Based on the results presented in section 4.1.2, learners with higher reading
proficiency did not significantly obtain better scores in vocabulary gain and retention
for the target words in both culturally familiar and unfamiliar texts. Table 4-12
demonstrated that the H group performed better than the L group in vocabulary gain
and even to a significant degree in vocabulary retention in test 3, which was rated as
less culturally familiar. This retention disparity was in keeping with the finding of
Hong’s study (2005) and could be explained by the assertion (Perfetti & Lesgold,
1977, 1979; Pulido, 2004) that more proficient readers’ larger sight vocabulary in the
texts, facility in detecting the interrelationships among words, phrases and sentences,
better automaticity in lower level processing and superior decoding skills provided
them with more contextual clues and extra attentional resources to infer the meanings
of target words with comparatively greater ease and precision. Nevertheless, in test 5
108
which was deemed more culturally familiar, the H group did worse than the L group.
This might be due to the fact that the H group could get the general idea of the
passage in test 5 more readily, which was proved by their higher mean scores in
reading comprehension (see Table 4-12). As a result, they might just neglect those
unknown target words on condition that the major reading purpose of the exercise was
already fulfilled. On the contrary, the L group had to depend more on the bottom-up
processing to get the gist of the passage and allotted more attention to target words in
test 5, which in turn led to better lexical development. One of interviewees in the
present study used to indicate that she would attend to difficult words more when she
found the text hard to understand.
Though proficient readers in this experiment gained more knowledge about
target lexical items than their less competent counterparts also in test 4 and in the
score for all target words (TW1-20), their scores were lower in vocabulary gain and
retention only to a slight degree in test 1, test 2, and test 5. To explain such
inconsistent phenomenon which contradicted my expectation, the researcher carries
out further detailed analysis to examine why proficient readers obtained lower mean
scores than their counterparts in vocabulary gain and retention in these three tests.
Some general findings are summarized as follows. First, the higher reading
proficiency group’s inferior scores in vocabulary gain and retention in these tests
109
could be attributed to the fact that the participants in this group might have already
known some target words quite well and thus have less room for improvement, while
the lower reading proficiency group performes worse in the pretest but its
understanding of these target lexical items later advances to the same level as the
higher reading proficiency group, which means they have improved more. Though
there are some differences between more proficient and less competent readers in their
vocabulary gain and retention of some target words, the disparity most does not reach
any statistical significance. This could be explained by the fact that these participants
belong to the same proficiency group (the intermediate level), which suggests that
they might possess similar competence in learning English. Sometimes the higher
reading proficiency group might be careless, which renders the lower reading
proficiency group the opportunity to surpass it. Most of the aforementioned
differences between these two groups does not reach significance, which manifests
that proficient readers do not gain significantly more knowledge about target words
than the less competent readers in most tests. This finding testifies that one’s reading
proficiency level does not bear a significant relationship with his or her incidental
word learning and also confirms the study of Nagy et al. (1987). Thirdly, from the
interview with nine participants representing high, intermediate and low levels, most
of them report that they tend to make more efforts to grasp the gist instead of paying
110
sufficient attention to target words because the task they receive mainly requires them
to get a general picture of the assigned passages. They would just underline the target
words and sometimes resort to the contextual clues for the inference of their meanings.
For instance, one participant at the high level reports as follows:
“I tend to invest more efforts to grasp the gist of the selected articles rather than to notice unknown words while processing those five texts. As for those unfamiliar lexical items, I would underline them to draw my attention to them and that was all I would do.”
One participant at the intermediate level used to indicate that
“I would allot more attention to achieve a general understanding of the main ideas of the assigned passages, while attending less to those unfamiliar lexical items by just underlining them and skipped them.”
A participant at the low level also made similar comments by saying that
“When I could readily get the general ideas of the text, I would just read on without bothering to recognize unfamiliar lexical items.”
The findings from the interviews echoes Bensoussan and Laufer’s claim (1984)
that general reading proficiency does not affect learners’ word guessing behavior and
the number of lexical items they successfully decipher.
The possible reasons underlying the high proficiency group’s inferior vocabulary
learning in some tests (test 1, test 2, and test 5) are generalized and induced from the
following detailed analysis and discussion. Table 4-24 is concerned with the mean
scores in pretest, immediate test and posttest of the participants at high and low
111
reading proficiency levels in every test.
Table 4-24: Vocabulary Learning in Each Test by Participants of High and Low Reading Proficiency
H L
M SD M SD t
Pretest (TW 1-20) 58.22 8.58 50.28 7.09 2.14**
Immediate test (TW 1-20) 70.33 8.40 60.67 6.84 2.68**
Post test (TW 1-20) 66.39 7.08 55.94 7.35 3.07**
Note. 1. H stands for participants with higher reading proficiency, L represents less proficient readers. N = 9 for both H (high) and L (low) groups. 2. TW means Target Word.
In Table 4-24, more proficient readers attain higher mean scores than the less
competent readers in pretest, immediate test and posttest in test 1, 2 and 5. However,
the H group does not make more progress than the L group in both vocabulary gain
and retention. To explore one step further, the researcher scrutinizes the performance
of these two groups in the target words of these three tests.
