Eakin, D.K., & Hertzog, C. (2012). Age invariance in feeling of knowing during implicit interference effects. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences,
10.1093/geronb/gbr150
Sciences, 67(5), 555–562, doi:10.1093/geronb/gbr150. Advance Access published on February 24, 2012
Age Invariance in Feeling of Knowing During Implicit
Interference Effects
Deborah K. Eakin1 and Christopher Hertzog2
1Department
2School
of Psychology, Mississippi State University.
of Psychology, Georgia Institute of Technology, Atlanta.
Objectives. Prior research found age invariance in accuracy of delayed judgments of learning accuracy (Eakin, D. K.,
& Hertzog, C. [2006]. Release from implicit interference in memory and metamemory: Older adults know that they can’t
let go. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 61, 340–347). We tested
whether aging affects accuracy of feeling of knowing (FOK) predictions under implicit interference. Discrepancies in the
literature suggest that FOKs sometimes are and sometimes are not affected by aging. In addition, because the effects of
implicit interference are different on recognition than on recall, older adults may have difficulty ignoring the impact of
interference on recall in order to accurately predict the lack of interference effects on recognition.
Method. Younger and older adults studied cue-target pairs and cue set size varied. After a cued recall test, they made
FOKs about future recognition of the target given the cue and then took a recognition test.
Results. Neither younger nor older adults were able to predict recognition of unrecalled items. FOKs were more
correlated with recall than with recognition for both age groups. Although both recall and recognition varied with age, no
age differences were obtained in FOK accuracy.
Discussion. FOK accuracy was not impaired with age, even when memory was. FOKs of both younger and older
adults reflected implicit interference effects in recall, not recognition.
Key Words: Associate set size—Feeling of knowing—Interference—Metamemory.
F
EELINGS of knowing (FOKs) are defined as assessments that information is available in memory even
when it has not been retrieved. FOKs can be useful in regulating successful remembering; for instance, by informing
decisions about whether to initiate a memory search or
whether to continue or terminate an unsuccessful retrieval
search (e.g., Nelson & Narens, 1990; Singer & Tiede, 2008).
FOK judgments ask individuals to predict future recognition
success after a failed recall attempt (Hart, 1965). Cue-target
paired associates (e.g., TICK-SPOON) are typically studied, and recall is tested by providing the cue word (e.g.,
TICK) and requiring recall of its newly learned associate
(e.g., SPOON). The question is whether individuals manifest higher FOKs for items they successfully recognize than
for items they do not. Such FOK accuracy (termed resolution) depends on multiple factors (e.g., Nelson, Gerler, &
Narens, 1984). Individuals apparently base FOKs on multiple sources of information, including the familiarity of the
cue and accessibility to information about the encoding
experience that may be diagnostic of target availability
(Dunlosky & Metcalfe, 2009; Koriat & Levy-Sadot, 2001;
Metcalfe, Schwartz, & Joaquim, 1993). People often have
better than chance accuracy in forecasting whether they
will recognize unrecalled information (Blake, 1973;
Costermans, Lories, & Ansay, 1992; Leonesio & Nelson,
1990; Schacter, 1983).
There is consensus that, relative to younger adults, older
adults maintain equivalent FOK resolution for semantic
information, such as world knowledge (e.g., Anooshian,
Mammarella, & Hertel, 1989; Butterfield, Nelson, & Peck,
1988; Marquie’ & Huet, 2000; Souchay, Moulin, Clarys,
Taconnat, & Isingrini, 2007). In contrast, several studies
have found age-related differences in FOK accuracy in episodic memory tasks (e.g., Souchay, Isingrini, & Espagnet,
2000; Souchay et al., 2007). This finding is not universal,
however. Other studies have found age equivalence in FOK
resolution for episodic memory of unrelated paired associates (Hertzog, Dunlosky, & Sinclair, 2010; MacLaverty &
Hertzog, 2009).
Recent research on a different metacognitive judgment, the
judgment of learning (JOL), has indicated no age differences
in resolution for episodic memory (Hertzog, Sinclair, &
Dunlosky, 2010). We previously showed that implicit
memory interference had similar effects on delayed JOLs of
younger and older adults (Eakin & Hertzog, 2006). This
study evaluated whether aging affected episodic FOK accuracy under conditions that generated implicit interference.
A demonstration that implicit interference has differential
impact on FOKs for older and younger adults would support the idea that older adults’ FOKs are based on different
cues than those of younger adults. Conversely, a finding
of age-related invariance in metamemory for implicit
©
©The
TheAuthor
Author 2012.
2012. Published
Published by
by Oxford
Oxford University
University Press
Press on
on behalf
behalf of
ofThe
The Gerontological
Gerontological Society
Society of
ofAmerica.
America.
All
All rights
rights reserved.
reserved. For
For permissions,
permissions, please
please e-mail:
e-mail: [email protected].
[email protected].
Received
ReceivedAugust
August 2,
2, 2010;
2010;Accepted
Accepted December
December 7,
7, 2011
2011
Decision
Decision Editor:
Editor: Robert
RobertWest,
West, PhD
PhD
5551
2556
EAKIN AND HERTZOG
interference would support the argument that FOKs for
adults of different ages are constructed in a similar manner.
