Psychological Research Psychologische Forschung
© Springer-Verlag 2003
DOI 10.1007/s00426-003-0138-5
Original Article
The word-frequency paradox for recall/recognition
occurs for pictures
Paul Johan Karlsen ( ') · Joan Gay Snodgrass
P. J. Karlsen
Institute of Psychology, University of Oslo, Box 1094, 0317 Oslo, Norway
J. G. Snodgrass
New York University, USA
P. J. Karlsen
'
Phone: + 47-22845000
Fax: [phone]
E-mail: [e-mail]
Received: 13 December 2003 / Accepted: 28 April 2003 / Published online: 25 June
2003
Abstract A yes-no recognition task and two recall tasks were conducted using pictures of
high and low familiarity ratings. Picture familiarity had analogous effects to word frequency,
and replicated the word-frequency paradox in recall and recognition. Low-familiarity pictures
were more recognizable than high-familiarity pictures, pure lists of high-familiarity pictures
were more recallable than pure lists of low-familiarity pictures, and there was no effect of
familiarity for mixed lists. These results are consistent with the predictions of the Search of
Associative Memory (SAM) model.
The word-frequency paradox for recall/recognition
occurs for pictures
The word-frequency paradox in recall and recognition memory refers to the fact that:
a) Common words are recalled better than rare words, while
1
b) Rare words are recognized better than common words (Balota & Neely, 1980; Dobbins,
Kroll, Yonelinas, & Liu, 1998; Kintsch, 1970; Glanzer & Bowles, 1976; Guttentag & Carroll,
1994; Mandler, Goodman, & Wilkes-Gibbs, 1982; Otani & Whiteman, 1993; Wallace &
Page, 1982; Wallace, Sawyer, Timothy, & Robertson, 1978; Whiteman, Nairne, & Serra,
1994)
The typical recall advantage for common words is limited to pure lists that are composed
of words of only one type of frequency. The finding that word-frequency is positively related
to recallability of pure lists has been replicated numerous times (Bousfield & Cohen, 1955;
Deese, 1960; Hall, 1954, 1985; Murdock, 1960; Raymond, 1969; Schulman, 1976; Sumby,
1963; Ward, Woodward, Stevens, & Stinson, 2003). For mixed lists, consisting of words of
both types of frequency, there is often an advantage for rare words (Delosh & McDaniel,
1996; Dewhurst, Hitch, & Barry, 1998; Gregg, Montgomery, & Castano, 1980; May & Tryk,
1971; May, Cuddy, & Norton, 1979; Van Overschelde, 2002), or no difference at all (Clark
& Burchett, 1994; Duncan, 1974; MacLeod & Kampe, 1996; Ward et al., 2003; Watkins,
LeCompte, & Kim, 2000), and at least one study showed an advantage for high-frequency
words (Balota & Neely, 1980). The finding that word-frequency is negatively related to
recognizability seems to occur for both mixed and pure lists (Allen & Garton, 1968; Gorman,
1961; Kinsbourne & George, 1974; Guttentag & Carroll, 1997; McCormack & Swenson,
1972; Schulman, 1967; Schulman & Lovelace, 1970; Shepard, 1967; Underwood & Freund,
1970).
The purpose of the present study is to test whether the word-frequency paradox applies
to the realm of pictures whose frequency in daily life is measured by ratings of familiarity
(Snodgrass & Vanderwart, 1980; Yoon, Luo, Mikels, Hedden, Gutchess, Jiao, & Park, 2002).
Memory for pictures may share fundamental properties with memory for words, so it would
be of interest to establish whether the effect of familiarity on pictures is analogous to the
effect of frequency on words.
Researchers typically rely on independent word counts (e.g., Dahl, 1979; Kucera & Francis,
1967) to distinguish between words of high and low frequency in natural language. Such
tabulations are based on a text corpus assumed to be reasonably representative of printed
American English. The tabulation is also thought to be positively correlated with occurrences
of words in natural speech. Boy, chair, and blue are words that occur with high frequency in
natural language whereas words like tandem, ocelot, and pagoda occur with low frequency.
2
Previous studies have shown that when participants are allowed to rate words with respect
to subjective familiarity, the effects of this factor in later memory tasks resemble typical
frequency effects (Schulman, 1976).