112
In test 1, these two groups have more differences in their scores of TW3
(aboriginal) and TW4 (tribe). As for vocabulary gain in TW3, more members in the H
group improve from total ignorance of word meaning to know how to use it
appropriately in the right context, while more L group members obtain lower scores
because their definition of this target word is acceptable in the pretest but
incompatible in the text in test 1. For example, they would proffer “primitive” as the
definition of “aboriginal”. However, its contextual meaning in the text in test 1 should
be related to tribes living in the mountains. So the L group’s scores in the immediate
test would be lower than those in the pretest, which makes their vocabulary gain of
TW3 slightly less than the H group’s. In TW4, most of the H group members could
provide the right definition or use the word in the pretest and they thus have little
room for improvement later. On the contrary, the L group only reports having seen the
word but it later could achieve the same level of understanding as the H group.
Consequently, it makes much more progress in this lexical item. In vocabulary
retention, both groups stick to their own definition for TW3 in the pretest and do not
acquire the contextual meaning. This occurs more in the L group and consequently the
L group’s scores in retention decrease more than the H group. In TW4, the H group
generally has a better command of this word, which leaves less room for gain in the
posttest. Contrary to this, a majority of L group members do not know TW4 quite well
113
in the beginning but they later obtain scores around 4 and 5. Due to such difference in
prior knowledge, the L group performs better than the H group in the vocabulary
retention of TW4.
In test 2, the inter-group differences consists mainly in TW6 (altar) and TW8
(preserve) of vocabulary gain and in TW5 (architecture) of vocabulary retention. In
TW6, both groups have or have not seen the word before. Whereas, the H group later
learns the word while the L group does not, which makes the former obtain higher
vocabulary gain than the latter. This could be accounted for by Stanovich’s remark
(1986) that more proficient readers could more effectively utilize the context to
decipher the meaning of the word. In addition, they could more readily recognize the
words in the surrounding text, which renders them more cognitive processing
resources to decode the word’s meaning. In TW8, more proficient readers know the
target word better in the beginning and do not gain much later after reading the text in
test 2. On the other hand, most L group learners improve from 2 to 4 or 5 in the
immediate test. Therefore, the L group outperforms the H group. The phenomenon in
TW8 also appears in TW5, which accounts for the reason why the less able readers
could attain higher vocabulary retention mean scores than more proficient readers.
In test 5, the H group differs from the L group in TW17 (abstract) and TW18
(anxiety) of vocabulary gain, TW17 (abstract) and TW19 (curriculum) of vocabulary
114
retention. Concerning their differences in vocabulary gain, some of the H group
members make improvement in their knowledge of TW17, while the L group does not
in this aspect. In TW18, the H group does not have a higher vocabulary gain due to
the ceiling effect. The L group does better because some of the members progressed
from two or three points in the pretest to four or five points in the immediate test. As
for the vocabulary retention, some participants in the H group achieved higher scores
in TW17, but the L group almost did not improve in their knowledge of the target
word. However, the opposite phenomenon is observed in TW19, in which the L group
outperforms the H group.
In conclusion, the high reading proficiency group’s less degree of progress
caused by its superior prior word knowledge together with these two groups’ similar
general English competence and word noticing and guessing behavior contributed to
such non-significance between reading proficiency and vocabulary learning of the
twenty target lexical items.
4.2.3 Vocabulary Gain and Retention and Participants’ Reading Comprehension
According to the results in section 4.1.3, learners who have better text
comprehension in both culturally familiar and unfamiliar texts do not absolutely
achieve more vocabulary gain and retention for the target words than those with less
well performance in reading comprehension. For instance, the former perform better
115
in vocabulary learning in the more familiar text (test5), but their mean scores of
vocabulary gain and retention were inferior to the latter in the less familiar article
(test3). In addition, the lexical development of the participants with better text
comprehension is eclipsed by their counterparts’ in most other tests (test1 and test2),
but the opposite is observed in test 4. Although there is no consistent pattern detected
from the aforementioned findings, the discrepancy between readers with higher scores
in reading comprehension and those with lower scores mostly do not reach
significance, which suggests that both groups of learners generally behave similarly in
vocabulary learning. Such findings could be attributed to the fact that they belong to
the same proficiency level define by their scores in reading and listening placement
tests. In addition, based on the participants’ responses from the interview as illustrated
in section 4.2.2, most interviewees report that they would pay more attention to get
the gist of the assigned passages because this is the required reading purpose of the
tests in the present study. In general, they just underline those unknown lexical items
and only attempt to decipher their meanings if necessary. This is in accordance with
Coady’s assertion (1993) that unfamiliar target words do not receive adequate
attention and are not processed deeply enough because of the lack of motivation when
their reading purpose is to get the main idea of the passage. Apart from this, the
process of constructing a mental representation of the text, according to Paribakht and
116
Wesche (1999), does not suffice for acquiring semantic features of new lexical items
during reading. Furthermore, like what Hong (2005) contends, reaching full
understanding of passages, which consists in macroprocessing, is conspicuously
insufficient for the amount of processing involved in the complicated vocabulary
learning. As a result, the improvement in text comprehension does not necessarily
entail the increase in learners’ vocabulary learning.