Implicit Interference Effects on Memory
Examining implicit interference is relevant to episodic
FOKs because interference influences accessibility to clues
about information held in memory that are an important
influence on FOKs (e.g., Koriat, 1995). All words are associated with other words, but the size of the associative set
varies. The impact of the number of associates on memory
is similar to explicit retroactive interference (e.g., Eakin,
2005). When a target is studied alone, but recall is later cued
by an associatively related word (i.e., extralist cueing),
recall is typically worse when the cue has a larger, rather
than a smaller associative set size. These set-size effects
are explained by the Processing Implicit and Explicit Representations Model (PIER2; Nelson, McKinney, Gee, &
Janczura, 1998), which posits that set-size effects are
obtained because there is more implicit interference from
competing associates when the associative set size is large.
Presentation of the extralist cue activates all of its associates, including the target. The nontarget associates compete with the target for access, thereby generating greater
implicit interference for larger versus smaller set sizes.
Cue-set-size effects can be eliminated in two ways. First,
when the cue and target are studied together as paired
associates—or using intralist cueing—cue-set-size effects
are not obtained (Eakin & Hertzog, 2006; Nelson & McEvoy,
1979). According to PIER2, providing both the cue and
target at study most strongly activate those associates shared
by both words (see also Nelson, McEvoy, & Schreiber,
1990). Intralist cueing reduces implicit interference because
competition arises only from associates in the intersection
of the cue and target set. Therefore, the size of the sampling
set for a large-set-size cue or target is effectively reduced to
that of the small-set-size cue or target. Second, cue-set-size
effects are eliminated when memory is tested using a forced
choice recognition test for which the foils are members of
the cue’s associative set (e.g., Eakin & Hertzog, 2006). The
leading explanation is that, when recognition is prompted
with a large-set-size cue, the functional sampling set is
reduced to the alternative and foils presented on the five
alternative forced-choice recognition test used in that study.
Prior research has found age differences in the impact of
interference on both recall and recognition memory (see
Eakin & Hertzog, 2006, for a review). Therefore, the paradigm is a good one for examining age differences in FOKs
to see whether both younger and older adults accurately
predict the influence of implicit interference on episodic
memory.
Implicit Interference Effects on Metamemory
Delayed judgments of learning (DJOLs) have been
shown to covary with implicit interference effects on recall.
DJOLs are predictions about future recall that are made
after studying an item, but before attempting to recall it.
Like FOKs, DJOLs are believed to be based on information
accessed at the time of the judgment (e.g., Koriat, 1995).
Because implicit interference influences retrieval of this
information, DJOLs vary in the same way as recall with
regard to implicit interference. In addition, this finding was
obtained for both younger and older adults (Eakin &
Hertzog, 2006); no age differences were observed.
Although the results from Eakin and Hertzog (2006)
found that DJOLs reflected the impact of implicit interference on recall, they did not reflect implicit interference
effects on recognition. One reason for DJOLs’ failure to
predict recognition in Eakin and Hertzog (2006) could be
primarily due to the different impact of implicit interference
on recognition than on recall; participants were told to
predict recall. Despite predicting different criterion tasks
(recall vs. recognition), one could argue that DJOLs and
FOKs are highly similar in nature, differing only with
respect to whether the prior retrieval search is covert and
implicitly required by the metacognitive judgment (DJOL;
Nelson, Narens, & Dunlosky, 2004) or overt and required
by the memory task preceding the metacognitive judgment
(FOK). If so, use of recognition memory as the task to be
predicted could limit FOK resolution because the information available at prediction would be derived from past recall
rather than future recognition. Indeed, we hypothesized that
FOKs would behave similarly to DJOLs, both in terms of
entrainment by recall outcomes and by sensitivity to associative set-size effects in recall, rather than recognition.
The forgoing arguments are relevant to whether manipulating associative set size would generate age differences in
episodic FOK resolution. If DJOLs and FOKs are based on
similar evidence at the time of a retrieval attempt, then one
would predict equivalent relationships of FOKs with cued
recall outcomes for both old and young adults. This relationship is rarely evaluated in FOK research. If individuals
generically rely on the same factors, such as accessibility,
when making FOKs that they do when making DJOLs, both
age groups would demonstrate low FOK resolution (as in
Eakin & Hertzog, 2006). Conversely, different information
is accessed when making FOKs as opposed to DJOLs—and
if younger adults access information that is actually diagnostic of subsequent recognition memory—then one might
observe age differences in FOK resolution for recognition,
even if none are observed for FOK resolution for cued
recall. Functionally, what would be required is for people to
override the influence of implicit interference on retrieval
when making an FOK, attending instead to other sources
of information that could be used to successfully predict
recognition memory outcomes. Doing so may not be easy,
even for the younger adults. Some studies suggest that younger
adults routinely fail to anticipate learning and forgetting
effects when making JOLs. However, older adults may
have even more difficulty doing so. Age-related differences
AGE
AGE INVARIANCE
INVARIANCE IN
IN FOKs
FOKs
in inhibition (e.g., Hasher & Zacks, 1988; Zacks & Hasher,
1994; see also Holley & McEvoy, 1996) and/or reduced
executive processing resources (e.g., Perrotin, Isingrini,
Souchay, Clarys, & Taconnat, 2006; Perrotin, Tournelle, &
Isingrini, 2008; Souchay, Isingrini, Clarys, Taconnat, &
Eustache, 2004) may reduce the likelihood that older adults
will seek and evaluate different sources of evidence when
making FOKs.
Method
Design and Participants
The design was a 2 × 2 × 2 mixed-model factorial design.
Cue set size (small, large) was manipulated within subjects.