Normative data on pictures are less extensively collected. The most commonly used norms
on characteristics of pictorial representations of concrete nouns are the Snodgrass and
Vanderwart norms (1980). Snodgrass and Vanderwart developed a set of 260 black-and-white
line drawings of concepts that were selected on the basis of three criteria:
a) That they be unambiguously picturable
b) That they include exemplars from the widely used category norms of Battig and Montague
(1969)
c) That they represent concepts at the basic level of categorization (Rosch, Mervis, Gray,
Johnson, & Boyes-Braem, 1976)
Among other characteristics of the pictures, Snodgrass and Vanderwart asked participants
to judge the familiarity of each picture  according to how usual or unusual the object is in
your realm of experience. Familiarity was defined as  the degree to which you come in
contact with or think about the concept. They were told to rate the concept itself, rather
than the way it was drawn. A 5-point scale was used in which 1 indicated very unfamiliar and
5 indicated very familiar. Pictures of body parts like arm, hand, and toe, were rated as highly
familiar, and people also agreed well on the names of these images. Pen, chair, and pants
are other examples of high-familiarity pictures, whereas ostrich, helicopter and pumpkin are
low-familiarity pictures that were named similarly across participants. Snodgrass and
Vanderwart found a positive but modest correlation between picture familiarity and concept
frequency in the Kucera-Francis word counts (r = .36). They hypothesized that this correlation
may be lower than expected because frequency values do not distinguish between different
meanings of a word. For example, well and saw both have high Kucera-Francis frequencies,
primarily because of non-noun usages. A recent study (Yoon et al., 2002) replicated the
picture familiarity norms with a group of 113 young adults, and so these new norms were
used in selecting pictures for the present studies as more than 20 years had elapsed since the
original Snodgrass and Vanderwart ratings were collected. Nonetheless, there is quite a high
(r = 0.92) correlation between the two sets of familiarity ratings. Yoon et al. provided no
measure of the correlation between familiarity and picture name frequency, but due to the
3
great stability in familiarity ratings over time it is likely to be similar to Snodgrass and
Vanderwort s finding.
The classical explanation of the effects of frequency on accuracy of performance in recall
and recognition was first proposed by Kintsch (1970) and extended by several other theorists.
They proposed a two-stage process of retrieval followed by a judgment of familiarity.
According to this two-stage model, common words have more associative links to other words
than do rare words. This makes them easier to retrieve, yielding a common-word advantage
in recall tasks, which rely heavily on retrieval processes. On the other hand, the two-stage
model proposes that a decision process favors rare words over common ones. It is easier to
make decisions about words that have relatively little irrelevant information from previous
encounters stored in long-term memory, and this leads to a rare-word advantage in recognition
tasks (Kintsch, 1970; Gillund & Shiffrin, 1984; Glanzer and Bowles, 1976; Mandler et al.,
1982).
The Search of Associative Memory (SAM) model follows the assumption that the memory
strength of an item is more fundamental than its number of possible meanings (Raaijmakers
& Shiffrin, 1981; Gillund & Shiffrin, 1984). To explain frequency effects Gillund and Shiffrin
proposed that long-term memory is composed of closely interconnected, relatively unitized,
permanent sets of features called images. Access to long-term memory occurs when probe
cues are assembled in short-term memory and used to activate an associated set of information
in long-term memory. The values of the retrieval structure at the time of recall or recognition
are the result of:
a) Rehearsal and coding processes that take place during the study of the list
b) Pre-experimental associations
c) The match of the cue encodings at study and test
The images associated with the retrieval cues are activated to differing degrees. It is
assumed that high-frequency items are more likely to access images in memory than
low-frequency items. That is, the retrieval strength between a high-frequency cue and any
image is stronger than that between a low-frequency cue and any image (Gillund & Shiffrin,
1984), and this difference is thought to be the result of long-term learning. It is also assumed
that inter-item strength within a list is based in part on extra-experimental semantic factors
that are sensitive to item frequency.
4
The SAM model predicts a pure-list high-frequency advantage in recall for two reasons.
Firstly, a given list item of any frequency, high or low, has a higher probability of recovery if
it is sampled by a high-frequency item as a cue than if it is sampled by a low-frequency item.
