The Effect of Exposure Frequency on Intermediate Language Learners' Incidental Vocabulary Acquisition and Retention through Reading

Abstract

Research has been investigating the role of reading, as one source of input, in language
learners' vocabulary development. The present study was designed to examine whether
intermediate learners incidentally (a) acquire and (b) retain unknown vocabulary as a result of
reading. The study further assessed (c) the effect of the text variable of exposure frequency.
Learners were exposed to unfamiliar words either two, four, or six times during reading.
Vocabulary acquisition and retention measured productive and receptive knowledge gain.
Results indicated that only two encounters with unfamiliar words during reading significantly
affected learners' vocabulary growth. Moreover, two or four exposure frequencies resulted
in fairly similar word gain, but six exposures produced significantly more vocabulary knowledge.
Retention measures showed mixed results: On productive vocabulary knowledge only half of the
subjects displayed a significant rate of retention. On receptive knowledge all but one
experimental group retained vocabulary over 4 weeks.

Full text

SSLA, 21, 589–619. Printed in the United States of America.
THE EFFECT OF EXPOSURE
FREQUENCY ON INTERMEDIATE
LANGUAGE LEARNERS’
INCIDENTAL VOCABULARY
ACQUISITION AND RETENTION
THROUGH READING
Susanne Rott
University of Illinois at Chicago
Research has been investigating the role of reading, as one source
of input, in language learners’ vocabulary development. The present
study was designed to examine whether intermediate learners inci-
dentally (a) acquire and (b) retain unknown vocabulary as a result of
reading. The study further assessed (c) the effect of the text variable
of exposure frequency. Learners were exposed to unfamiliar words
either two, four, or six times during reading. Vocabulary acquisition
and retention measured productive and receptive knowledge gain.
Results indicated that only two encounters with unfamiliar words dur-
ing reading significantly affected learners’ vocabulary growth. More-
over, two or four exposure frequencies resulted in fairly similar word
gain, but six exposures produced significantly more vocabulary
knowledge. Retention measures showed mixed results: On productive
vocabulary knowledge only half of the subjects displayed a significant
rate of retention. On receptive knowledge all but one experimental
group retained vocabulary over 4 weeks.
The expectation of furthering foreign and second language learners’ (hencefor-
ward L2 learners) lexical growth by engaging them in extensive reading (e.g.,
Brown, 1994; Day & Bamford, 1998; Krashen, 1993; Nation, 1990) is based on
I would like to thank Jim Lee, John Lalande II, David Weible, and the anonymous SSLA reviewers for
their valuable comments.
Address correspondence to Susanne Rott, Department of German (MC 189), University of Illinois
at Chicago, 601 S. Morgan, Chicago, IL 60607; e-mail: srott@uic.edu.
 1999 Cambridge University Press 0272-2631/99 $9.50
589

590 Susanne Rott
input-oriented language acquisition theory. It holds that language develop-
ment takes place when learners are involved in the processing of meaningful
and contextualized input integrating new forms into their L2 language systems
by making form-meaning connections (e.g., Lee & VanPatten, 1995). The nu-
merous benefits of reading as a source of input are apparent. Extensive read-
ing potentially provides learners with the opportunity to process an
unfamiliar word in its various natural contexts in order to acquire the com-
plex properties of the lexical item, such as its polysemic and registerial as-
pects (Judd, 1978) as well as its syntactic, lexical, and semantic properties
(e.g., Nation, 1990; Richards, 1976). Reading is furthermore considered an es-
sential source for the acquisition of less frequent lexical items (Coady, 1993,
1997; Ellis, 1994) that are usually encountered in texts only; yet these words
are part of advanced and superior learners’ lexicons. Similarly, Krashen (1989)
suggested that a substantial part of the L2 lexicon is gained through reading
because the large vocabulary body of a superior or near native speaker can-
not be learned explicitly through word lists or by looking up words in the dic-
tionary. Further comparing reading with oral input, Ellis described reading as
the “ideal medium” for vocabulary acquisition. Print material allows learners
more time for processing a new word, “whereas in speech it passes ephemer-
ally” (1994, p. 40). Concerning the learning context, it is evident that, particu-
larly for learners in a foreign language setting, as compared to a second
language setting, reading is a primary source for processing new forms be-
cause the main access to target language (TL) sources is often limited to print
material.
The importance of reading for L2 learners’ lexical development is generally
acknowledged by researchers and theorists and has been repeatedly verified
by studies that found that learners incidentally gained a statistically signifi-
cant amount of unfamiliar words while reading text for global comprehension
(Day, Omura, & Hiramatsu, 1991; Dupuy & Krashen, 1993; Hulstijn, 1992; Huls-
tijn, Hollander, & Greidanus, 1996; Joe, 1995; Knight, 1994; Luppescu & Day,
1995; Paribakht & Wesche, 1996; Pitts, White, & Krashen, 1989). However, an
ongoing debate on the effectiveness of reading for L2 learners’ vocabulary de-
velopment has pinpointed pedagogical and methodological shortcomings of
research to date (e.g., Huckin & Haynes, 1993; Hulstijn et al. 1996; So
¨
kmen,
1997). Describing lexical growth as a “by-product” (Diakidoy, 1993) of reading
or as “accidental learning of information without intention of remembering
that information” (Hulstijn et al., 1996, p. 327) deems word gain through read-
ing an unpredictable process that cannot be influenced either through instruc-
tion and the selection of materials or through learners’ approach to the
reading task.
1
Such a negative assessment is understandable considering that,
to date, SLA research that focused explicitly on incidental word gain through
reading (Day et al., 1991; Dupuy & Krashen, 1993; Pitts et al., 1989) has failed
to explain, describe, and account for factors that favorably affected the condi-
tion in which incidental word learning took place. But in order to be able to
increase the likelihood of incidental word gain during reading, research needs

Exposure Frequency, Acquisition, and Retention 591
to determine the learner and text factors that have an impact on the acquisi-
tion process.
The present study was conducted to (a) shed more light on the effective-
ness of reading for incidental word gain by examining its effect on retention
of word knowledge and (b) further explore the effect of the text variable expo-
sure frequency on the word learning process.
REVIEW OF LITERATURE
Three quantitative investigations that exclusively focused on incidental vocab-
ulary acquisition while reading text for meaning (Day et al., 1991; Dupuy &
Krashen, 1993; Pitts et al., 1989) demonstrated that intermediate learners ac-
quired a small but significant amount of words through reading. Dupuy and
Krashen and Day et al. established reading as a source for lexical gain by com-
paring word knowledge of learners who had encountered the chosen target
words (TW) while reading to word knowledge of learners of a control group
who had not encountered the TWs in text. But neither of the studies used a
pretest-posttest design to ensure that subjects were not familiar with any of
the TWs. Dupuy and Krashen used colloquialisms as TWs “which were proba-
bly not familiar to intermediate students of French” (1993, p. 57; emphasis
added). Day et al. identified their TWs through a pilot study using “words that
the subjects did not know or found difficult” (1991, p. 272; emphasis added).
Moreover, Pitts et al. (1989) established their findings neither through com-
parison with a control group nor through a pretest-posttest design but as-
sumed that the TWs of their study (Russian slang words) could have been
encountered by their subjects only during the reading treatment.
In all three studies, word knowledge gain was measured immediately after
learners had read the authentic or modified input passages. None of the stud-
ies assessed long-term effects of reading for lexical growth. In fact, Day and
colleagues (1991) questioned long-term knowledge gain of words encountered
in only one text even though the text might provide multiple encounters. Con-
sidering the incidental learning process during reading as an accumulation of
instances of a variety of contexts, the researchers suggested that only re-
peated exposure through extensive reading leads to integration of new words
into the L2 lexicon. Moreover, research results of each of the three studies
were based on only one multiple-choice task assessing receptive word knowl-
edge. The researchers themselves pointed out that their instrumentation was
not sensitive to partial knowledge gain (Dupuy & Krashen, 1993), which might
have underscored their findings. These studies furthermore did not tap
productive word knowledge to develop a more complete picture of the effec-
tiveness of reading for vocabulary growth. Unfortunately, none of the investi-
gations accounted for any learner variables or for textual features of the input
passages that might have affected word acquisition positively or negatively.
The researchers reported only the range of TW occurrences in the texts read
by learners but did not examine the specific effect of multiple exposures on