Cueing procedure (extralist, intralist) and age group
(younger, older) were between-subjects factors. Fifty-one
younger adults (M = 19.58, SD = 2.05) from Mississippi
State University participated in exchange for course credit.
Forty-two older adults (M = 68.09, SD = 4.45) were
recruited from the Atlanta, Georgia community and were
reimbursed for participation.
Materials
Stimulus materials.—We created a list of 44 related cue
and target word pairs using the University of South Florida
word association norms (Nelson et al., 1990). Half of the
word pairs had small-set-size cues (5–9 associates, M =
6.79, SD = 0.16) and half had large-set-size cues (16–24
associates, M = 19.75, SD = 0.10). We equated forward
association strength (M = 0.12, SD = .03) and backward
association strength (M = 0.03, SD = 0.01) across cue set
size and list. Note that the association strength was relatively low; the target was never the most highly associated
member of the cue’s semantic set. We also equated target
set size, printed word frequency (Kučera & Francis, 1967),
concreteness, and connectivity (Nelson, McEvoy, & Schreiber,
1998). We used the ListChecker Pro 1.2 program (Eakin,
2010) to ensure that each cue was associatively related only
to its intended target and not to any other target or cue on
the list.
The paper-and-pencil recognition test consisted of five
alternatives, one of which was the target. A Scantron form was
provided for response collection. Each of the 44 cues was
presented (e.g., TWIG) along with its intended target (TREE)
and four semantically associated foils (LEAF, BRANCH,
BREAK, STICK) in random order. The intended target had
the strongest forward association strength with the cue only
about 15% of the time. Four versions of the recognition test
for each list were created with the cue order and the order of
the alternatives differing randomly across versions.
Additional materials.—We used the Mini-Mental State
Exam (MMSE; Cockrell & Folstein, 1988) to screen
5573
participants for impaired cognitive functioning; we excluded eight older adults as a result of their low performance (MMSE score < 27). The mean score on the MMSE
was 29.9 for the older adults. We administered a demographic questionnaire, including questions about age, education, overall health, and ethnicity. Younger adults (M =
4.01, SEM = 0.12) reported significantly better overall
health, F(1, 91) = 6.91, p = .009, ηp2 = .04, than the older
adults (M = 2.31, SEM = 0.15). The Shipley (1940) vocabulary test was administered to both groups; older adults (M =
0.90, SEM = 0.01) scored significantly better than the younger adults (M = 0.75, SEM = 0.01), F(1, 91) = 62.15, p <
.001, ηp2 = .29. Education level was numerically, but not
significantly, higher for the older (M = 15.28, SEM = 0.33)
than the younger (M = 13.37, SEM = 0.26) adults. The
groups did not vary on any other demographic factor.
Procedure
The experimental tasks were programmed using E-prime
version 1.1 (Psychology Software Tools, Inc.) and were
executed on standard PCs. The computerized experiment
consisted of three distinct phases: (a) the encoding phase,
(b) the recall phase, and (c) the FOK prediction phase. The
recognition phase was presented using a paper/pencil test.
A six-item practice session preceded each phase.
Intralist cueing procedure.—During the encoding phase,
participants viewed cue target word pairs (e.g., TWIGTREE) for 8 s. They were instructed to form an interactive
image of the two words; participants pressed the space bar
to indicate when an image was formed. After the 8 s, participants rated their image as vivid (clear with lots of detail) or
neutral (unclear and vague) or pressed the enter key if they
were not able to form an image (compliance rates and proportion of items rated vivid were not correlated with memory outcome). All 44-list pairs were presented for encoding
in random order. The recall phase followed; each of the
44 cues from the word pairs was presented in random order
and participants typed in the target word with which the
presented word had been paired during the encoding phase.
We encouraged participants to try hard to recall and to guess
if they were able to do so. If unable to recall or guess, participants typed “NEXT” on the keyboard. All items were
presented for recall before proceeding to the FOK phase.
During the FOK phase, regardless of recall outcome, the
same 44 cues were again randomly presented, and participants provided an FOK for each item. We told them to base
their FOKs on their sense of whether they would be able to
recognize the item from among five alternatives on a future
recognition test. Time was not limited, but we asked participants to make their predictions quickly using a percentage
confidence rating scale of 0–100 (0 = certain not to recognize, 100 = certain to recognize), and we encouraged them
to use the full range of the scale.
4558
EAKIN AND HERTZOG
Table 1. Recall and Recognition Memory Performance
Younger adults
Small set
Cueing procedure
Older adults
Large set
Small set
Large set
M
SD
M
SD
M
M
SD
M
SD
M
Extralist
Intralist
M
0.56
0.79
0.68
0.20
0.14
0.41
0.81
0.61
0.15
0.15
0.48
0.80
0.64
0.48
0.64
0.56
0.17
0.20
0.46
0.64
0.55
0.15
0.23
0.47
0.64
0.55
Extralist
Intralist
M
Extralist
Intralist
M
0.81
0.89
0.85
0.73
0.55
0.64
0.11
0.27
0.80
0.90
0.85
0.75
0.59
0.67
0.26
0.11
0.81
0.89
0.85
0.74
0.57
0.65
0.67
0.74
0.70
0.59
0.50
0.55
0.29
0.28
0.69
0.78
0.73
0.70
0.61
0.66
0.33
0.18
0.68
0.76
0.72
0.65
0.55
0.60
Recall
Recognition
All items
Unrecalled
0.28
0.31
0.25
0.30
0.31
0.33
0.30
0.25
Note. Recognition means are listed both for all items and for unrecalled items only. Marginal means are included.