This is due to the higher strength values between high-frequency probes and any image, than
between low-frequency probes and any image. All items that are recalled during the test
phase function as cues for additional recall. With a pure high-frequency list only
high-frequency items are used as cues, so recall is superior to that of a pure low-frequency
list, where all cues are low-frequency items. Secondly, there is more self-sampling for pure
lists of low-frequency items because the strength of associations between items is lower in a
list of low-frequency items. That is, the participant is more likely to retrieve the same item
over and over again. Self-sampling reduces recall because it reduces the opportunities for
recovery of a new word. For mixed lists there is no effect of frequency because high-frequency
items are just as likely to sample a low-frequency item as a high-frequency item from the list,
and thus either type is equally likely to be recovered. Similarly, when a low-frequency item
is used to retrieve other list items, it is equally likely to sample either a high-frequency or a
low-frequency item from the list.
The explanation for the low-frequency advantage in recognition memory tests is quite
different. Here, there is an advantage for low-frequency items in both mixed and pure lists
because the distribution shapes and placements are determined by the frequency of the test
item rather than the frequency composition of the list. Because the strength of
pre-experimental associations is weaker between low-frequency items than between
high-frequency items, the residual inter-item parameters are lower for low-frequency items,
so the increase in familiarity from distractor to target distributions is larger for low-frequency
items, relative to the corresponding variances of the distributions. This difference in density
distributions means that there is a greater shift in the familiarity value for the presented
low-frequency items than for the presented high-frequency items. In short, unstudied
low-frequency items appear less familiar than unstudied high-frequency items, while studied
low-frequency items appear more familiar than studied high-frequency items. This leads to
a higher hit rate for critical low-frequency items and a lower false alarm rate for low-frequency
distractors (e.g., Glanzer & Bowles, 1976). Glanzer and colleagues dubbed the empirical
finding that for two classes of stimuli, the class with the higher hit rate also has a lower false
alarm rate  the mirror effect of recognition memory (Glanzer & Adams, 1985; 1990; Glanzer,
Adams & Iverson, 1991).
5
In summary, the SAM model predicts the following effects for word frequency:
1. Recognition will be better for low-frequency words than for high-frequency words
2. The mirror effect will occur in recognition
3. Recall of high-frequency words will be equal or worse than recall of low-frequency words
when they are presented in mixed lists
4. Recall of high-frequency words will be better than recall of low-frequency words when
they are presented in pure lists
The purpose of the present studies was to test the generality of these empirical effects by
determining whether they also occur for pictures of low and high familiarity. The first
experiment tests for recognition and recall of mixed lists to determine whether predictions
1, 2, and 3 are supported. The second experiment tests for recall of pure lists to determine
whether prediction 4 is supported.
Experiment 1
This experiment tests whether the word-frequency effect applies to pictures of high and low
familiarity. The pictures were presented in mixed lists in the study phase and tested for both
recognition and recall.
Method
Participants
Twenty-six participants were recruited from the population of undergraduate psychology
students at New York University. They were all native speakers of English with normal or
corrected vision.
Design
The experiment had a 2×2 within-subjects design. The independent variables were:
1. Memory task free recall versus yes/no recognition
2. Rated familiarity of pictures high versus low
The dependent variable in the recall task was the proportion of correctly recalled items.
The four dependent variables in the recognition task were hit and false alarm rates, the
measure of discriminability d , and the bias parameter C, i.e., the placement of the criterion
between the Yes and No responses. Arcsine transformations were performed on HIT and
6
FA rates, which were also corrected according to the recommendation by Snodgrass and
Corwin (1988).
Materials
Pictures of high and low familiarity were selected from the Snodgrass and Vanderwart
standardized pictures (1980) using the updated norms (Yoon et al., 2002). The concept
agreement of all items was above 90%, indicating that people generally agree well on the
names of the pictures. Average familiarity values for high-familiarity and low-familiarity
pictures were 4.78 and 2.56, with standard deviations of .19 and .45, respectively. Three lists
of items were created and matched with respect to number of semantic categories and category
members.
The three lists contained 20 high-familiarity pictures and 20 low-familiarity pictures each.
Two lists were used as critical items in the study lists; the third list constituted the 40 distractor
items in the recognition test. At the beginning and end of each study list, 5 filler items were
presented to reduce primacy and recency effects. These items were repeated across the two
lists. An additional 30 filler items were mixed randomly with the critical items of the study
list in the recognition task to reduce performance. Thus, 40 critical pictures and 40 filler
pictures were presented during the study-phase of the recognition task, 40 critical pictures
and 40 new items (distractors) were presented during the test phase of the recognition task,
and 40 critical pictures and 10 filler pictures were presented during the study-phase of the
recall task. The 40 distractor items were equally divided between high and low familiarity
and the 40 filler items were of medium familiarity according to the norms (Yoon et al., 2002).