592 Susanne Rott
learning. Dupuy and Krashen (1993) reported a frequency ranging between
one and six times, in the study by Pitts et al. (1989) TW exposure ranged from
1 to 27, and Day et al. reported that the TWs occurred with “ample frequen-
cies and in sufficient context to allow for reasonable guesses about meaning”
(1991, p. 543).
Other studies (Hulstijn et al., 1996; Knight, 1994) that reconfirmed that inci-
dental word acquisition occurs during reading, but whose main focus was to
assess the effectiveness of vocabulary enhancement techniques, provide some
insight into variables that may have an impact on incidental word gain. Huls-
tijn et al., who controlled the text variable of exposure frequency, observed
that multiple exposures often failed to have a positive impact on acquisition.
Although learners more readily recognized words they had encountered three
times during reading than words they had encountered only once, they often
were not able to infer the correct meaning at any of the three encounters.
Knight further found that high verbal ability learners gained significantly more
receptive as well as productive vocabulary knowledge than did low verbal
ability learners. Both research studies were also concerned with the retention
of words acquired during reading. In contrast to Day and colleagues’ (1991)
suggestion mentioned above, results revealed that readers had retained a sig-
nificant amount of word knowledge 1 hour (Hulstijn et al., 1996) and 2 weeks
(Knight, 1994) after the reading encounter. Knight further recorded an in-
crease in productive vocabulary knowledge after 2 weeks. She explained this
delayed word knowledge gain with the intensive learner-text interaction taking
place during the meaning assignment process.
Although both studies reported incidental word gain through reading, they
further showed that a combination of vocabulary enhancement techniques
with reading yielded significantly more word gain than reading alone. In fact,
an increasing number of investigations have found superior word gain through
vocabulary enhancement techniques. This line of research makes use of the
benefits of learning words by processing L2 text, as outlined in the introduc-
tion, and addresses and accounts for the problems that learners encounter
when trying to assign meaning to unfamiliar words during reading (for de-
tailed overviews, see Huckin & Haynes, 1993; Hulstijn et al., 1996; So
¨
kmen,
1997).
The effectiveness of reading as an important source of L2 vocabulary devel-
opment has been questioned because a number of learner and text variables
have been identified that aid, but more often impede, beginning L2 readers’
word inferencing process. The four major factors that can have an impact on
the outcome of inferencing are: (a) learners’ knowledge about the linguistic
properties of an unknown word (Bensoussan & Laufer, 1984; Haynes, 1993;
Na & Nation, 1985), (b) context properties in which the unknown word ap-
pears (Bensoussan & Laufer, 1984; Haynes, 1993; Huckin & Bloch, 1993; Mon-
dria & Wit-DeBoer, 1991), (c) the approach taken by the language learner to
infer meaning (Chern, 1993; Lee & Wolf, 1997; Walker, 1983), and (d) cognitive

Exposure Frequency, Acquisition, and Retention 593
processes that influence L2 readers’ awareness of and attention to unfamiliar
words (Coady, 1993, 1997; Ellis, 1994; Perry, 1993; Schmidt, 1995).
Recently conducted vocabulary studies have responded to these difficul-
ties. They aimed at enhancing and simplifying word inferencing to speed up
the vocabulary acquisition process through dictionary access (Hulstijn et al.,
1996; Knight, 1994; Luppescu & Day, 1995), glosses (Hulstijn, 1992; Hulstijn et
al., 1996; Watanabe, 1997), or postreading vocabulary activities (Paribakht &
Wesche, 1996, 1997; Zimmerman, 1997). Taken together, the findings of these
studies do not provide conclusive evidence, but they do strongly suggest that
vocabulary growth through reading can be increased by providing L2 readers
with a variety of enhancement techniques. However, even though the studies
indicate that enhancement techniques result in more vocabulary knowledge
than does incidental acquisition during normal reading, the pedagogically rele-
vant question is whether such enhancement techniques can account for the
extensive vocabulary corpus of a near-native speaker; that is, how many vo-
cabulary enhancement exercises can we engage L2 learners in during three to
four 50-minute sessions per week, how many glossed texts can we provide,
and how much can we rely on students reading short texts at the computer
with dictionary access?
A further issue has been raised by Nation and Warig (1997), who stressed
the importance of assessing learners’ vocabulary needs to establish a realistic
goal in terms of quantity and kind of vocabulary to be taught and targeted for
learners to know. The researchers posited that 3,000–5,000 word families are
a good basis for text comprehension. But if L2 learners aspire to attain higher
levels of language abilities, the targeted vocabulary size is much larger, al-
though it is smaller than the previously suggested 40,000 lexical items (Ander-
son & Nagy, 1992). Recent studies have shown that an educated American
native speaker’s lexicon ranges between 17,000 and 20,000 base words
(D’Anna, Zechmeister, & Hall, 1991; Goulden, Nation, & Read, 1990; Zechmeis-
ter et al.1993). Hazenberg and Hulstijn (1996) further established that, in order
to successfully comprehend academic texts, learners need to have a lexical
body of about 11,000 base words, at least for Dutch texts. Looking at basic
foreign language courses in the United States that are filled with a large num-
ber of students who focus only on fulfilling their language requirement, a basic
vocabulary of 3,000–5,000 words gained through reading in combination with
enhancement activities might be a realistic goal.
2
But in the same classes are
learners who are (hopefully) planning to pursue academic studies in the for-
eign language. These learners, therefore, require a larger and broader lexicon
after only 3 or 4 years of language study.
The logical hypothesis underlying the present investigation is that L2 learn-
ers need to be engaged in reading under both circumstances: reading for
meaning under an enhanced condition, to ensure a basic lexicon, as well as
under a normal condition, to advance beyond the basic requirement. There-
fore, the present study was set up to gain further insight into long-term effects

594 Susanne Rott
of exposing learners to text while controlling for the number of encounters
with unfamiliar words. The aim of this line of research is to influence and es-
tablish conditions conducive to increasing the rate of vocabulary growth in a
normal reading situation.
This research study investigated the following questions:
1. Do intermediate foreign language learners acquire vocabulary incidentally through
reading?
2. If acquired, is vocabulary retained over a longer period of time: 1 week later and
1 month later?
3. Does frequency of exposure to a lexical item have an effect on incidental acquisi-
tion and retention of vocabulary?
METHOD
Subjects
Subjects participating in this study were 95 learners of German as a foreign
language enrolled in fourth-semester classes at the University of Illinois. Four
treatment groups came from the Urbana-Champaign campus and two treat-
ment groups came from the Chicago campus. All students enrolled in fourth-
semester language classes were initially included in the study. However, 28
subjects’ data had to be excluded from the final analysis because they were
either familiar with the TWs or did not complete all treatments or measures.
Materials
The 12 lexical items selected as TWs for the present study involved items and
actions of everyday life and did not entail new or target culture-specific con-
cepts for learners. Two sets (set I and set II) of TWs, each with three nouns
and three verbs, were created (Appendix A). Each set was distributed to a
different experimental group to limit the quantity of reading to six paragraphs
per treatment day. This procedure furthermore allowed comparing acquisition
and retention across the two-word sets for greater generalizability of the re-
sults.
For each of the 12 TWs, six different paragraphs were written, one for each
treatment day. For the creation of the reading passages the following criteria
were applied: (a) The paragraphs were each 4–6 sentences long, that is, each
treatment package contained between 24 and 36 sentences. The individual
paragraphs for each TW were kept short because learners were asked to im-
mediately recall the content of each passage read in as much detail as possi-
ble. This procedure ensured that learners focused on text comprehension
while reading and it allowed measurement of reading comprehension. The re-
sults of the effect on reading comprehension are reported elsewhere (Rott,
1997). (b) The ideas in each paragraph were developed clearly, with concrete

Exposure Frequency, Acquisition, and Retention 595
subjects and situations. (c) No synonyms or explicit definitions of the TW
were used, but the context provided sufficient clues to infer meaning. This
had been tested with native speakers. (d) Each paragraph presented the TW
in a different context, so that learners received, with each paragraph, a new
opportunity to make an inference. The TW Bude (“apartment”), for example,
was mentioned in the context of moving out from home, looking for the right
place, describing the furniture that is needed, a party in the new apartment,
cleaning the apartment, and parents visiting the apartment (Appendix B). For
variation, the TWs appeared in the beginning, in the middle, or at the end of
the different passages. (e) The chosen TWs concerned key issues in the input
passages. For the reading comprehension analysis reported in Rott (1997), the
text had been divided into two levels (1 and 2) of weighed idea units (Bern-
hardt, 1991), which made a distinction between more (level 2) and less (level
1) important propositions. Idea units containing the TW received a level-2 cat-
egorization. (f) To make reading more meaningful for learners, a contextual
setting was created for the TWs. The paragraphs were introduced as the diary
entries of the German student Andreas writing about everyday situations.
Testing Materials
Vocabulary Checklist Test.
A vocabulary pretest in the form of a checklist
was developed to verify that the selected TWs were unfamiliar to all subjects.
Learners received a list of 50 lexical items (12 TWs and 38 distracters). They
were instructed to attempt a translation of each of the words and skip only
those items they had never encountered before. This test format, which had
been successfully used in other studies (e.g., Knight, 1994), accounted for
learners’ partial word knowledge.
Vocabulary Acquisition and Retention Measures.
To assess word acquisi-
tion and retention, two types of vocabulary tests were administered in order
to measure different levels of word learning, to prevent a task effect, and to
receive more generalizable data. For these reasons, learners performed both
a recognition and a production task. First, subjects encountered the TWs de-
contextualized in the form of a list. This list consisted of the 12 TWs and 8
distracters. Distracters also involved actions and things encountered in daily
life, as did the TWs. To assess full or partial acquisition and retention, learn-
ers were asked to supply the meaning of all words in their native language
(English). Next, learners engaged in a select-definition task. This task also con-
tained the 12 TWs and 8 distracters as in the supply-definition task. Subjects
were requested to select the correct definition of the word from four possible
meanings and a “don’t know” option. The four multiple-choice answers, pro-
vided in English, were written to emphasize semantic difference. For example,
for the TW Glotze (“television set”) learners received the following choices:
(a) word processor, (b) television set, (c) garbage disposal, (d) cellular phone,
or (e) don’t know.