After all 44 items were presented, the recognition phase
began. A paper version of the recognition test was provided
on a clipboard, along with a Scantron form on which to
mark responses. Each of the 44 cues was presented along
with five alternatives, one of which was the target with
which the cue was paired during the encoding phase. Participants filled in the bubble corresponding to the alternative
(a, b, c, d, or e) to indicate their response. The recognition
test was self-paced.
Extralist cueing procedure.—All instructions and procedures for extralist cueing were the same as for intralist cueing
except that, during the encoding phase, only the target from
each cue target pair was presented. Then the same 44 associated cues were presented during the recall, FOK, and recognition phases, in a different random order for each phase.
overall recognition memory by younger adults (see Table 1).
Age did not interact with cue set size F < 1.00 or cueing
procedure, F < 1.00.
Probability of Recognition—Unrecalled Items Only
Recognition of unrecalled items did not differ between
the two age groups, F(1, 84) = 1.08, p = .30. Set-size effects
were not obtained under extralist cueing for either age
group (see Table 1). None of the interactions approached
significance.
Results
FOK Sensitivity
Our primary focus was on whether older adults manifested impaired FOK accuracy in either (a) predicting subsequent recognition memory or (b) influence of recall
success on FOKs. However, mean FOKs are reported first
(see Table 2).
Probability of Recall
A general linear model produced two key results. First,
the expected interaction between cue set size and cueing
procedure was significant, F(1, 89) = 15.31, p < .001, ηp2 =
.15. Cue-set-size effects were obtained for extralist cueing
and eliminated for intralist cueing. Second, age influenced
the effect of cue set size on recall. The main effect of age
was significant, F(1, 89) = 6.70, p = .01, ηp2 = .07, and the
three-way interaction between age, cueing procedure, and
cue set size was significant, F(1, 89) = 9.87, p = .002, ηp2 =
.10. Cue-set-size effects were obtained under extralist
cueing and were eliminated under intralist cueing for the
younger adults, not the older adults (see Table 1). For the
older adults, cue-set-size effects were not obtained for
either cueing procedure.
FOKs for all items.—The main effect of recall outcome
was significant, F(1, 83) = 31.63, p < .001, ηp2 = .28. FOKs
were higher for recalled (M = 90.56, SEM = 0.80) than for
unrecalled (M = 55.09, SEM = 6.46) targets. This finding
did not differ for the two age groups; the interaction was not
significant, F(1, 83) = 3.19, p = .08. Similar to recall, cue set
size interacted with cueing procedure, F(1, 89) = 14.36, p <
.001, ηp2 = .14; cue-set-size effects were obtained for
extralist cueing and were eliminated in intralist cueing. This
finding was obtained for both younger and older adults.
Although recall varied with age, FOKs for all items did
not, F(1, 89) < 1.00. Age did not interact significantly with
cueing procedure, F(1, 89) < 1.00, or with cue set size,
F(1, 89) < 1.00, and the three-way interaction was not
significant, F(1, 89) = 2.98, p = .09.
Probability of Recognition—All Items
The mean probability of recognition varied between
age groups, F(1, 89) = 8.15, p = .005, ηp2 =.08, with better
FOKs for unrecalled items only.—FOK research traditionally evaluates FOKs for unrecalled items only. Variations in
FOKs did not coincide with recognition of unrecalled items.
5595
AGE
AGE INVARIANCE
INVARIANCE IN
IN FOKs
FOKs
Table 2. Mean Feeling of Knowing Sensitivity for All Items and for Unrecalled Items
Younger adults
Older adults
Small set
Cueing procedure
All
Extralist
Intralist
M
Recalled
Extralist
Intralist
M
Unrecalled
Extralist
Intralist
M
Large set
Small set
M
SD
M
SD
M
73.35
85.35
79.35
15.55
12.56
61.31
84.53
72.92
18.43
15.05
67.33
84.94
76.13
94.30
93.88
94.09
5.54
6.54
93.06
89.50
91.28
8.93
10.09
45.25
52.86
49.06
24.85
23.01
40.14
48.75
44.44
24.83
27.17
M
Large set
SD
M
SD
M
70.73
82.80
76.76
15.65
20.41
63.12
79.39
71.25
19.64
19.61
66.92
81.10
74.01
91.69
93.68
92.68
87.96
90.94
89.45
10.91
6.90
85.03
92.18
88.61
10.98
7.54
91.56
86.50
89.03
42.70
50.81
46.75
52.88
56.78
54.83
25.53
21.35
44.34
49.93
47.14
25.49
24.84
48.61
53.36
50.99
Note. Marginal means are included.
Although recognition was equal for unrecalled targets given
small- and large-set-size cues, FOKs were higher for smallthan for large-set-size cues, F(1, 84) = 11.41, p = .001,
ηp2 = .12 (M = 51.95, SEM = 2.55 and M = 45.79, SEM =
2.73, respectively). FOKs also did not vary with cueing procedure (F(1, 84) = 1.69, p = .20) and did not interact with
age (F(1, 84) < 1.00). None of the other interactions were
significant.
FOK Resolution
Although FOK sensitivity findings suggest that—similar
to DJOLs—FOKs were influenced by the impact of implicit
interference on recall, rather than an examination of resolution indicates whether FOKs were correlated with recognition outcome and whether aging impaired FOK accuracy at
predicting the impact of implicit interference on recognition.
