M = 79.67, SEM = 2.15) than for extralist (M = 61.78, SEM = 2.77) cues. There were reliable age differences, F(1, 156 ) = 5.74, p <.05, p2 =.04, with younger adults' POKs being significantly higher (M = 73.51, SEM = 2.58) than those of older adults (M = 67.93, SEM = 3.05). The reliable Cue Set Size x Cueing Procedure interaction, F(1, 156 ) = 5.74, p <.05, p2 =.12, revealed that POKs for extralist cues were more positive for small- than for large-set-size cues (difference, M = 10.88), whereas this difference was virtually eliminated for intralist cues (difference, M = 3.76). However, the three-way Cue Set Size x Cueing Procedure x Age Group interaction was also significant, F(1, 156 ) = 6.83, p =.01, p2 =.04. Cue-set-size effects in metamemory were obtained for extralist cueing for both age groups, but eliminated in the intralist cueing procedure only for younger adults. Mean POKs for small- and large-set-size cues differed by 12.30 for young adults in the extralist cueing procedure but only 1.20 for the intralist cueing procedure. For older adults, however, the corresponding mean differences in POK magnitude were 9.45 and 6.32, respectively. Thus, POK sensitivity varied with cue set size in the same way as did probability of recall.
|
correlations between POKs and recall (see Table 3). The mean for younger (M = 0.75, SEM = 0.03) and older (M = 0.73, SEM = 0.03) adults did not differ, F(1, 152) = 0.23, p >.05, p2 =.001. None of the other main effects or interactions was significant. Thus, POKs discriminated recalled and unrecalled items well for both younger and older adults and were unaffected by the experimental manipulations.
|
p2 =.002, with equivalent recognition for small-set-size (M = 0.81, SEM = 0.01) and large-set-size (M = 0.80, SEM = 0.01) cues. Reliable main effects emerged for age, F(1, 156) = 54.30, p <.001, p2 =.26, and cueing procedure, F(1, 156) = 8.36, p =.004, p2 =.05. Overall, recognition was better for younger (M = 0.89, SEM = 0.02) than for older (M = 0.71, SEM = 0.02) adults, and better for intralist (M = 0.84, SEM = 0.02) than for extralist (M = 0.77, SEM = 0.02) cues. The Age x Cue Set Size interaction was significant, F(1, 156) = 4.28, p <.05, p2 =.02. However, as shown, the actual mean differences were extremely small, and for practical purposes the two conditions were equal. The Cueing Procedure x Cue Set Size interaction was not significant. No cue-set-size effects were obtained for recognition, even in the extralist cueing procedure. Apparently, having access to the target at recognition ameliorates the effect of implicit interference. None of the other interactions were significant.
|
Four younger adults had perfect recall, so we excluded them from this analysis. Probability of recognition for unrecalled items (see the lower half of Table 4) was lower than that of recognition of all items, and the main effects of age and cueing procedure followed the same pattern as that for recognition of all items. Recognition was significantly lower for older (M = 0.53, SEM = 0.02) than for younger (M = 0.74, SEM = 0.02) adults, F(1, 152) = 42.08, p <.001, p2 =.22, and recognition was higher for intralist (M = 0.70, SEM = 0.02) than for extralist (M = 0.57, SEM = 0.02) cues, F(1, 152) = 16.10, p <.001, p2 =.10. Unlike recognition for all items, the main effect of cue set size was significant when unrecalled items were examined. Recognition was better for large-set-size (M = 0.70, SEM = 0.02) than for small-set-size (M = 0.57, SEM = 0.02) cues, F(1, 152) = 41.51, p <.001, p2 =.21. There was also a Cue Set Size x Cueing Procedure interaction such that, for both age groups, there was no difference between small-set-size (M = 0.70, SEM = 0.03) and large-set-size (M = 0.73, SEM = 0.02) cues in recognition for extralist cueing, suggesting that implicit interference was eliminated when the target was available at test, F(1, 152) = 12.83, p <.001, p2 =.08. For intralist cueing, cue-set-size effects were reversed such that large-set-size cues (M = 0.67, SEM = 0.02) resulted in better recognition than did small-set-size cues (M = 0.47, SEM = 0.03). Apparently, implicit interference was reduced as a result of intralist encoding, but availability of the target at recognition also provided an additional benefit for large-set-size cues. The same pattern of results were obtained for both age groups; the three-way interaction was not significant, F < 1.