596 Susanne Rott
To ensure that reading for meaning would be the source of the subjects’
vocabulary growth, they were tested on the TWs of both word sets—that is,
they were tested for acquisition and retention of words they had encountered
during reading as well as words they had not encountered during reading.
That way, the group that had encountered the TWs of set I in reading pas-
sages functioned as a control group for TW set II and the group that had en-
countered TWs of set II in reading passages functioned as a control group for
TW set I.
Because learners received the vocabulary tests three times—namely, im-
mediately after reading (acquisition), after 1 week (retention 1), and after 4
weeks (retention 2)—vocabulary items were presented in a different order
each time. During all three administrations, subjects first completed the sup-
ply-definition task and then the select-definition task so that they would not
copy answers or glean hints from the multiple-choice test onto the supply-
definition measure.
Scoring
For the supply-definition task, two independent raters (the researcher and a
graduate teaching assistant) scored the definitions according to how accu-
rately subjects could define the TW. Two points were allotted for an entirely
correct answer, one point for a definition showing partial knowledge, and zero
points for an incorrect answer. The raters were in full agreement in all in-
stances. The select-definition multiple-choice task was scored as follows: A
correct answer received two points and an incorrect answer received zero
points.
Study Design
Subjects were divided in two treatment groups, one for word set I and one for
word set II. Each treatment group was further divided into three groups of
two, four, or six TW exposure frequencies (Figure 1). Each experimental group
was randomly assigned to one treatment condition and consisted of an intact
class. No true randomization of treatment conditions was possible because of
the variation in treatment and testing schedules for the two-, four-, and six-
exposure conditions.
The study was conducted over a period of 13 weeks (Table 1). Data collec-
tion took place during regular class time. During week 1, subjects were asked
to participate in a study on second language acquisition and signed a partici-
pation agreement. In week 2 a vocabulary checklist test was administered to
ensure that the TWs were unfamiliar to all participants. Treatment and post-
tests took place during weeks 4–13. During each treatment session instructors
distributed a folder containing the day’s materials to each subject. Six reading
passages, each containing one of the six different TWs, were printed on sepa-
rate pages. After each reading passage, one extra page was added for immedi-

Exposure Frequency, Acquisition, and Retention 597
Figure 1. Design of the study.
ate written recall of the content. This process was repeated during the
ensuing weeks, once each week: for a period of 6 weeks for the six-exposure
groups, 4 weeks for the four-exposure groups, and 2 weeks for the two-expo-
sure groups. Each exposure group (two, four, or six exposures) received three
posttests: the first on the last day of treatment to measure acquisition, and
the second and third 1 week and 1 month, respectively, after the last day of
treatment to measure retention.
RESULTS
The descriptive statistics (means, standard deviations, and number of sub-
jects) of acquisition (immediately after reading), retention 1 (1 week after
reading), and retention 2 (4 weeks after reading) scores for the supply-defini-
tion scores are reported in Table 2. Table 3 reports descriptive statistics for
the select-definition scores.
The statistical analyses of the research questions were based on two re-

598 Susanne Rott
Table 1. Time frame of the study
Treatment Groups
Set I Set II
Week
number 2 Exposure 4 Exposure 6 Exposure 2 Exposure 4 Exposure 6 Exposure
1 Participation agreement for all groups
2 Vocabulary checklist test for all groups
3——————
4XXXXXX
5 X, A X X X, A X X
6RXXRXX
7 — X, A X — R X, A
8— R X — R X
9 R — X, A R — X, A
10 — — R — — R
11 — R — — R —
12——————
13 — — R — — R
Note. X = treatment with reading passage, A = test for acquisition, R = test for retention.
Table 2. Means and standard deviations of 0-, 2-, 4-,
and 6-exposure supply-definition measure scores for
acquisition and retention
Acquisition Retention 1 Retention 2
Condition n M SD M SD M SD
Set I
0 Exposure 50 0.00 0.00 0.00 0.00 0.40 0.28
2 Exposure 12 2.42 1.68 3.17 1.85 2.33 1.50
4 Exposure 17 2.71 2.34 2.24 2.33 1.82 1.67
6 Exposure 16 5.38 3.40 4.75 3.22 4.44 2.73
Set II
0 Exposure 45 0.00 0.00 0.14 0.51 0.23 0.64
2 Exposure 18 2.28 1.64 2.33 2.40 1.44 1.42
4 Exposure 16 4.06 3.90 3.81 2.81 4.18 2.54
6 Exposure 16 5.38 2.19 6.00 2.90 4.00 2.53
Note. 12 possible points.
peated-measures ANOVAs with a mixed 2 (word set) × 3 (time) × 4 (exposure
frequency) design. The first ANOVA analyzed select-definition task vocabulary
scores, and the second ANOVA analyzed supply-definition task vocabulary
scores (dependent variables). The independent variables were word set (I and
II), time of vocabulary assessment (immediately after reading, 1 week later,
and 4 weeks later), and exposure frequency (zero, two, four, or six expo-
sures). For all statistical analyses the alpha level was set at .05. To account

Exposure Frequency, Acquisition, and Retention 599
Table 3. Means and standard deviations of 0-, 2-, 4-,
and 6-exposure select-definition measure scores for
acquisition and retention
Acquisition Retention 1 Retention 2
Condition n M SD M SD M SD
Set I
0 Exposure 50 0.59 0.92 1.02 1.09 1.53 1.10
2 Exposure 12 6.67 3.34 6.33 2.94 6.33 2.81
4 Exposure 17 5.65 2.47 5.41 2.09 5.41 2.43
6 Exposure 16 8.13 2.60 8.13 2.87 7.38 3.24
Set II
0 Exposure 45 1.09 1.33 1.73 1.34 2.36 1.38
2 Exposure 18 5.44 2.36 5.33 2.47 4.56 2.26
4 Exposure 16 6.50 2.78 6.13 1.71 6.00 2.07
6 Exposure 16 8.38 2.34 9.13 1.93 7.88 2.12
Note. 12 possible points.
Table 4. Analysis of variance for
supply-definition scores measured at
time (T) of acquisition, retention 1 and
retention 2 as a function of exposure
frequency (F) and word set (S)
Source df F
Between-subjects
Set 1 1.62
Frequency 3 91.45*
S × F 3 3.70*
Error 182 (7.39)
Within-subjects
Time 2 15.96*
F × T 6 7.64*
S × T 2 0.85
F × S × T 6 4.42*
Error 364 (0.72)
Note. Values enclosed in parentheses represent mean square
errors. *p < .05.
for repeated analyses of the subjects’ data, the error terms of the omnibus F
tests were used for all follow-up analyses (Keppel, 1991).
The results of the repeated-measures ANOVA for supply-definition mea-
sures (see Table 4) showed that the amount of vocabulary gained varied as a
function of exposure frequency, F(3, 182) = 91.45 and as a function of time of
vocabulary measurement, F(2, 364) = 15.96. The main effects were qualified by
significant interactions of exposure frequency and word set, F(3, 182) = 3.70,

600 Susanne Rott
Table 5. Analysis of variance for
select-definition scores measured at
time (T) of acquisition, retention 1 and
retention 2 as a function of exposure
frequency (F) and word set (S)
Source df F
Between-subjects
Set 1 0.39
Frequency 3 151.67*
S × F 3 25.89*
Error 182 (9.26)
Within-subjects
Time 2 1.90
F × T 6 12.73*
S × T 2 1.44
F × S × T 6 0.68
Error 364 (0.96)
Note. Values enclosed in parentheses represent mean square er-
rors. *p < .05.
exposure frequency and time, F(6, 364) = 7.46, and a three-way interaction of
exposure frequency, word set, and time, F(6, 364) = 4.42. Also, the results for
the repeated-measures ANOVA for select-definition measures (see Table 5)
demonstrated that the amount of vocabulary acquired varied as a function of
exposure frequency, F(3, 182) = 151.67. Again, the result was qualified by inter-
actions. Word set and exposure frequency interacted significantly, F(3, 182) =
25.89, and the interaction of exposure frequency and time reached a level of
significance, F(6, 364) = 12.73.
To summarize: The omnibus F tests for supply- as well as select-definition
measures indicated that the frequency with which learners encountered unfa-
miliar vocabulary had an impact on the amount of vocabulary gained. The
quantity of vocabulary gained, however, varied at the times vocabulary knowl-
edge was assessed (immediately after reading, 1 week later, and 1 month
later) as well as across the two word sets.
To answer the first research question, which addressed the effect of read-
ing for language learners’ vocabulary development, main effects and interac-
tions of the omnibus F tests were followed up with a two-way between-
subjects ANOVA. Data were isolated to compare scores of the zero-exposure
condition with two-, four-, or six-exposure condition scores at either time of
acquisition, retention 1, or retention 2. The independent variables were word
set (I or II) and exposure frequency (0/2, 0/4, or 0/6). The dependent variables
were supply- and select-definition scores. Results are summarized in Table C1
(Appendix C). The analyses clearly demonstrated that after two exposures
subjects in both word set treatment groups gained more productive word
knowledge immediately after the reading exposure, F(1, 56) = 8.57; 1 week