A general linear model was conducted on within-person
ordinal Goodman–Kruskal gamma correlations (G) between
FOKs and both recognition and recall (see Table 3).
FOK resolution with recognition.—FOK resolution for
all items, regardless of recall outcome, was reliably above
chance, t(63) = 9.01, p < .001. Notably, resolution did not
vary with age group (F(1, 61) < 1.00), indicating that younger and older adults were equally accurate at predicting recognition outcome when all items were included in the
analysis, even though that outcome differed for younger and
older adults. For items that were not recalled, however,
it was clear that neither younger nor older adults could
accurately predict recognition outcomes. When recall
failed, resolution was poor for both age groups; the overall
mean G was essentially zero.
FOKs resolution with cued recall.—To determine the
influence of prior experience with implicit interference
obtained during recall, FOKs were correlated with recall
success. Resolution was reliably greater than chance,
t(92) = 39.76, p < .001, with an overall mean of 0.80 (SEM =
0.02). Small age differences were obtained. Younger adults
(M = 0.87, SEM = 0.03) were more accurate than older
Table 3. Feeling of Knowing Resolution (Feeling of Knowing × Recognition) for All and Unrecalled Items Only
Younger adults
Small set
FOK × Recognition
All
Extralist
Intralist
M
Unrecalled
Extralist
Intralist
M
FOK × Recall
Extralist
Intralist
M
M
Older adults
Large set
Small set
Large set
SD
M
SD
M
M
SD
M
SD
M
0.29
0.68
0.48
0.54
0.26
0.22
0.29
0.26
0.61
0.92
0.25
0.49
0.37
0.27
0.56
0.41
0.58
0.23
0.35
0.45
0.40
0.50
0.60
0.31
0.50
0.41
−0.09
0.02
−0.03
0.49
0.71
−0.04
−0.22
−0.13
0.61
0.92
−0.06
−0.10
−0.08
0.20
−0.11
0.04
0.77
0.63
−0.01
0.22
0.13
0.66
0.78
0.11
0.06
0.08
0.86
0.86
0.86
0.05
0.07
0.86
0.88
0.87
0.05
0.07
0.87
0.86
0.73
0.65
0.69
0.07
0.07
0.80
0.77
0.79
0.06
0.06
0.76
0.74
Notes. FOK resolution for recall (FOKs × recall) is included in the lower section of the table for comparison purposes. Marginal means are included. FOK =
feeling of knowing.
6560
EAKIN AND HERTZOG
adults (M = 0.74, SEM = 0.03), F(1, 83) = 6.44, p = .01,
ηp2 = .07.
Recall gammas versus recognition gammas.—A statistical
comparison of mean G for FOK-by-recall to mean G for
FOK-by-recognition highlighted the differential findings
depending on the type of memory test. Cued recall was
much more highly correlated with FOKs (M = 0.80, SEM =
0.02) than recognition performance (M = 0.46, SEM =
0.05), F(1, 84) = 55.57, p < .001, ηp2 = .40.
Discussion
We used implicit interference effects on memory as a
basis for evaluating FOKs and FOK resolution in younger
and older adults. Age differences were obtained both for
recall and for recognition. For recall, cue-set-size effects
were obtained under extralist cueing and eliminated under
intralist cueing and recognition for younger adults. For
older adults, cue-set-size effects were not obtained for
extralist cueing. Therefore, set-size effects on recall and
recognition were different for the two age groups, and these
effects could have, in principle, impacted FOKs and FOK
resolution differentially for younger and older adults.
Both younger and older adults provided FOKs that were
inconsistent with the lack of implicit interference effects on
recognition memory. Although both age groups showed
reliable FOK correlations with recognition memory for all
items (when successfully recalled items were included in
the correlations), neither group showed above-chance FOK
resolution when the analysis was restricted only to unrecalled items. This finding suggests that neither age group
was able to override the implicit interference at the time of
the FOK in order to accurately reflect the fact that implicit
interference would be resolved on the recognition test.
Rather, mean FOKs manifested implicit interference effects
that mirrored those influences of set size on recall. This
finding is consistent with prior research with young adults
showing that DJOLs and FOKs both mirrored cue-set-size
(Schreiber & Nelson, 1998) and target-set-size (Schreiber,
1998) effects in cued recall. Indeed, FOKs were much more
highly correlated with recall than with recognition, pointing
to the strong influence of recall on subsequent FOKs, even
though FOKs were collected in a block that was delayed
until after the recall test.
The most plausible reason that FOKs were more highly
correlated with recall than with recognition outcomes is that
people were using retrieval of a candidate answer both
at recall and at the time of the FOK and that candidate
retrieval was highly correlated across these two retrieval
attempts (Krinsky & Nelson, 1985). We tested this by eliminated commission errors and recalculated gammas between
FOKs and recognition, which resulted in a mean gamma of
.45 (SE = 0.07) with no significant main effects or interactions. Accuracy was somewhat higher when commission
errors were eliminated than when they were included (G =
.39). Apparently, when a response was accessible (even
when it was not correct), people had high confidence in
future recognition. Gammas for omission errors only were
low (G = .12, SE = 0.16), indicating that when no response
came to mind, FOKs were indiscriminate with regard to
future recognition. Alternatively, FOKs may have been
based on retrieval of information about whether the item
had been previously recalled, as in memory for past test
heuristic by Finn and Metcalfe (2008). In either case, the
fact that FOKs for unrecalled items were not correlated with
subsequent recognition indicated that either (a) few other
cues were utilized when making the FOKs or (b) any other
cues that were being accessed for such trials lacked predictive validity for recognition performance. Although people
were explicitly trying to predict recognition, the conditions
at the time of the FOK prediction more closely matched
those at recall than at recognition. For both FOKs and
recall, a cue was presented in isolation and, regardless of
cueing procedure, the implicit interference experienced
when the cue is processed for FOK was prescriptive of the
impact of implicit interference on recall. However, presumably because this influence is implicit, people did not predict that it would be reduced for recognition.