Procedure
The experiment was conducted twice with each participant, with a 10-min break between
each session. During this break, the participant solved anagram puzzles. The orders of the
two study lists and the two memory tasks were counterbalanced. Participants were told in
advance to expect a recall or recognition test after each study phase, but they were not told
which one of the two they would receive. To increase motivation a $25 award was promised
to the person with the best performance. Each picture was presented for .05 s followed by
an inter-stimulus interval of .05 s. After the study phase, the participant was given five
arithmetic problems to further reduce recency effects.
7
In the recognition test, 40 critical items and 40 distractor items were presented sequentially,
and the participant was instructed to make a yes/no decision as to whether the item had been
in the most recent study list. In the recall test the participant was given 3 min to write down
the names of as many pictures as possible from the most recent study list. The participant
then typed the names into the computer. The names were later corrected for spelling mistakes
and naming variations.
Results and discussion
Mean accuracy in recall and recognition are presented in Table 1. We performed two analyses
of variance (ANOVAs) on the arcsine transformations of memory scores.
[Table 1. will appear here. See end of document.]
One analysis compared recall scores to hit rates in recognition, another compared recall
scores to correct rejection rates. In both cases there was a two-way interaction between
Memory task and Rated familiarity; F (1, 25) = 11.06, p < .01, and F (1, 25) = 4.90, p < .05.
The difference between high- and low-familiarity pictures in proportion of correct recall was
not significant; t (51) = 1.05, p > .10, and the difference between high- and low-familiarity
pictures in hit and correct rejection rates were both significant; t (51) = 3.91, p < .001, and
t (51) = 2.30, p < .05. This leads to a difference in discrimination (d ); t (51) = 4.49, p < .01,
indicating that the less familiar critical pictures were easier to recognize, and the less familiar
distractor pictures easier to reject, than the more familiar items. This is consistent with
previous findings that the frequency factor affects the accuracy of both types of decision (e.g.,
Glanzer & Bowles, 1976), i.e., the mirror effect was obtained for picture familiarity. The
difference in C value between the two levels of familiarity is not significant; t (51) = 1.70, p
= .09.
The effects of rated familiarity of pictures in the first study are consistent with the
word-frequency paradox and the SAM model for recall and recognition (Gillund & Shiffrin,
1984). An advantage occurs for low-familiarity pictures in recognition, while no difference
is found in recall when high- and low-familiarity pictures are presented in mixed study lists.
These results are consistent with predictions 1, 2, and 3 of the SAM model as described in
the introduction. The second experiment will test prediction 4 of the SAM model; i.e., whether
recall is better for high-familiarity than low-familiarity pictures when they are presented in
pure lists.
8
Experiment 2
In this experiment pictures of high and low familiarity are presented in separate blocks and
tested for recall performance only. The SAM model predicts better recall of high- than
low-familiarity pictures in this paradigm.
Method
Participants
Twenty-five participants were recruited from the population of undergraduate psychology
students at New York University. They were all native speakers of English with normal or
corrected vision.
Design
The experiment was run according to a within-subjects design. The independent variable was
rated familiarity of pictures high versus low. The dependent variable was the proportion
of correctly recalled items.
Materials
The set of pictures employed in this experiment was identical to that used in Experiment 1.
Four lists of pictures were created and matched with respect to number of semantic categories
and category members. Two lists were composed of high-familiarity pictures, two lists were
composed of low-familiarity pictures, and each list contained 20 pictures. At the beginning
and end of each list, five additional pictures were presented as filler items to reduce primacy
and recency effects. These changed across lists and were similar to the other items with
respect to familiarity.
Procedure
The experiment consisted of four study-test sequences, with a 2-min break between each
sequence. During the break, the participant solved anagram puzzles. Participants were told
in advance to expect an unspecified memory test after each study phase. To increase
motivation a $25 award was promised to the person with the best performance. The order
of the four lists was counterbalanced. Each picture within a list was presented for .5 s and
9
followed by an inter-stimulus-interval of .5 s. After the study phase, the participant was given
five arithmetic problems to further reduce recency effects.