Exposure Frequency, Acquisition, and Retention 601
later, F(1, 56) = 10.37; and 4 weeks later, F(1, 56) = 4.72, than subjects who had
not encountered the TWs during reading. Similar findings were made compar-
ing vocabulary gain of four encounters with zero encounters during reading.
Learners in the four-exposure reading condition gained significantly more
words immediately after reading, F(1, 62) = 25.55; 1 week later, F(1, 62) =
20.40; and 4 weeks later, F(1, 62) = 18.56. Likewise, after six exposures, learn-
ers gained more words than after zero exposures during reading as measured
immediately after text exposure, F(1, 62) = 62.55; 1 week later, F(1, 62) = 61.10;
and 4 weeks later, F(1, 62) = 38.53.
Also in the select-definition vocabulary tasks, learners in the reading expo-
sure condition outperformed learners in the nonreading condition. Results are
summarized in Table C2 (Appendix C). After two exposures during reading,
learners displayed more word knowledge than after zero exposures. The two-
way between-subjects ANOVA comparing learners’ receptive word knowledge
of two and zero exposures during reading resulted in a significant effect imme-
diately after reading, F(1, 56) = 43.77; 1 week later, F(1, 56) = 29.95; and 4
weeks later, F(1, 56) = 18.75. Learners with four exposures to the TWs outper-
formed learners in the nonreading condition immediately after reading, F(1,
62) = 50.87; 1 week later, F(1, 62) = 35.87; as well as 4 weeks later, F(1, 62) =
24.40. Six exposures to the TWs during reading also resulted in significantly
more receptive word gain than zero exposures as measured immediately after
reading, F(1, 62) = 90.82; 1 week later, F(1, 62) = 89.26; and 4 weeks later, F(1,
62) = 57.13.
With the second research question, the present study examined whether
the vocabulary gained incidentally during reading was also retained over time,
namely over a period of 4 weeks. The research question was first addressed
with two repeated-measures ANOVAs for select-definition and supply-defini-
tion vocabulary scores. The two between-subjects (independent) variables
were word set (I and II) and exposure frequency (two, four, and six expo-
sures). The three levels within-subjects dependent variable was time of vocab-
ulary measurement (acquisition, retention 1, and retention 2). The ANOVA for
supply-definition scores (see Table C3, Appendix C) resulted in a main effect
for exposure frequency, F(2, 89) = 11.52, and a main effect for time of vocabu-
lary measurement, F(2, 178) = 10.25. These main effects were further qualified
by interactions of exposure frequency and time, F(4, 178) = 2.63, and a three-
way interaction of word set, exposure frequency, and time of vocabulary mea-
surement, F(4, 178) = 3.27, which indicates that subjects of the two word-set
groups demonstrated different amounts of vocabulary knowledge after two,
four, and six exposures at the three different times that word gain was as-
sessed.
Follow-up analyses (see Table C4, Appendix C) of simple main effects for
the two-exposure frequency condition showed that subjects in the set I condi-
tion retained productive word knowledge over 1 and 4 weeks, F(2, 22) = 2.17,
nonsignificant, whereas the subjects’ vocabulary knowledge in the set II condi-
tion changed significantly within 4 weeks, F(2, 34) = 3.64. Subjects retained sig-
nificantly fewer words 4 weeks after reading (M = 1.44) as compared to words

602 Susanne Rott
acquired immediately after reading (M = 2.28) and words measured after 1
week (M = 2.33). Although all subjects in the four-exposure treatment condi-
tions retained productive word knowledge over 4 weeks, F(2, 32) = 1.96, non-
significant, in the set I as well as the set II, F(2, 30) = 0.34, nonsignificant,
condition, subjects who had encountered the TWs six times showed a change
in word knowledge within 4 weeks for the set I, F(2, 30) = 5.85, and the set II,
F(2, 30) = 9.45, treatment conditions. A statistically significant decrease in
word knowledge was recorded for learners of set I after 1 week (M = 5.38 at
acquisition, M = 4.75 after 1 week) but not between 1 week and 4 weeks (M =
4.44). Also, in the set II condition, a trend of word loss was identified. Al-
though learners demonstrated word gain after 1 week (M = 5.38 at acquisition,
M = 6.00 after 1 week), an overall significant word loss was noted after 4 weeks
(M = 4.00).
The ANOVA for the select-definition measure (see Table C5, Appendix C)
yielded a main effect for exposure frequency, F(2, 89) = 10.70, and a main ef-
fect for time of vocabulary measurement, F(2, 178) = 7.93. Follow-up analyses
(see Table C6, Appendix C) of simple main effects showed that subjects’ word
knowledge in the two-exposure condition of set II changed significantly, F(2,
32) = 4.19. Learners’ word knowledge significantly decreased within 4 weeks
(M = 2.28 at acquisition, M = 2.33 at retention 1, M = 1.44 at retention 2). Sub-
jects in all other treatment conditions retained receptive word knowledge
over 4 weeks, F(2, 22) = 0.42, nonsignificant, for the two-exposure set I condi-
tion; F(2, 32) = 0.35, nonsignificant, for the four-exposure set I condition F(2,
30) = 1.00, nonsignificant, for the four-exposure set II condition). Results for
the six-exposure set I condition were close to reaching a significant level of
word loss over 4 weeks. The ANOVA reached a level of significance, F(2, 30) =
5.00, which indicates the difference in word knowledge immediately after read-
ing (M = 8.13), 1 week later (M = 8.13), and 2 weeks later (M = 7.38). The analy-
sis of simple effects of word knowledge 1 week after reading and 4 weeks later
resulted in a level of significance, F(1, 15) = 9.00, but did not reach a level of
significance immediately after reading and 4 weeks later (although it was
close), F(1, 15) = 4.35, nonsignificant. In the six-exposure set II condition, sub-
jects also performed differently at the three times of word measurement, F(2,
30) = 4.13. The significant result can be explained with the curvilinear develop-
ment of word knowledge. Although word knowledge after 4 weeks (M = 7.88)
was significantly lower than knowledge assessed after 1 week (M = 9.13), word
knowledge was not significantly lower than immediately after reading (M = 8.38).
The third research question was guided by the assumption that the fre-
quency of exposure to an unknown word during reading has an effect on inci-
dental acquisition and retention. The ANOVAs for supply- and select-definition
measures reported for the second research question had resulted in main ef-
fects for exposure frequency and time and in additional interactions for the
supply-definition scores. To follow up on these results and answer the third
research question, data were further isolated and a between-subjects ANOVA
was used to compare subjects’ performance on production and recognition
tasks after two, four, and six exposures to the TWs. The two between-subjects

Exposure Frequency, Acquisition, and Retention 603
factors (independent variables) were two levels of word set (I and II) and two
levels of exposure frequency (2/4, 2/6, or 4/6). The dependent variable was
vocabulary knowledge at the three times of vocabulary assessment (acquisi-
tion, retention 1, or retention 2). Results for the production task are summa-
rized in Table C7 (Appendix C). There was no significant difference in
vocabulary gain on supply-definition tasks between subjects who had been ex-
posed to the TWs two or four times. The between-subjects ANOVA did not
reach a level of significance immediately after reading, F(1, 59) = 1.11, nonsig-
nificant; 1 week later, F(1, 59) = 0.08, nonsignificant; or 4 weeks later, F(1, 59)
= 1.28, nonsignificant. However, a significant interaction between word set and
exposure frequency indicated that the nonsignificant effect was not the same
for the two word sets, F(1, 59) = 5.80, 4 weeks after the reading treatment. Sub-
jects in the four-exposure set II condition (M = 4.18) displayed higher word
gains than did learners in the set I condition (M = 1.82), whereas subjects in
the two-exposure set I condition (M = 2.33) displayed higher word gains than
did learners in the set II condition (M = 1.44). Learners who had been exposed
to the TWs six times significantly outperformed learners who had been ex-
posed to the TWs two times on the production task immediately after reading,
F(1, 58) = 9.29; 1 week later, F (1, 58) = 6.99; and 4 weeks later, F(1, 58) = 5.50.
Likewise, the six-exposure condition yielded higher word knowledge than the
four-exposure condition immediately after reading, F(1, 61) = 4.31, and 1 week
later, F(1, 61) = 6.00. Four weeks later, the learners’ word knowledge was not
higher after six encounters, F(1, 61) = 1.60, nonsignificant. This result can be
explained with higher word gain scores in the four-exposure set II condition
(M = 4.18) compared to the four-exposure set I condition (M = 1.82) and the
overall low word gain scores of the six-exposure set I (M = 4.44) and set II con-
ditions (M = 4.00).
The analyses of select-definition tasks (see Table C8, Appendix C) showed
that two or four encounters with the TWs did not result in significantly differ-
ent word gain immediately after reading, F(1, 59) = 0.00, nonsignificant; 1 week
later, F(1, 59) = 0.00, nonsignificant; or 4 weeks later, F (1, 59) = 0.06, nonsignif-
icant. Six exposures to the TWs, however, resulted in more vocabulary knowl-
edge than two exposures immediately after reading, F(1, 58) = 4.31; 1 week
later, F(1, 58) = 6.97; as well as 4 weeks after reading, F(1, 58) = 4.25. Likewise,
six exposures to the TWs yielded significantly higher receptive word gain than
four exposures. This was the case immediately after text exposure, F(1, 61) =
4.54; 1 week later, F(1, 61) = 7.82; and 4 weeks after the reading task, F(1, 58)
= 4.73.
Summary of Main Findings
The findings of the present study have demonstrated:
1. Reading text for meaning is a significant source of intermediate L2 learners’ vocab-
ulary growth. Students who encountered an unfamiliar word two, four, or six times
during reading demonstrated significantly more word knowledge than students

604 Susanne Rott
Figure 2. Overall receptive word knowledge gain.
who had not encountered the words during reading. The effect of reading for word
learning had measurable effects immediately after reading exposure, 1 week later,
and 1 month later.
2. Furthermore, the data presented mixed results for the retention of word knowl-
edge gained incidentally during reading. All learners (except one treatment group)
retained a significant amount of receptive vocabulary knowledge over a period of
4 weeks. However, about half of the learners (both six-exposure groups and one
two-exposure group) showed a significant decrease in productive word knowledge
over 4 weeks. Yet, as can be seen in Figures 2 and 3 (average word gain pooled
over the two word set conditions), learners who had acquired words through
reading demonstrated more receptive as well as productive word knowledge after
4 weeks than learners who had not encountered the TWs through reading.
3. Again, mixed support was found for the impact of exposure frequency on inciden-
tal vocabulary growth. The study indicated that four encounters with unfamiliar
words during reading did not result in significantly more productive and receptive
word knowledge than two encounters. Six encounters, on the other hand, resulted
in significantly more receptive as well as productive word knowledge (with the ex-
ception of one experimental condition). Figures 4 and 5 show average word gain
pooled over the two word set conditions
DISCUSSION
The overall result of the present research study—namely, that L2 learners in-
cidentally acquired vocabulary during reading—replicated and confirmed find-