Implicit interference effects on FOKs support the accessibility view of FOK construction (Koriat, 1995; Koriat &
Levy-Sadot, 2001). Varying degrees of implicit interference
from associates of the cue influenced the accessibility of
information available when FOKs were made. Accessibility
of information affected recall but was neutralized by presenting alternatives on the forced-choice recognition test.
It is interesting to note that there was one indication of an
age difference in FOK resolution; gamma correlations of
FOKs with recall were reliably higher for younger adults,
even though they were much higher for both age groups
than the correlations of FOKs with recognition memory.
This finding is not material with respect to the issue of age
differences in episodic FOK resolution because other
studies have focused exclusively on resolution for unrecalled items, given their analog to feeling of knowing as
originally intended (e.g., Hart, 1965). Nevertheless, this age
difference is intriguing. We suggest that one influence on
the difference is that older adults are less confident than
younger adults that they will correctly recognize successfully recalled items (see MacLaverty & Hertzog, 2009, for a
similar finding). This effect probably diluted the correlation
of FOKs with prior recall outcomes. This outcome also suggests at least some divergence between FOKs and DJOLs
because it was not observed for DJOLs in the Eakin and
Hertzog (2006) study.
Adults of both age groups produced chance levels of
FOK resolution for recognition memory for unrecalled
items. In contrast, other published studies of age differences
in episodic FOK resolution using unrelated items (e.g.,
Hertzog et al., 2010; MacLaverty & Hertzog, 2009) have
AGE
AGE INVARIANCE
INVARIANCE IN
IN FOKs
FOKs
found above-chance FOK resolution for both age groups.
MacLaverty and Hertzog found equivalent resolution for
younger and older adults in both immediate and delayed
FOKs (delayed FOKs were used in the present study), so
the issue appears to be one of low resolution for related
items. Resolution of immediate JOLs has also been shown
to be better for unrelated than for related pairs (Connor,
Dunlosky, & Hertzog, 1997; Hertzog, Kidder, PowellMoman, & Dunlosky, 2002; Hertzog, Sinclair, & Dunlosky,
2010). This study, in combination with results from
MacLaverty and Hertzog, suggests that associative relatedness also influences the resolution of FOKs.
The present study adds to the growing evidence that age
differences in episodic FOKs found in some studies (e.g.,
Souchay et al., 2007) are far from universal. Instead, the age
differences in FOK resolution found by Souchay and
colleagues (2007) may depend instead on differences in
methods (e.g., type of recognition memory test) or realized
samples of older adults. For instance, older adults of
Souchay et al. (2000) had much lower mean education and
lower scores on the MMSE than those in the present study.
In addition, episodic FOK accuracy is impaired by Alzheimer’s disease (Souchay, Isingrini, & Gil, 2002), mild
cognitive impairment (Anderson & Schmitter-Edgecombe,
2010), and low scores on neuropsychological tests of frontal function (e.g., Perrotin et al., 2008; Souchay & Isingrini,
2004; Souchay et al., 2000). Such findings may indicate that
studies finding age differences in episodic FOK resolution
have included older adults with nonnormative brain pathology in their sample (see Eakin, Harris, & Hertzog,
2010; MacLaverty & Hertzog, 2009, for further discussion).
Conclusions
Many factors such as the number of associates of a cue,
cueing procedure, and the type of memory test can
metamemory. However, age does not appear to be one of
those factors. As in the case of predicting implicit interference effects on recognition, sometimes people of both age
groups are equally inaccurate. Accurate metamemory predictions derive from people attending to diagnostic sources
of information when making predictions. When metamemory judgments are based on invalid sources of information
or ignore valid sources, dissociation between metamemory
judgments and memory outcomes have been obtained (e.g.,
Eakin, 2005; Koriat & Bjork, 2005, 2006). As this study
demonstrated, the influence of implicit interference is different on FOKs than on recognition, and people of both age
groups failed to predict this differential impact. However,
metamemory judgments reflected the impact of implicit
interference on recall, and in this case, both younger and
older adults’ FOKs were highly correlated with recall outcomes. Even though age differences were obtained in recall,
both age groups relied on accessibility during the recall test
as a basis for FOKs, leading to high correlations of FOKs
5617
with recall for both age groups but minimizing correlations
of FOKs with subsequent recognition memory.
Funding
This research was supported in part by a postdoctoral fellowship
awarded to D. K. Eakin by the National Institute on Aging (NIA; Cognitive
Aging Training Grant NIA 23T2 AG00175-11) at the Georgia Institute of
Technology. This research was also supported by Grant NIA R37 AG13148
awarded to C. Hertzog.
Acknowledgments
The authors would like to thank Elizabeth Crommelin, Rosalyn Daniels,
Shuntal Dean, Lindsey McKay, Miranda Morris, Leigh Nadel, Jared Pryor,
Helen Saputra, and Mary Wheeler for their assistance with data collection
and processing.