POK correlations with recognition of all items
Although POKs inquired about future recall, we wanted to determine to what degree POKs were correlated with recognition memory outcomes. POK accuracy was lower for recognition than for recall. The mean for younger (M = 0.51, SEM = 0.05) and older (M = 0.45, SEM = 0.04) adults did not differ reliably, F(1, 116) =.74, p >.05, p2 =.05 (see Table 5). Contrary to correlations with recall accuracy, POK correlations with recognition accuracy varied with both cue set size and cueing procedure. The mean was higher for small-set-size (M = 0.56, SEM = 0.04) than for large-set-size (M = 0.40, SEM = 0.05) cues, F(1, 116 ) = 8.90, p <.05 , p2 =.07, and higher for the intralist (M = 0.61, SEM = 0.05) than for the extralist (M = 0.34, SEM = 0.05) cues, F(1, 116 ) = 16.17, p <.001, p2 =.12. The interaction between these two factors was also significant, F(1, 116 ) = 4.60, p <.05, p2 =.04, because mean varied with set size only in the intralist cueing procedure.
|
was greater for small-set-size (M = 0.36, SEM = 0.09) than for large-set-size (M = 0.003, SEM = 0.10) cues only in the intralist cueing procedure, F(1, 96) = 6.31, p =.01, p2 =.06. Most striking was the finding that people of both age groups were most accurate in the small-cue, intralist condition. Apparently something other than retrievability was influencing their predictions, and this influence was diagnostic of recognition. |
DISCUSSION |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
), this finding indicates that implicitly activated semantic associates of the cue resulted in more or less implicit interference at retrieval, depending on the number of associates of the cue. These effects were eliminated by standard paired-associate cued recall (intralist cueing) for younger adults. However, the study identified intriguing age differences in these effects. Older adults showed cue-set-size effects with extralist cueing similar to those of younger adults, a finding that could be taken as evidence against an inhibition deficit (e.g., Hasher & Zacks, 1988; Zacks & Hasher, 1994) because older adults did not have more difficulty with implicit interference from a larger sampling set than did younger adults. However, the findings regarding age differences in intralist cueing tell a different story. When cues and targets were studied together, cue-set-size effects were eliminated for younger but not for older adults. According to PIER2, when cue and target are studied together, the sampling set of activated associates should have been reduced to include only associates of both the cue and the target, thereby equating the amount of implicit interference at retrieval for large- and small-set-size cues. Perhaps an age-related deficit in inhibitory mechanisms contributed to older adults' inability to eliminate irrelevant associates from the sampling set, resulting in some persistence of the cue-set-size effects for older adults' intralist cued recall.
The lower probability of recall for older adults with intralist cueing is consistent with prior findings of an age-related associative learning deficit (Dunlosky, Hertzog, & Powell-Moman, 2005; Kausler, 1994; Naveh-Benjamin, Guez, & Shulman, 2004). For intralist cueing, we instructed participants to produce a mediator (an interactive image) in order to minimize variability that was due to differences in imagery mediator production (Dunlosky & Hertzog, 1998). Hence, cue presentation at recall affords two possible routes to the target. Either the target can be activated as one element of a sampling set of semantic associates of the cue, or the cue can prompt explicit retrieval of the mediator formed at encoding. Older adults typically are able to use interactive imagery for encoding under instructions (Dunlosky & Hertzog, 1998, 2001), and they were able to do so in the present experiment, judging from the results with the imagery ratings. However, as demonstrated by Dunlosky and colleagues, older adults might have had more difficulty retrieving the interactive image mediator at test in order to facilitate recall (Dunlosky et al.), resulting in the older adults' residual set-size effects under intralist cueing. Consistent with the findings of Dunlosky and colleagues, small set sizes did not lead to reports of better quality mediators by older adults as indicated by an equal proportion of items receiving ratings of "vivid." Hence, it is more plausible to interpret the effect as reflecting residual interference effects at test. When mediator retrieval fails, participants can recover the target by the same kind of indirect activation as in extralist cueing (e.g., McEvoy, Nelson, & Komatsu, 1999). Thus, a reasonable explanation of the residual set-size effect is that older adults are more susceptible to the implicit interference from large-set-size cues when mediator access is unsuccessful.
Both younger and older adults' POKs were sensitive to the experimental manipulations of cue set size and cueing procedure, tracking their effects on recall. Apparently, younger and older adults' metacognitive monitoring is equally sensitive to implicit interference effects. Furthermore, both younger and older adults' POKs predicted future item recall with equal accuracy, with mean unaffected by cueing procedure or cue set size. The finding that metamemory tracked memory in terms of implicit interference is inconsistent with earlier research on metamemory during explicit interference (e.g., Eakin, 2005a; Metcalfe, Schwartz, & Joaquim, 1993). Differences in the type of metacognitive judgment collected (e.g., Metcalfe et al.) and the source of interference could have contributed to discrepancies. Perhaps implicit interference influences are consistent during prediction and retrieval, resulting in increased accuracy as compared with explicit influences, which may be less consistent.