Exposure Frequency, Acquisition, and Retention 605
Figure 3. Overall productive word knowledge gain.
Figure 4. Overall select-definition scores.
ings of previous investigations (Day et al., 1991; Dupuy & Krashen, 1993;
Hulstijn et al, 1996; Knight, 1994; Pitts et al., 1989). Present findings expanded
the existing research through a controlled setup that ensured that learners
did not have any previous knowledge of the selected TWs and established
reading as a source for word gain by comparing word knowledge of learners
who had encountered the TWs in text with word knowledge of learners who
had not encountered the TWs in text. Further substantiating the significance
of normal reading as an important source of input for L2 learners’ vocabulary
development, the present investigation indicated that only two exposures to
an unknown word resulted in significant receptive as well as productive word
gain; that is, incidental acquisition took place even though it can be assumed
that learners encountered a variety of problems during the meaning making

606 Susanne Rott
Figure 5. Overall supply-definition scores.
process (as mentioned in the review of literature) that might not have led to
a successful inference and therefore not to acquisition or retention of the
TWs. The amount of word gain was modest, especially if compared to studies
with enhanced reading conditions (Hulstijn et al., 1996; Knight, 1994; Pari-
bakht & Wesche, 1996, 1997; Watanabe, 1997; Zimmerman, 1997). Yet, the
present data approximated findings of L1 vocabulary research, which regards
reading as a major source for lexical development during early school years.
L1 researchers who have established lexical acquisition during reading as
a cumulative incremental process (e.g., Nagy, Anderson, & Herman, 1987)
showed that the probability of gaining receptive word knowledge of an unfa-
miliar word during reading ranged between 0.08 and 0.22 when measured im-
mediately after text exposure (e.g., Nagy, Herman, & Anderson, 1985; Shu,
Anderson, & Zhang, 1995). Nagy et al. (1985) further assessed an “absolute
amount” (p. 261) of word learning based on the assumption that if a word is
retained over the period of six days it is committed to long-term memory. The
researchers found that the probability of word knowledge retention six days
after reading exposure was lower than when assessed immediately after read-
ing, namely, 0.05. The direct comparison of L1 research and present findings
is not possible because of the differences in study setup and because the L1
studies did not strictly account for exposure frequency or report on inferabil-
ity of the selected TWs. But considering the present data in the light of L1
findings, reading seems to have a similarly important impact on L2 as on L1
lexical development. The present data analysis, based on two to six expo-
sures, reflected the probability range of these L1 studies on receptive word
knowledge (0.06–0.13 across the three frequency ranges over a period of 4
weeks) and furthermore showed the probability of gaining productive knowl-
edge, which was lower and ranged between 0.03 and 0.09 (see Table 6).
Other L2 studies that assessed L2 learners’ receptive and productive word
knowledge gain through reading resulted in mixed findings. Paribakht and
Wesche (1997) noted that their learners gained almost exclusively receptive

Exposure Frequency, Acquisition, and Retention 607
Table 6. Probability* of words learned incidentally
Condition Acquisition Retention 1 Retention 2
Productive knowledge
2 exposures 0.03 0.04 0.03
4 exposures 0.05 0.05 0.05
6 exposures 0.09 0.09 0.07
Receptive knowledge
2 exposures 0.09 0.08 0.06
4 exposures 0.09 0.08 0.07
6 exposures 0.13 0.13 0.11
Note. 6 possible TWs. *Probability that an unfamiliar word will be learned was calculated as
suggested by Shu et al. (1995): “[T]he difference between the proportion of words known by
the group that has read the text and the group that has not, divided by the proportion of
words not known by students who have not read the text.” (pp. 85 ff)
Table 7. Percentage* of words learned incidentally
Condition Acquisition Retention 1 Retention 2
Receptive knowledge
2 exposures 43% 34% 28%
4 exposures 44% 36% 32%
6 exposures 61% 60% 48%
Productive knowledge
2 exposures 19% 22% 14%
4 exposures 28% 25% 24%
6 exposures 45% 45% 34%
Note. 6 possible TWs. *Percentage of words learned was calculated as suggested by Knight
(1994): The difference between the proportion of the words known by the group that has
read the text and the group that has not, divided by the total number of target words.
word knowledge through reading. Knight’s (1994) learners, on the other hand,
gained receptive knowledge as well as a significant but lower amount of pro-
ductive word knowledge. Likewise, in the present investigation, learners
gained a higher percentage of receptive word knowledge than productive
word knowledge. The percentage range of word gain of the present data col-
lection (see Table 7), however, was overall at least 10% higher than that re-
corded by Knight.
3
Two possible factors might have contributed to this
difference—namely, exposure frequency and inferenceability of the TWs.
Whereas most of the TWs in the study by Knight occurred only once, learners
in the present study encountered the TWs at least twice. Additionally, the
present study provided sufficient contextual clues for learners to be able to
infer meaning. The selected TWs were furthermore related to the main idea of
the passage, which possibly made readers feel the need to comprehend the
words. Even though Knight did not provide any information about the infer-
enceability of the TWs in her study, it can be assumed that the authentic texts
used as input passages did not provide sufficient contextual clues to allow
readers to assign meaning to each and every one of the TWs.

608 Susanne Rott
The interpretation of present findings strongly suggests that vocabulary
growth through reading has a measurably stronger effect on receptive than on
productive vocabulary knowledge. Although the relationship between re-
ceptive and productive knowledge is still unclear, it is generally held that re-
ceptive knowledge is an initial step toward productive knowledge on a word
knowledge continuum (e.g., Melka, 1997). In this framework, the finding that
almost all learners retained a significant amount of receptive word knowledge
and half of the subjects also retained productive word knowledge over 4
weeks implied that word learning during reading is a cumulative process.
This investigation further expanded research in the area of word retention.
The data analysis revealed that the encounter with unfamiliar words in text
had long-term effects on vocabulary growth. Even 4 weeks after vocabulary
exposure, learners demonstrated productive (14–34% over two to six expo-
sures) and receptive (28–48% over two to six exposures) word knowledge (see
Table 7). None of the previously conducted studies assessed word retention 4
weeks after the reading treatment by comparing the word knowledge of learn-
ers who encountered the TWs in text with the word knowledge of learners
who did not encounter the TWs in text. This evidence was essential because
it allowed for further interpretation of the findings. Although an overall trend
of word loss over a period of 4 weeks was evident, learners of all but one
experimental condition retained receptive word knowledge, but only about
50% of the learners retained a significant amount of productive vocabulary
knowledge. Yet, as discussed with respect to the first research question,
learners still demonstrated more productive word knowledge after encounter-
ing the TWs during reading than did learners who had not encountered the
words in text. Moreover, it can be assumed that a further decrease in word
knowledge should be relatively small because it is generally assumed that a
“decline in retention is sharpest immediately after learning and becomes in-
creasingly gradual” (Nagy et al., 1987, p. 261).
The current result did not confirm observations made by Knight (1994). In
her study, subjects showed higher productive word knowledge 2 weeks after
the reading treatment than immediately after reading exposure. She explained
this delayed effect of word gain with an intensive learner-text interaction tak-
ing place during the word inferencing process. Although for 2 (out of 12) treat-
ment groups of the present study an increase in word knowledge was
registered after 1 week, learners did not demonstrate more word knowledge
after 4 weeks than they did immediately after reading. Instead, learners who
encountered the TWs six times during reading showed a significant decrease
of productive word knowledge after 4 weeks. For those learners, learner-text
interaction had potentially been more frequent than for learners who had en-
countered the TWs four times during reading and who retained word knowl-
edge over 4 weeks. There seem to be two possible explanations for this
inconsistency in the findings. It could have been due to the variations in the
studies’ setups and reading materials used, and possibly also due to learners’

Exposure Frequency, Acquisition, and Retention 609
overall text comprehension. Both studies required learners to immediately re-
call the content of the text. But, whereas readers in Knight’s study recalled
the entire text at once, the text of the present study was divided into six para-
graphs. Because of this, learners had to cope with a smaller amount of proposi-
tional content at a time, which might have resulted in less intensive learner-text
interaction. Also, the content of the texts used for this investigation might have
contributed to the intensity of learner-text interaction. Present text passages
provided situations of daily life and had been created to be easily understood
by intermediate foreign language learners; that is, easy content comprehensibil-
ity of the context of TWs might have required less intensive learner-text interac-
tion than the reading of a literary text as in the study by Knight.
Furthermore, the present study sought to assess the effect of exposure fre-
quency on learners’ vocabulary growth. Previous L2 investigations either did
not account for the variable of exposure frequency (Day et al., 1991; Dupuy &
Krashen, 1993; Knight, 1994; Pitts et al., 1989), suggested a minimum of 5–15
exposures based on less controlled study setups and instrumentations (e.g.,
Crothers & Suppes, 1967; Kachroo, 1962; Saragi, Nation, & Meister, 1978), or
compared low frequencies such as three versus one encounter within a single
text (Hulstijn et al., 1996). Hulstijn et al. found that there was no significantly
higher word gain after three exposures than after one exposure. Similarly,
present findings were based on multiple exposures that were chronologically
distinct (one exposure per week), which did not result in a difference in word
gain after either two or four exposures. The percentage of receptive word
knowledge gain was very close immediately after reading (43% vs. 44%) as
well as 4 weeks later (28% vs. 32%; Table 7). The percentage of productive
word gain was not quite as close after two or four exposures ranging from
19% versus 28% immediately after reading to 14% versus 24% 4 weeks later
(Table 7). Six exposures to an unfamiliar word during reading, on the other
hand, resulted in significantly higher receptive word knowledge immediately
after text exposure (61% vs. 43% and 44%) as well as 4 weeks later (48% vs.
28% and 32%; Table 7). Productive word gain was also significantly higher
after six exposures immediately after reading (45% vs. 19% and 28%; Table 7),
whereas the difference between four and six exposures was smaller (34% vs.
24%) 4 weeks later. These findings confirmed the logical assumption that mul-
tiple exposures to an unfamiliar lexical item positively affect learning. More-
over, the findings demonstrated that even a lower frequency such as six
exposures, which can take place over a period of weeks, makes a significant
difference in L2 readers’ lexical development.
To summarize: Present findings suggest that reading can have a similarly
favorable long-term effect on adult L2 vocabulary growth as on child L1 lexical
development, given texts that are sufficiently rich as to allow readers to as-
sign meaning to unfamiliar words. Although two exposures during reading al-
ready result in word knowledge gain, the likelihood of speeding up lexical
growth is considerably increased through six encounters.