Correspondence
Correspondence should be addressed to Deborah K. Eakin, PhD, Department
of Psychology, Mississippi State University, P.O. Box 6161, Mississippi
State, MS 39762. E-mail: [email protected].
References
Anooshian, L. J., Mammarella, S. L., & Hertel, P. T. (1989). Adult age
differences in knowledge of retrieval processes. International Journal
of Aging and Human Development, 29, 39–52. doi:10.2190/ERPX9NU8-HNRM-QXTR
Blake, M. (1973). Prediction of recognition when recall fails: Exploring
the feeling-of-knowing phenomenon. Journal of Verbal Learning and
Verbal Behavior, 12, 311–319. doi:10.1016/S0022-5371(73)80075-1
Butterfield, E. C., Nelson, T. O., & Peck, G. (1988). Developmental
aspects of the feeling of knowing. Developmental Psychology, 24,
654–663. doi:10.1037/0012-1649.24.5.654
Cockrell, J. R., & Folstein, M. F. (1988). Mini-Mental State Examination
(MMSE). Psychopharmacology Bulletin, 24, 689–692.
Connor, L. T., Dunlosky, J., & Hertzog, C. (1997). Age-related differences
in absolute but not relative metamemory accuracy. Psychology and
Aging, 12, 50–71. doi:10.1037/0882-7974.12.1.50
Costermans, J., Lories, G., & Ansay, C. (1992). Confidence level and
feeling of knowing in question answering: The weight of inferential
processes. Journal of Experimental Psychology: Learning, Memory,
and Cognition, 18, 142–150. doi:10.1037/0278-7393.18.1.142
Dunlosky, J., & Metcalfe, J. (2009). Metacognition. Beverly Hills, CA: SAGE.
Eakin, D., Harris, W., & Hertzog, C. (2010). Age invariance in metamemory
for both semantic and episodic memory: Both younger and older adults
provide accurate feeling of knowing predictions. Paper presented at the
51st annual meeting of the Psychonomic Society, St Louis, MO.
Eakin, D. K. (2005). Illusions of knowing: Metamemory and memory
under conditions of retroactive interference. Journal of Memory and
Language, 52, 526–534. doi:10.1016/j.jml.2005.01.009
Eakin, D. K. (2010). ListChecker Pro 1.2: A program designed to facilitate
creating word lists using the University of South Florida word association norms. Behavior Research Methods, 42, 1012–1021.
doi:10.3758/BRM.42.4.1012
Eakin, D. K., & Hertzog, C. (2006). Release from implicit interference in
memory and metamemory: Older adults know that they can’t let go.
The Journals of Gerontology, Series B: Psychological Sciences and
Social Sciences, 61, 340–347.
Finn, B., & Metcalfe, J. (2008). Judgments of learning are influenced by
memory for past test. Journal of Memory and Language, 58, 19–34.
doi:10.1016/j.jml.2007.03.006
Hart, J. T. (1965). Memory and the feeling-of-knowing experience. Journal
of Educational Psychology, 56, 208–216. doi:10.1037/h0022263
Hasher, L., & Zacks, R. (1988). Working memory, comprehension, and
aging: A review and a new view. In G. H. Bower (Ed.), The psychology of learning and motivation: Advances in research and theory
8562
EAKIN AND HERTZOG
(Vol. 22, pp. 193–225). San Diego, CA: Academic Press. doi:10.1016/
S0079-7421(08)60041-9
Hertzog, C., Dunlosky, J., & Sinclair, S. M. (2010). Episodic feeling-ofknowing resolution derives from the quality of original encoding.
Memory & Cognition, 46, 939–948. doi:10.3758/MC.38.6.771
Hertzog, C., Kidder, D., Powell-Moman, A., & Dunlosky, J. (2002). Aging
and monitoring associative learning: Is monitoring accuracy spared
or impaired? Psychology and Aging, 17, 209–225. doi:10.1037/08827974.17.2.209
Hertzog, C., Sinclair, S. M., & Dunlosky, J. (2010). Age differences in the
monitoring of learning: Cross-sectional evidence of spared resolution
across the adult life span. Developmental Psychology, 46, 939–948.
doi:10.1037/a0019812
Holley, P. E., & McEvoy, C. L. (1996). Aging and inhibition of unconsciously processed information: No apparent deficit. Applied Cognitive
Psychology, 10, 241–256. doi:10.1002/(SICI)1099-0720(199606)
10:3<241::AID-ACP377>3.0.CO;2–3
Koriat, A. (1995). Dissociating knowing and the feeling of knowing: Further evidence for the accessibility model. Journal of Experimental
Psychology: General, 124, 311–333. doi:10.1037/0096-3445.124.3.311
Koriat, A., & Bjork, R. (2005). Illusions of competence in monitoring
one’s knowledge during study. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 31, 187–194. doi:10.1037/02787393.31.2.187
Koriat, A., & Bjork, R. (2006). Mending metacognitive illusions: A comparison of mnemonic-based and theory-based procedures. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 32,
1133–1145. doi:10.1037/0278-7393.32.5.1133
Koriat, A., & Levy-Sadot, R. (2001). The combined contributions of the
cue-familiarity and accessibility heuristics to feelings of knowing.
Journal of Experimental Psychology: Learning, Memory, and Cognition, 27, 34–53. doi:10.1037/0278-7393.27.1.34
Krinsky, R., & Nelson, T. O. (1985). The feeling of knowing for different
types of retrieval failure. Acta Psychologica, 58, 141–158.
doi:10.1016/0001-6918(85)90004-6
Kučera, H., & Francis, W. N. (1967). Computational analysis of present-day American English. Providence, RI: Brown University Press.