Finally, cue-set-size effects were not found for recognition tests for either cueing procedure, when the target word was one of several alternatives in a forced-choice test. One possible explanation is that presentation of the cue in conjunction with five members of its set effectively reduces all cues to a small set status. That the largest difference between recall and recognition was for large-set-size cues provides some support for this idea that the recognition test represents an explicit example of the implicit set reduction resulting from intralist cueing. In both cases, a reduction of the set size results in better retrievability of the target. A small, but significant, interaction showed that older adults were better at recognition for large- than for small-set-size cues. It may be that the older adults benefited even more from the reduction of large cue sets provided by the recognition test.
These findings also support the accessibility view of metacognitive judgments (Koriat, 1997; cf. Schreiber & Nelson, 1998; also see Maki, 1999). By this account, implicit interference influences the accessibility of information available to make POKs, impacts cued recall, and is neutralized by the recognition test. People were generally accurate at predicting recognition except in the absence of retrievability of the target. For unrecalled items, predictions were completely undiagnostic of recognition performance, except in the case of small-set-size cues in the intralist cueing procedure. The differences were striking. For both age groups, metamemory accuracy increased for unrecalled, small-set-size cues from = 0 in the extralist cueing procedure to around =.35 for intralist cueing. Accuracy may have increased so dramatically for unrecalled, small-set-size cues in the intralist cueing procedure because interference effects in that condition during the prediction phase may have paralleled the context at recognition better than any of the other conditions. The result was more accurate predictions, which could be explained by a kind of process-based transfer-appropriate monitoring (Weaver & Kelemen, 2003).
The findings from the present experiment contribute support to a growing literature showing no major age differences in metamemory accuracy (e.g., Connor et al., 1997; Eakin, 2005b; Hertzog & Dixon, 1994; Lovelace & Marsh, 1985). Although older adults did not perform as well on either memory task, their performance was accurately reflected in their metamemory predictions. Such findings support the argument that preserved metacognition can be a basis for memory-compensation strategies for older adults (e.g., Dunlosky, Kubat-Silman, & Hertzog, 2003).
Along with the present study, several other studies (e.g., Mazzoni & Nelson, 1995) have provided evidence that memory and metamemory are impacted differentially by experimental manipulations. In particular, dissociations have been obtained between memory and metamemory under conditions of explicit interference (Chandler, 1994; Eakin, 2005a; Metcalfe et al., 1993). In contrast, metamemory tracked memory in this implicit interference context. The distinction between explicit and implicit interference effects may prove important in understanding metacognitive monitoring in general, memory performance of older adults in particular, and the relationship between the two.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
). 
2. Although the norms were collected for younger adults, previous studies have demonstrated stable response sets across the adult age span (e.g., Bowles, Williams, & Poon, 1983; but see Burke & Peters, 1986 for a caveat), especially for older adults who perform as well or better than younger adults on vocabulary tests, such as the Shipley (Burke & Peters, 1986; Lovelace & Cooley, 1982). In addition, the South Florida norms specifically have been used in previous studies with older adults (e.g., Holley & McEvoy, 1996).
3. The ListChecker Pro 1.2 program uses the South Florida word association norms (D. L. Nelson et al., 1998) to make a variety of comparisons that are helpful in list construction, such as identification of direct and indirect within-list and between-list associations. For more information on ListChecker Pro 1.2, contact D. Eakin.
4. Feeling-of-knowing judgments were also collected after each recall attempt during the recall phase. However, because of space constraints, the findings regarding these judgments are not reported here.
Decision Editor: Thomas M. Hess, PhD
Received for publication August 3, 2005. Accepted for publication April 17, 2006.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
era, H., Francis, W. N. (1967). Computational analysis of present-day American English. Providence, RI: Brown University Press.This article has been cited by other articles:
 |
|
 |
|
  D. Pushkar, J. Chaikelson, M. Conway, J. Etezadi, C. Giannopoulus, K. Li, and C. Wrosch Testing Continuity and Activity Variables as Predictors of Positive and Negative Affect in Retirement J Gerontol B Psychol Sci Soc Sci, October 29, 2009; (2009) gbp079v1. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