610 Susanne Rott
LIMITATIONS AND SUGGESTIONS FOR FUTURE RESEARCH
The setup of the present study did not engage L2 learners in a fully ecologi-
cally valid reading situation
4
and did not control for all confounding variables.
Therefore, findings should be followed up with research that reconfirms the
present results with varied materials and expands the research body by as-
sessing the effect of reading on learners’ L2 vocabulary development by con-
trolling and systematically manipulating additional learner and text factors.
First, concerning the treatment and the materials, neither the texts used
nor the reading tasks were part of the treatment groups’ class syllabi. This,
therefore, meant an interruption of normal classroom procedures. Each treat-
ment took about half a class session per week. A potential impact of height-
ened attention to the treatment or fatigue of participation was especially
strong for six-exposure groups who read texts for the study once a week over
6 consecutive weeks. Moreover, the interpretation of the present findings is
limited in that the texts used for the treatment were specifically created for
the investigation and furthermore provided readers with contextual clues to
assign meaning to the TWs. But, as summarized in the review of literature,
texts are not always sufficiently rich to allow learners to assign the correct
meaning to unfamiliar words. To further underscore the effectiveness of read-
ing for L2 readers’ vocabulary growth, future research needs to use authentic
readings or textbook readings that are fully integrated into the L2 curriculum.
Additionally, the present results were based on only six nouns and six
verbs concerning objects and activities of daily life. Second-year learners in
most language programs, however, are usually exposed to a variety of topics
containing conceptually more difficult words than those used for the present
study. Consequently, research needs to further investigate whether the re-
sults of this study also hold for words that are conceptually more abstract.
Second, concerning the assessment of word knowledge, future research
needs to apply a more refined and varied instrumentation to assess the effect
of reading on the depth of L2 readers’ incidental word knowledge gain (as for
example in Joe, 1995; Paribakht & Wesche, 1996). Summarizing previous stud-
ies, Melka (1997) explained that there are different degrees of recognition and
production skills ranging from recognizing visual features to being able to use
a word appropriately in all possible contexts and situations. On a word-knowl-
edge continuum scale, the assessment tasks used for the present study mea-
sured a low degree of both receptive and productive word knowledge.
Although the multiple-choice recognition task emphasized only semantic dif-
ference, the production task required learners to produce the TW out of con-
text.
Recent theoretical and research literature has demonstrated the colloca-
tional network of learning and storing words. Moon (1997) pointed out that
lexical items should not be looked at in isolation because as soon as words
are part of a text (written or oral) they form “meaningful and inseparable
units” (p. 43) with other words (multiword items).
5
In fact, Schouten-van Par-

Exposure Frequency, Acquisition, and Retention 611
reren (1989) found that learners retrieved word meaning by recollecting the
words from the immediate context of that particular TW. Additional research,
therefore, needs to further develop vocabulary tests that assess learners’ L2
word gain through reading by their appropriate use and interpretation of the
TWs as part of multiword items in various contexts.
Third, concerning the effects of learner and text variables on L2 vocabulary
development through reading, research on incidental word acquisition has so
far separately investigated the process of word inferencing and the product of
word inferencing, namely, acquisition. Future investigations should focus on
combining both lines of research and systematically manipulate learner and
text variables in order to get a clearer picture of the impact that variables
affecting the process of inferencing have on incidental word gain. The learner
variable—reading comprehension—is of particular interest because it com-
prises a whole range of text-based and learner-based factors that have an im-
pact on the product of text comprehension. A separate line of research needs
to focus on the relationship between textual features and acquisition, concen-
trating on issues such as the importance of TWs to overall text comprehen-
sion, whether TWs are implicated in any causal relations that make them
more memorable than other words, and the impact of connections to other
text forms through antecedent reference, synonymy, or inclusion relations.
Additionally, this line of investigation needs to further account for the effect
of word properties. Laufer (1997), for example, summarized intralexical fac-
tors that can have a facilitating or a difficulty-inducing affect on word learning
in general. Their impact on incidental word gain through reading has not yet
been systematically investigated.
Moreover, the present findings were based on data from 12 treatment
groups of intact fourth-semester language classes. Thus, for logistical reasons,
this study did not randomly assign learners to treatment groups nor did the
study ensure that the treatment groups were equally equivalent intermediate
L2 learners. Assessing and accounting for learner factors such as reading abil-
ity and vocabulary size and the relationship to incidental word gain could fur-
ther show whether there is a threshold below which reading as a source for
lexical development is less effective for second-year learners.
(Received 15 September 1998)
NOTES
1. Recent research and theoretical literature have attempted to demonstrate and describe the
mental processes involved in incidental word gain during reading. Although some researchers limit
incidental word gain to the strategic interaction between reader and unedited text (Ellis, 1994), oth-
ers expand the concept of incidental to enhanced reading conditions such as glossed texts (Hulstijn,
1992; Watanabe, 1997). For the present investigation, the former concept of incidental learning was
used.
2. Based on four 16-week-long semesters with 4 contact hours per week and a target vocabulary
of 5,000 base words, learners have to gain about 20 words per contact hour.
3. For the supply-definition task, Knight (1994) received a percentage range of 5–11% and, for the
select-definition task, the percentage of word gain ranged between 20% and 35%.

612 Susanne Rott
4. As one of the reviewers suggested, the present study was not set up to draw any immediate
pedagogical conclusions. Implications are, therefore, merely suggestive. Approximating the probabil-
ity of L1 word learning, reading for meaning appears to be a suitable basis for accelerating learners’
L2 vocabulary growth. Second-year learners who seek to obtain advanced language skills and plan
to pursue academic studies in the L2 should be advised to engage in the massive reading of texts to
speed up the development of their L2 lexicon. Keeping in mind learners’ limited L2 reading abilities
and the potential problems with assigning meaning to unfamiliar words, it cannot be assumed that
each text has an equally positive impact for all learners. In fact, the present data showed that learn-
ing varied among the individual experimental groups. Extensive reading of a variety of texts, as for
example suggested by Krashen (1993), is therefore supported by the present study. This pedagogical
suggestion and the national trend to place a focus on reading in second-year language classes, how-
ever, needs to be further qualified. The present investigation demonstrated that repeated encounters
with unfamiliar words, more specifically six versus two or four encounters, resulted in higher effects
of reading as a source for vocabulary growth. This observation should be considered in curriculum
design and the selection of reading materials for intermediate language classes. Many of the text-
books presently used aim to cover as many different topics as possible in order to expose learners
to a wide variety of vocabulary and structures. Often, one text covers one aspect or one particular
content area. Taking present findings into account, this would mean that learners should engage in
reading several texts on the same topic, what Krashen (1993) calls “narrow reading,” to increase the
likelihood of encountering words pertaining to one semantic field multiple times and in that way
promote the rate of vocabulary growth. Day and Bamford (1998), for example, created a list of read-
ing materials suitable for extensive reading as well as intensive reading on one particular topic for
learners of various backgrounds and interests. Easily accessible materials are specialized journals
(fashion, automobile, gardening, etc.), series written for adolescents (also suggested by Krashen,
1993), children’s books, and newspapers. An investigation conducted by Paribakht and Wesche
(1997) further supports the idea of narrow reading. The researchers reported a “surprisingly” high
incidental word gain during reading, attributing their finding to repeated encounters and themati-
cally related readings.
Furthermore, the present study found that words acquired incidentally during reading could be
better assessed through learners’ receptive vocabulary knowledge than their productive knowledge.
This finding needs to be kept in mind when assessing the effect of reading on vocabulary develop-
ment and when assessing lexical proficiency in general. Language instructors as well as language
learners tend to assess lexical proficiency through oral and written output tasks. As the present
study has shown, however, learners’ vocabulary gain during reading as measured through a re-
ceptive task yielded higher results than when measured with a production task. This result does not
warrant discarding reading as an effective source for vocabulary growth by assessing its potential
only through immediate output tasks. In the assessment process, teachers as well as learners need
to keep in mind that intermediate-level language learners can comprehend more language than they
can produce.
5. Furthermore, cognitive psychologists Landauer and Dumais (1997) simulated vocabulary learn-
ing with latent semantic analysis and found that word acquisition did not result only through direct
word exposure but also through exposure of collocational networks, which opens up a whole new
line of research.
REFERENCES
Anderson, R. C., & Nagy, W. E. (1992). The vocabulary conundrum. American Educator: The Profes-
sional Journal of the American Federation of Teachers, 16(4), 14–18, 44–46.
Bensoussan, M., & Laufer, B. (1984). Lexical guessing in context in EFL reading comprehension. Jour-
nal of Research in Reading, 7, 15–32.
Berhnardt, E. B. (1991). Reading development in a second language: Theoretical, empirical, and class-
room perspectives. Norwood, NJ: Ablex.
Brown, H. D. (1994). Teaching by principles. An interactive approach to language pedagogy. Englewood
Cliffs, NJ: Prentice Hall.
Chern, C. L. (1993). Chinese students’ word solving strategies in reading in English. In T. Huckin, M.
Haynes, & J. Coady (Eds.), Second language reading and vocabulary learning (pp. 67–85). Nor-
wood, NJ: Ablex.
Coady, J. (1993). Research in ESL/EFL vocabulary acquisition: Putting it in context. In T. Huckin, M.
Haynes, & J. Coady (Eds.), Second language reading and vocabulary learning (pp. 3–23). Nor-
wood, NJ: Ablex.