In the recall test the participant was requested to write down the names of as many pictures
as possible within a time limit of 3 min. The participant then typed the names into the
computer and the names were later corrected for spelling mistakes and naming variations.
Results and discussion
The mean proportion of low-familiarity and high-familiarity pictures recalled correctly was
.25 and .36 respectively, with standard deviations of .10 and .11. The statistical comparison
on the arcsine transformations of the memory scores showed that the high-familiarity
advantage was significant; t (24) = 6.03, p < .001, indicating that when pictures were blocked
with respect to familiarity, the more familiar pictures were better recalled than the less
familiar pictures. The finding that picture-familiarity is positively related to the recallability
of pure lists is analogous to the advantage previously established for high-frequency words
(Bousfield & Cohen, 1955; Deese, 1960; Hall, 1954, 1985; Murdock, 1960; Raymond, 1969;
Schulman, 1976; Sumby, 1963).
General discussion
The classical word-frequency paradox in recall and recognition involves four empirical results
that need to be explained:
1. Recognition is better for low-frequency words than for high-frequency words
2. The mirror effect occurs in recognition
3. Recall of high-frequency words is equal to or worse than recall of low-frequency words
presented in mixed lists
4. Recall is better for pure lists of high-frequency words than for pure lists of low-frequency
words
The current study indicates that all four effects also apply to pictures whose frequency in
daily life is measured by ratings of familiarity. To our knowledge this is the first demonstration
that the recall/recognition paradox also applies to pictures. An interpretation of the finding
is that an explanation of the word-frequency effect should hold for the effects of picture
familiarity. The SAM model by Gillund and Shiffrin (1984) provides one such explanation
that can be extended to picture memory. According to the SAM model, the strength of the
10
total amount of information activated by an item and context is used to make a familiarity
decision in recognition. In recall the same information is used to determine sampling and
recovery during an extended search.
The fact that the dependent measure in our recall tasks was the written names of pictures
represents one important limitation in the experimental design. The correlation between
ratings of familiarity and frequency of picture names is known to be positive (Snodgrass &
Vanderwart, 1980), and consequently it is difficult to interpret whether our recall data are
due to picture familiarity or the frequency of picture names. The more strenuous and
time-consuming task of drawing the pictures as a recall response is also likely to be mediated
by verbal responses, i.e.,  I saw an apple, so I ll draw an apple , but although it may be difficult
to avoid verbal responses in the test phase, certain steps may be taken to prevent (implicit)
naming of pictures during study.
The word-frequency paradox is restricted to items of sufficient familiarity and does not
occur with extremely low-frequent words or nonwords (Mandler et al., 1982; Rao, 1983;
Schulman, 1976; Zechmeister, Curt, & Sebastian, 1978). One possible interpretation of the
current findings with pictures is that the frequency/familiarity paradox forms a general
phenomenon of human memory associated with any stimulus material that activates sufficiently
specific representations in long-term memory. If the paradox is indeed a general effect of
frequency/familiarity, it should be present in other domains than visual presentation, on
which all studies of the paradox in the current literature seem to rely. An intrinsic problem
with this interpretation, however, is that such material is likely to be easily named; thus it
may be difficult to disentangle the effects of linguistic frequency from familiarity or frequency
per se.
In summary, these studies have shown that memory for pictures and memory for words
share important empirical effects: that the word-frequency paradox in recall and recognition
also occurs as the picture-familiarity paradox in recall and recognition, and that both show
four important characteristics predicted by the SAM model. It remains to be seen whether
a model of memory other than SAM can account for these parallel results.
Acknowledgements The authors gratefully acknowledge the assistance of Clifford A. Larochel,
Andrea J. Ardouin, and Rossanna Lee in testing the participants.
11
References
Allen, L. R., & Garton, R. F. (1968). The influence of word knowledge on the word-frequency effect
in recognition memory. Psychonomic Science, 10, 401 402.
Balota, D. A., & Neely, J. H. (1980). Test-expectancy and word-frequency effects in recall and
recognition. Journal of Experimental Psychology: Human Learning and Memory, 6, 576 587.
Battig, W. F., & Montague, W. E. (1969). Category norms for verbal items in 56 categories: A replication
and extension of the Connecticut category norms. Journal of Experimental Psychology Monograph,
80 (3, Pt. 4).