Leonesio, R. J., & Nelson, T. O. (1990). Do different metamemory judgments tap the same underlying aspects of memory? Journal of Experimental Psychology: Learning, Memory, and Cognition, 16, 464–470.
doi:10.1037/0278-7393.16.3.464
MacLaverty, S. N., & Hertzog, C. (2009). Do age-related differences in
episodic feeling of knowing accuracy depend on the timing of the
judgment? Memory, 17, 860–873. doi:10.1080/09658210903374537
Marquie’, J. C., & Huet, N. (2000). Age differences in feeling-of-knowing
and confidence judgments as a function of knowledge domain. Psychology and Aging, 15, 451–461. doi:10.1037/0882-7974.15.3.451
Metcalfe, J., Schwartz, B. L., & Joaquim, S. G. (1993). The cue-familiarity
heuristic in metacognition. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 19, 851–861. doi:10.1037/02787393.19.4.851
Nelson, D. L., & McEvoy, C. L. (1979). Encoding context and set size.
Journal of Experimental Psychology: Human Learning and Memory,
5, 292–314. doi:10.1037//0278-7393.5.3.292
Nelson, D. L., McEvoy, C. L., & Schreiber, T. A. (1990). Encoding context
and retrieval conditions as determinants of the effects of natural
category size. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 16, 31–41. doi:10.1037/0278-7393.16.1.31
Nelson, D. L., McEvoy, C. L., & Schreiber, T. A. (1998). The University of
South Florida word association, rhyme, and word fragment norms
Retrieved from http://www.usf.edu/FreeAssociation/
Nelson, D. L., McKinney, V. M., Gee, N. R., & Janczura, G. A. (1998).
Interpreting the influence of implicitly activated memories and recognition. Psychological Review, 105, 299–324. doi:10.1037/0033295X.105.2.299
Nelson, T. O., Gerler, D., & Narens, L. (1984). Accuracy of feeling-ofknowing judgments for prediction perceptual identification and
relearning. Journal of Experimental Psychology. General, 113, 282–
300. doi:10.1037/0096-3445.113.2.282
Nelson, T. O., & Narens, L. (1990). Metamemory: A theoretical framework
and new findings. The Psychology of Learning and Motivation, 26,
125–141. doi:10.1016/S0079-7421(08)60053-5
Nelson, T. O., Narens, L., & Dunlosky, J. (2004). A revised methodology
for research on metamemory: Pre-judgment recall and monitoring
(PRAM). Psychological Methods, 9, 53–69. doi:10.1037/1082989X.9.1.53
Perrotin, A., Isingrini, M., Souchay, C., Clarys, D., & Taconnat, L. (2006).
Episodic feeling-of-knowing accuracy and cued recall in the elderly:
Evidence for double dissociation involving executive functioning
and processing speed. Acta Psychologica, 122, 58–73. doi:10.1016/
j.actpsy.2005.10.003
Perrotin, A., Tournelle, L., & Isingrini, M. (2008). Executive functioning
and memory as potential mediators of the episodic feeling-of-knowing
accuracy. Brain and Cognition, 67, 76–87. doi:10.1016/j.bandc.
2007.11.006
Schacter, D. L. (1983). Amnesia observed: Remembering and forgetting in
a natural environment. Journal of Abnormal Psychology, 92, 236–
242. doi:10.1037//0021-843X.92.2.236
Schreiber, T. A. (1998). Effects of target set size on feelings of knowing
and cued recall: Implications for the cue effectiveness and partialretrieval hypotheses. Memory & Cognition, 26, 553–571. doi:10.3758/
BF03201162
Schreiber, T. A., & Nelson, D. L. (1998). The relation between feelings of
knowing and the number of neighboring concepts linked to the test
cue. Memory & Cognition, 26, 869–883. doi:10.3758/BF03201170
Shipley, W. C. (1940). A self-administered scale for measuring intellectual
impairment and deterioration. Journal of Psychology, 9, 371–377.
doi:10.1080/00223980.1940.9917704
Singer, M., & Tiede, H. (2008). Feeling of knowing and duration of unsuccessful memory search. Memory & Cognition, 36, 588–597.
doi:10.3758/MC.36.3.588
Souchay, C., Isingrini, M., Clarys, D., Taconnat, L., & Eustache, F. (2004).
Executive functioning and judgment-of-learning versus feeling-ofknowing in older adults. Experimental Aging Research, 30, 47–62.
doi:10.1080/03610730490251478
Souchay, C., Isingrini, M., & Espagnet, L. (2000). Aging, episodic memory
feeling-of-knowing and frontal functioning. Neuropsychologia, 14,
299–309. doi:10.1037/0894-4105.14.2.299
Souchay, C., Moulin, C. J. A., Clarys, D., Taconnat, L., & Isingrini, M.
(2007). Diminished episodic memory awareness in older adults:
Evidence from feeling-of-knowing and recollection. Consciousness
and Cognition, 16, 769–784. doi:10.1016/j.concog.2006.11.002
Zacks, R. T., & Hasher, L. (1994). Directed ignoring: Inhibitory regulation
of working memory. In D. Dagenbach & T. H. Carr (Eds.), Inhibitory
processes in attention, memory, and language (pp. 241–264). San
Diego, CA: Academic Press.