Exposure Frequency, Acquisition, and Retention 613
Coady, J. (1997). L2 vocabulary acquisition through extensive reading. In J. Coady & T. Huckin (Eds.),
Second language vocabulary acquisition (pp. 225–237). New York: Cambridge University Press.
Crothers, E., & Suppes, P. (1967). Experiments in second-language learning. New York: Academic
Press.
D’Anna, C. A., Zechmeister, E. B., & Hall, J. W. (1991). Toward a meaningful definition of vocabulary
size. Journal of Reading Behavior, 13, 109–122.
Day, R. R., & Bamford, J. (1998). Extensive reading in the second language classroom. New York: Cam-
bridge University Press.
Day, R., Omura, C., & Hiramatsu, M. (1991). Incidental EFL vocabulary learning and reading. Reading
in a Foreign Language, 7, 541–551.
Diakidoy, I. A. N. (1993). The role of reading comprehension and local context characteristics in word
meaning acquisition during reading. Unpublished doctoral dissertation, University of Illinois, Ur-
bana-Champaign.
Dupuy, B., & Krashen, S. (1993). Incidental vocabulary acquisition in French as a foreign language.
Applied Language Learning, 4, 55–63.
Ellis, N. (1994). Implicit and explicit learning of languages. London: Academic Press.
Goulden, R., Nation, P., & Read, J. (1990). How large can a receptive vocabulary be? Applied Linguis-
tics, 11, 341–363.
Haynes, M. (1993). Patterns and perils of guessing in second language reading. In T. Huckin, M.
Haynes, & J. Coady (Eds.), Second language reading and vocabulary learning (pp. 46–66). Nor-
wood, NJ: Ablex.
Hazenberg, S., & Hulstijn, J. H. (1996). Defining a minimal receptive second-language vocabulary for
non-native university students: An empirical investigation. Applied Linguistics, 17, 145–163.
Huckin, T., & Bloch, J. (1993). Strategies for inferring word meaning in context: A cognitive model. In
T. Huckin, M. Haynes, & J. Coady (Eds.), Second language reading and vocabulary learning (pp.
153–180). Norwood, NJ: Ablex.
Huckin, T., & Haynes, J. (1993). Summary and future directions. In T. Huckin, M. Haynes, & J. Coady
(Eds.), Second language reading and vocabulary learning (pp. 289–298). Norwood, NJ: Ablex.
Hulstijn, J. H. (1992). Retention of given and inferred word meanings: Experiments in incidental vo-
cabulary learning. In P. J. L. Arnaud & H. Bejoint (Eds.), Vocabulary and applied linguistics (pp.
113–125). London: Macmillan.
Hulstijn, J. H., Hollander, M., & Greidanus, T. (1996). Incidental vocabulary learning by advanced
foreign language students: The influence of marginal glosses, dictionary use, and reoccurrence
of unknown words. Modern Language Journal, 80, 327–339.
Joe, A. (1995) Text-based tasks and incidental vocabulary learning. Second Language Research, 11,
149–158.
Judd, E. L. (1978). Vocabulary teaching and TESOL: A need for reevaluation of existing assumptions.
TESOL Quarterly, 12, 71–75.
Kachroo, J. N. (1962). Report on an investigation into the teaching of vocabulary in the first year of
English. Bulletin of the Central Institute of English, 2, 67–72.
Keppel, G. (1991). Design and analysis. A researcher’s handbook. Englewood Cliffs, NJ: Prentice Hall.
Knight, S. (1994). Dictionary, the tool of last resort effects on comprehension and vocabulary acquisi-
tion for students of different verbal abilities. Modern Language Journal, 78, 285–298.
Krashen, S. (1989). We acquire vocabulary and spelling by reading: Additional evidence for the Input
Hypothesis. Modern Language Journal, 73, 450–464.
Krashen, S. (1993). The power of reading. Englewood, NJ: Libraries Unlimited.
Landauer, T. K., & Dumais, S. T. (1997). A solution to Plato’s problem: The latent semantic analysis
theory of acquisition, introduction, and representation of knowledge. Psychological Review, 101,
211–240.
Laufer, B. (1997). What’s in a word that makes it hard or easy: Some intralexical factors that affect
the learning of words. In N. Schmitt & M. McCarthy (Eds.), Vocabulary: Description, acquisition,
and pedagogy (pp. 140–155). New York: Cambridge University Press.
Lee, J. F., & Wolf, D. (1997). A quantitative and qualitative analysis of the word-meaning inferencing
strategies of L1 and L2 readers. Spanish Applied Linguistics, 1, 24–46.
Lee, J. F., & VanPatten, B. (1995). Making communicative language teaching happen. New York: Mc-
Graw Hill.
Luppescu, S., & Day, R. R. (1995). Reading, dictionaries, and vocabulary learning. In B. Harley (Ed.),
Lexical issues in language learning (pp. 229–251). Amsterdam: Benjamins.
Melka, F. (1997). Receptive vs. productive aspects of vocabulary. In N. Schmitt & M. McCarthy (Eds.),
Vocabulary: Description, acquisition, and pedagogy (pp. 84–102). New York: Cambridge University
Press.

614 Susanne Rott
Mondria, J., & Wit-DeBoer, M. (1991). The effects of contextual richness on the guessability and the
retention of words in a foreign language. Applied Linguistics, 12, 249–267.
Moon, R. (1997). Vocabulary connections: Multi-word items in English. In N. Schmitt & M. McCarthy
(Eds.), Vocabulary: Description, acquisition, and pedagogy (pp. 40–63). New York: Cambridge Uni-
versity Press.
Na, L., & Nation, I. S. P. (1985). Factors affecting guessing vocabulary in context. RELC Journal, 16,
33–42.
Nagy, W. E., Anderson, R. C., & Herman, P. A. (1987). Learning word meanings from context during
normal reading. American Educational Research Journal, 24, 237–270.
Nagy, W. E., Herman, P. A., & Anderson, R. C. (1985). Learning words from context. Reading Research
Quarterly, 20, 233–253.
Nation, I. S. P. (1990). Teaching and learning vocabulary. Rowley, MA: Newbury House.
Nation, P., & Warig, R. (1997). Vocabulary size, text coverage, and word lists. In N. Schmitt & M.
McCarthy (Eds.), Vocabulary: Description, acquisition, and pedagogy (pp. 6–19). New York: Cam-
bridge University Press.
Paribakht, T. S., & Wesche, M. (1996). Enhancing vocabulary acquisition through reading: A hierar-
chy of text-related exercise types. The Canadian Modern Language Review, 52, 154–178.
Paribakht, T. S., & Wesche, M. (1997). Vocabulary enhancement activities and reading for meaning in
second language vocabulary acquisition. In J. Coady & T. Huckin (Eds.), Second language vocabu-
lary acquisition (pp. 174–200). New York: Cambridge University Press.
Perry, K. (1993). Too many words: Learning the vocabulary of an academic subject. In T. Huckin, M.
Haynes, & J. Coady (Eds.), Second language reading and vocabulary learning (pp. 109–129). Nor-
wood, NJ: Ablex.
Pitts, M., White, H., & Krashen, S. (1989). Acquiring second language vocabulary through reading: A
replication of the Clockwork Orange study using second language acquirers. Reading in a For-
eign Language, 5, 271–275.
Richards, J. (1976). The role of vocabulary teaching. TESOL Quarterly, 10, 77–89.
Rott, S. (1997). The effect of exposure frequency and reading comprehension on incidental vocabulary
acquisition and retention through reading for learners of German as a foreign language. Unpub-
lished doctoral dissertation, University of Illinois, Urbana-Champaign.
Saragi, T., Nation, I. S. P., & Meister, G. F. (1978). Vocabulary learning and reading. System, 6(2),
72–78.
Schmidt, R. (1995). Consciousness and foreign language learning: A tutorial on the role of attention
and awareness in learning. In R. Schmidt (Ed.), Attention & awareness in foreign language learn-
ing (pp. 1–65). Honolulu: University of Hawai‘i Press.
So
¨
kmen, A. J. (1997). Current trends in teaching second language vocabulary. In N. Schmitt & M.
McCarthy (Eds.), Vocabulary: Description, acquisition, and pedagogy (pp. 237–257). New York:
Cambridge University Press.
Schouten-van Parreren, C. (1989). Vocabulary learning through reading: Which conditions should be
met when presenting words in text? AILA Review, 6, 75–85.
Shu, H., Anderson, R. C., & Zhang, H. (1995). Incidental learning of word meanings while reading: A
Chinese and American cross-cultural study. Reading Research Quarterly, 30, 76–95.
Walker, L. J. (1983). Word identification strategies in reading a foreign language. Foreign Language
Annals, 16, 293–299.
Watanabe, Y. (1997). Input, intake, and retention: Effects of increased processing on incidental learn-
ing of foreign language vocabulary. Studies in Second Language Acquisition, 19, 287–308.
Zechmeister, E. B., D’Anna, C. A., Hall, J., Paus, C. H., & Smith, J. A. (1993). Metacogntitive and other
knowledge about the lexicon: Do we know how many words we know? Applied Linguistics, 14,
188–206.
Zimmermann, C. B. (1997). Do reading and interactive vocabulary instruction make a difference? An
empirical study. TESOL Quarterly, 31, 121–140.