Bousfield, W. A., & Cohen, B. H. (1955). The occurrence of clustering in the recall of randomly
arranged words of different frequencies-of-usage. Journal of General Psychology, 52, 83 95.
Clark, S. E., & Burchett, R. E. R. (1994). Word frequency and list composition effects in associative
recognition and recall. Memory and Cognition, 22, 55 62.
Dahl, H. (1979). Word frequency of spoken American English. Essex, CT: Verbatim.
Deese, J. (1960). Frequency of usage and number of words in free recall: The role of association.
Psychological Reports, 7, 337 344.
Delosh, E. L., & McDaniel, M. A. (1996). The role of order information in free recall: Application
to the word-frequency effect. Journal of Experimental Psychology: Learning, Memory, and Cognition,
22, 1136 1146.
Dewhurst, S. A., Hitch, G. J., & Barry, C. (1998). Separate effects of word frequency and age of
acquisition in recognition and recall. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 24, 284 298.
Dobbins, I. G., Kroll, N. E. A., Yonelinas, A. P., & Liu, Q. (1998). Distinctiveness in recognition and
free recall: The role of recollection in the rejection of the familiar. Journal of Memory and Language,
38, 381 400.
Duncan, C. P. (1974). Retrieval of low-frequency words from mixed lists. Bulletin of the Psychonomic
Society, 4, 137 138.
Gillund, G., & Shiffrin, R. M. (1984). A retrieval model for both recognition and recall. Psychological
Review, 91, 1 67.
Glanzer, M., & Adams, J. K. (1985). The mirror effect in recognition memory. Memory and Cognition,
13, 8 20.
Glanzer, M., & Adams, J. K. (1990). The mirror effect in recognition memory: Data and theory.
Journal of Experimental Psychology: Learning, Memory, and Cognition, 16, 5 16.
Glanzer, M., & Bowles, N. (1976). Analysis of the word-frequency effect in recognition memory.
Journal of Experimental Psychology: Human Learning and Memory, 2, 21 31.
Glanzer, M., Adams, J. K., & Iverson, G. (1991). Forgetting and the mirror effect in recognition
memory: Concentering of underlying distributions. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 17, 81 93.
Gorman, A. M. (1961). Recognition memory for nouns as a function of abstractness and frequency.
Journal of Experimental Psychology, 61, 23 29.
Gregg, V. H., Montgomery, D. C., & Castano, D. (1980). Recall of common and uncommon words
from pure and mixed lists. Journal of Verbal Learning and Verbal Behavior, 19, 240 245.
Guttentag, R. E., & Carroll, D. (1994). Identifying the basis for the word frequency effect in recognition
memory. Memory, 2, 255 273.
Guttentag, R. E., & Carroll, D. (1997). Recollection-based recognition: Word frequency effects.
Journal of Memory and Language, 37, 502 516.
Hall, J. F. (1954). Learning as a function of word-frequency. American Journal of Psychology, 67,
138 140.
Hall, J. F. (1985). Free recall as a function of type of encoding and word frequency. Bulletin of the
Psychonomic Society, 23, 368 370.
12
Kinsbourne, M., & George, J. (1974). The mechanism of the word-frequency effect on recognition
memory. Journal of Verbal Learning and Verbal Behavior, 13, 63 69.
Kintsch, W. (1970). Learning, memory and conceptual processes. New York: Wiley.
Kucera, H., & Francis, W. N. (1967). Computational analysis of present-day American English.
Providence, RI: Brown University Press.
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.
Mandler, G., Goodman, G. O., Wilkes-Gibbs, D. L. (1982). The word-frequency paradox in recognition.
Memory and Cognition, 10, 33 42.
May, R. B., & Tryk, H. E. (1971). Word sequence, word frequency, and free recall. Canadian Journal
of Psychology, 24, 299 304.
May, R. B., Cuddy, L. J., & Norton, J. M. (1979). Temporal contrast and the word frequency effect.
Canadian Journal of Psychology, 33, 141 147.
McCormack, P. D., & Swenson, A. L. (1972). Recognition memory for common and rare words.
Journal of Experimental Psychology, 95, 72 77.
Murdock, B. B., Jr. (1960). The immediate retention of unrelated words. Journal of Experimental
Psychology, 60, 222 234.