Exposure Frequency, Acquisition, and Retention 615
APPENDIX A
LIST OF TARGET WORDS
Set I Set II
Nouns Nouns
Knauser “scrooge” Glotze “television” (colloquial)
Bude “apartment” (colloquial) Inserat “advertisement”
Klamotten “clothes” (colloquial) Karre “car” (colloquial)
Verbs Verbs
miefen “to stink” beschaffen “to get hold of something”
quasseln “to chat” bu
¨
ffeln “to study”(colloquial)
besudeln “to besmear” meckern “to complain”
APPENDIX B
PASSAGES WITH TARGET WORD BUDE
17. Ma
¨
rz: Heute fru
¨
h mußte ich schon um 5 Uhr aufstehen, um mir die Zeitung zu
kaufen und nach Anzeigen durchsehen. Ich brauche endlich meine eigene Bude. Ich mag
einfach nicht mehr bei meinen Eltern wohnen. Wenn ich jetzt anfange zu studieren,
werde ich bestimmt viele neue Leute kennenlernen, die ich zu mir einladen will. Wenn
ich bei meinen Eltern wohne, mu
¨
ssen meine Freunde schon um 10 Uhr gehen und wir
ko
¨
nnen keine richtigen Partys feiern.
2. April: Ich bin am Ende meiner Nerven. Seit drei Wochen suche ich jetzt schon nach
einer Bude und habe noch nichts gefunden. Entweder sind sie scho
¨
n mit großen Zim-
mern, in der Na
¨
he der Uni und viel zu teuer, oder die Zimmer sind klein und ha
¨
ßlich
und weit von der Uni entfernt. Nu
¨
chste Woche fa
¨
ngt die Uni an. Ich hoffe, ich finde bald
etwas.
4. April: Ich wohne jetzt also in der Goethestraße. Mein Zimmer ist nicht so groß,
aber hell. Ich muß mir noch ein Bu
¨
cherregal und ein paar Pflanzen kaufen. Peter und Uli
sind meine Mitbewohner. Sie wohnen schon seit zwei Semestern in dieser Bude und ha-
ben alles in der Ku
¨
che, was man so braucht.
8. April: Die Uni hat jetzt angefangen, und ich habe auch schon ein paar nette Leute
kennengelernt. Sie werden am Freitag zu mir kommen. Wir haben na
¨
mlich eine Semes-
ter-Anfangs-Party in unserer Bude. In der Ku
¨
che gibt es das Essen, im Flur gibt es Ge-
tra
¨
nke und in Ulis Zimmer wird getanzt. Uli hat na
¨
mlich eine gute Stereoanlage.
9. April: Die Party war einfach super. Wir haben die ganze Nacht durchgefeiert und
morgens zusammen gefru
¨
hstu
¨
ckt. Uli und ich, wir haben die Bro
¨
tchen vom Ba
¨
cker geh-
olt und die anderen haben die Bude aufgera
¨
umt. Wir saßen noch bis mittags zusammen
und haben geredet. So eine Party sollten wir jedes Wochenende machen.
20. April: Heute haben mich meine Eltern zum ersten Mal in meiner Bude besucht.
Ich habe natu
¨
rlich extra vorher aufgera
¨
umt und die Ku
¨
che sauber gemacht. Wir haben
zusammen mit Peter und Uli Kaffee getrunken. Meine Mutter hat gemerkt, daß wir keine
Kaffeemaschine haben und mir sofort Geld gegeben, damit wir uns eine kaufen ko
¨
nnen.

616 Susanne Rott
APPENDIX C
Table C1. ANOVA summary table of supply-definition
scores for time (A, R1, R2), frequency (F0/2, F0/4, F0/6)
and word set
Source df SS MS F
Between-subjects
F0/2 at A 1 79.34 79.34 8.57*
F0/2 at R1 1 96.10 96.10 10.37*
F0/2 at R2 1 34.84 34.84 4.72*
F0/4 at A 1 188.79 188.79 25.55*
F0/4 at R1 1 150.74 150.74 20.40*
F0/4 at R2 1 137.13 137.13 18.56*
F0/6 at A 1 462.25 462.25 62.55*
F0/6 at R1 1 451.56 451.56 61.10*
F0/6 at R2 1 284.77 284.77 38.53*
Note. *p < .05.
Table C2. ANOVA summary table of select-definition
scores for time (A, R1, R2), frequency (F0/2, F0/4, F0/6)
and word set
Source df SS MS F
Between-subjects
F0/2 at A 1 405.34 405.34 43.77*
F0/2 at R1 1 277.38 277.38 29.95*
F0/2 at R2 1 173.61 173.61 18.75*
F0/4 at A 1 471.06 471.06 50.87*
F0/4 at R1 1 332.15 332.15 35.87*
F0/4 at R2 1 225.95 225.95 24.40*
F0/6 at A 1 841.00 841.00 90.82*
F0/6 at R1 1 826.56 826.56 89.26*
F0/6 at R2 1 529.00 529.00 57.13*
Note. *p < .05.

Exposure Frequency, Acquisition, and Retention 617
Table C3. Repeated measures analysis of
variance for supply-definition scores
measured at acquisition, retention 1, and
retention 2 (T) as a function of exposure
frequency (F) and word set (S)
Source df F
Between-subjects
Set 1 1.05
Frequency 2 11.52*
S × F 2 2.27
Error 89 (14.95)
Within-subjects
Time 2 10.25*
S × T 2 0.49
F × T 4 2.63*
S × F × T 4 3.27*
Error 178 (1.37)
Note. Values enclosed in parentheses represent mean square er-
rors. *p < .05.
Table C4. ANOVA summary table of supply-definition
measures for time at level of frequency (F2, F4, or F6)
and word set (1 and 2)
Source df SS MS F
Within-subjects
F2 at set 1 2 2.97 1.49 2.17
F2 at set 2 2 4.99 2.49 3.64*
F4 at set 1 2 2.69 1.34 1.96
F4 at set 2 2 0.47 0.23 0.34
F6 at set 1 2 8.01 4.00 5.85*
F6 at set 2 2 12.94 6.47 9.45*
Note. *p < .05.

618 Susanne Rott
Table C5. Repeated-measures analysis of
variance for select-definition scores
measured at acquisition, retention 1, and
retention 2 (T) as a function of exposure
frequency (F) and word set (S)
Source df F
Between-subjects
Set 1 0.00
Frequency 2 10.70*
S × F 2 1.81
Error 89 (16.35)
Within-subjects
Time 2 7.93*
S × T 2 1.26
F × T 4 2.01
S × F × T 4 0.61
Error 178 (1.03)
Note. Values enclosed in parentheses represent mean square er-
rors. *p < .05.
Table C6. ANOVA summary table of select-definition
measures for time at level of frequency (F2, F4, or F6)
and word set (1 and 2)
Source df SS MS F
Within-subjects
F2 at Set 1 2 0.43 0.21 0.42
F2 at Set 2 2 4.32 2.16 4.19*
F4 at Set 1 2 0.36 0.18 0.35
F4 at Set 2 2 1.03 0.51 1.00
F6 at Set 1 2 25.75 12.88 5.00*
F6 at Set 2 2 4.23 2.11 4.13*
Note. *p < .05.

Exposure Frequency, Acquisition, and Retention 619
Table C7. ANOVA summary table of supply-definition
scores for time (A, R1, R2), exposure frequencies (2/4,
2/6, 4/6) and word set
Source df SS MS F
Between-subjects
F2/4 at A 1 16.53 16.53 1.11
F2/4 at R1 1 1.15 1.15 0.08
F2/4 at R2 1 19.17 19.17 1.28
F2/6 at A 1 138.96 138.96 9.29*
F2/6 at R1 1 104.45 104.45 6.99*
F2/6 at R2 1 82.28 82.28 5.50*
F4/6 at A 1 64.36 64.36 4.31*
F4/6 at R1 1 89.76 89.76 6.00*
F4/6 at R2 1 23.90 23.90 1.60
Note. *p < .05.
Table C8. ANOVA summary table of select-definition
scores for time (A, R1, R2), exposure frequencies (2/4,
2/6, 4/6) and word set
Source df SS MS F
Between-subjects
F2/4 at A 1 0.00 0.00 0.00
F2/4 at R1 1 0.06 0.06 0.00
F2/4 at R2 1 1.05 1.05 0.06
F2/6 at A 1 72.99 72.99 4.31*
F2/6 at R1 1 118.13 118.13 6.97*
F2/6 at R2 1 72.07 72.07 4.25*
F4/6 at A 1 76.92 76.92 4.54*
F4/6 at R1 1 132.51 132.51 7.82*
F4/6 at R2 1 59.08 59.08 4.73*
Note. *p < .05.