Otani, H., & Whiteman, H. L. (1993). Word frequency effect: A test of processing-based explanation.
Psychological Record, 43, 317 327.
Raaijmakers, J. G., & Shiffrin, R. M. (1981). Search of associative memory. Psychological Review,
88, 93 134.
Rao, K.V. (1983). Word frequency effect in situational frequency estimation. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 9, 73 81.
Raymond, B. (1969). Short-term storage and long-term storage in free recall. Journal of Verbal
Learning and Verbal Behavior, 8, 567 574.
Rosch, E., Mervis, C. B., Gray, W. D., Johnson, D. M., & Boyes-Braem, P. (1976). Basic objects in
natural categories. Cognitive Psychology, 8, 382 439.
Schulman, A. I. (1967). Word length and rarity in recognition memory. Psychonomic Science, 9, 21 22.
Schulman, A. I. (1976). Memory for rare words previously rated for familiarity. Journal of Experimental
Psychology: Human Learning and Memory, 2, 301 307.
Schulman, A. I., & Lovelace, E. A. (1970). Recognition memory for words presented at a slow or rapid
rate. Psychonomic Science, 21, 99 100.
Shepard, R. N. (1967). Recognition memory for words, sentences, and pictures. Journal of Verbal
Learning and Verbal Behavior, 6, 156 163.
Snodgrass, J. G., & Corwin, J. (1988). Pragmatics of measuring recognition memory: Applications to
dementia and amnesia. Journal of Experimental Psychology: General, 117, 34 50.
Snodgrass, J. G., & Vanderwart, M. (1980). A standardized set of 260 pictures: Norms for name
agreement, image agreement, familiarity, and visual complexity. Journal of Experimental Psychology:
Human Learning and Memory, 6, 174 215.
Sumby, W. H. (1963). Word frequency and serial position effects. Journal of Verbal Learning and
Verbal Behavior, 1, 443 450.
Underwood, B. J., & Freund, J. S. (1970). Word frequency and short-term recognition memory.
American Journal of Psychology, 83, 343 351.
Van Overschelde, J. P. (2002). The influence of word frequency on recency effects in directed free
recall. Journal of Experimental Psychology: Learning, Memory, and Cognition, 28, 611 615.
Wallace, W. P., & Page, K. A. (1982). Recognition test trials without distractors: A comparison of test
trials and study trials on recognition and recall. Bulletin of the Psychonomic Society, 20, 245 248.
Wallace, W. P., Sawyer, T. J., Robertson, L. C. (1978). Distractors in recall, distractor-free recognition,
and the word-frequency effect. American Journal of Psychology, 91, 295 304.
13
Ward, G., Woodward, G., Stevens, A., & Stinson, C. (2003). Using overt rehearsals to explain word
frequency effects in free recall. Journal of Experimental Psychology: Learning, Memory, & Cognition,
29, 186 210.
Watkins, M. J., LeCompte, D. C., & Kim, K. (2000). Role of study strategy in recall of mixed lists of
common and rare words. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26,
239 245.
Whiteman, H. L., & Nairne, J. S., & Serra, M. (1994). Recognition and recall-like processes in the
long-term reconstruction of order. Memory, 2, 275 294.
Yoon, C., Luo, T., Mikels, J. A., Hedden, T., Gutchess, A. H., Jiao, S., & Park, D. C. (2002). Picture
norming: A cross-culturally standardized set of 260 pictures in older and younger adults: American
and Chinese norms for name agreement and familiarity. Internet site of the University of Michigan
Cognition and Aging Lab. Retrieved February, 2002, from http://agingmind.isr.umich.edu/ca4c.html
Zechmeister, E. B., Curt, C., & Sebastian, J. A. (1978). Errors in recognition memory task are a
U-shaped function of word frequency. Bulletin of the Psychonomic Society, 11, 371 373.
14
Table 1. Mean proportion recalled and recognized for mixed lists of low- and high-familiarity pictures in Experiment 1. Hit = mean proportion of hits, CR = mean proportion
of correct rejections. The discriminability index d and the criterion C are also shown
Low High
M SD M SD
Recall 0.28 0.12 0.29 0.12
Recognition
Hit 0.83 0.14 0.77 0.13
CR 0.79 0.14 0.76 0.10
d 2.00 1.58
C  0.08  0.03
15