Although the discrepancy in lexical development between the participants with
better or inferior text comprehension do not reach statistical significance, the
performance of the former group in this aspect is surpassed by that of the latter group
in test 1, test2, test 3, and test 4, which is contradictory to my assumptions. To explore
the reasons behind, the researcher further analyzes the participants’ scores in terms of
pretest, immediate test and posttest in test 1, test 2, test 3, and test 4. Some factors are
summarized as follows. To begin with, the group with better text comprehension
might originally have superior understanding of certain target words and thus had less
room for improvement, which was the so-called ceiling effect, or it simply makes less
amount of progress than the group with lower mean scores in reading comprehension
in the word learning progress though both groups have the same entry behavior. This
is also possible because the participants in the present study are classified as the
intermediate students, which suggests their similar competence and capacity to learn
117
English. Secondly, more subjects in the group with less text comprehension might
self-evaluate their vocabulary from having never seen the word in the pretest (by
which they got 1 point) to having seen the word but unaware of its meaning (by which
they obtained two points). The accumulation of such improvement might amount to
much more gain than the group with better text comprehension, especially when the
participants in the latter group does not increase their knowledge (from 2 to 2). Last
but not least, the participants in the group with higher reading comprehension might
have offered in the pretest a certain definition which could win the participants three
to four points, but such definition is not compatible with the context in the immediate
tests and they are only awarded with two points, which in turn decreases their
vocabulary gain and retention. These concluding remarks are supported and drawn
from the following exhaustive analysis and discussion.
Based on the Table 4-13, readers with better text comprehension obtain less
mean scores in vocabulary gain and retention in test 1, 2, and 3, and in vocabulary
retention in test 4. To understand why this phenomenon might arise from, the
researcher examines the mean scores in the pretest, immediate test and posttest
between groups with higher and lower text comprehension. The results are illustrated
in Table 4-25.
118
Table 4-25: Scores in Pretest, Immediate Test, and Posttest of the Participants with High and Low Text Comprehension in Test 1, 2, 3 and 4
Tests H L
Pretest 14.85 2.11 11.67 2.23 3.42**
Immediate Test 15.70 2.59 15.29 2.89 0.35
Note. 1. H stands for the group of participants who obtained higher scores in reading comprehension (4 and 5), L
represents those who got lower scores (1 and 2). 2. Test 1: H (N = 16), L (N = 7); Test 2: H (N = 10), L (N = 12);
Test 3: H (N = 9), L (N = 12); Test 4: H (N = 8), L (N = 10).
From Table 4-25, the researcher finds that in test 1 the H group’s mean scores in
pretest, immediate test and posttest are lower than those of the L group’s, which
explains why the H group fails to achieve better vocabulary learning for target words
in this test. As for test 2, more proficient readers obtain higher scores in the pretest
and the immediate test, but lower scores in the posttest than less competent readers.
The H group consequently attains lower vocabulary gain and retention due to less
improvement. In test 3, the H group has a superior performance in the pretest, but it
obtains a less mean score in the immediate test, which leads to lower vocabulary gain.
119
With regard to the vocabulary retention, the H group performs less well even though it
has higher mean scores in the pretest and the posttest. Similar phenomenon is also
detected in the H group’s vocabulary retention in test 4 due to less progress. In brief,
the H group does less well than the L group in incidental vocabulary learning in these
tests mainly due to two reasons: 1. participants in H group might possess similar prior
knowledge about target words, but they fail to make more progress later. 2. the
participants might have more advanced prior knowledge, which leaves them less room
for improvement later.
To have a more precise insight into these findings, the reseacher further analyzes
the data to find out which target words in each test result in these phenomena by
soliciting two groups of subjects: participants who get a score of 4 and 5 belong to the
H group, while others who only obtain 1 and 2 are classified into the L group.
For the analysis of the disparity of vocabulary learning between the H and L
groups in test 1, the mean scores of all the target words in pretest, immediate test and
posttest are illustrated in Table 4-26. Table 4-27 summarizes their vocabulary gain and
retention in the four target lexical items in test 1.
120
Table 4-26: Scores of the First Four Targets in Pretest, Immediate Test and Posttest by Participants' of Different Reading Comprehension
H L
(aboriginal) 2.53 1.45 3.00 1.53 -0.70
Pretest
(aboriginal) 2.94 1.24 3.29 1.60 -0.57
Immediate test
(aboriginal) 2.94 1.34 3.71 1.60 -1.21
Posttest
TW4
(tribe) 4.31 1.08 4.71 0.76 -0.89
Note. 1. H represents participants with higher reading comprehension, L stands for those with lower reading comprehension. 2. H (N = 16), L (N = 7). 3. TW means Target Word.
121
Table 4-27: Vocabulary Gain and Retention of the Target Words in Test 1 by Participants of Different Reading Comprehension
Vocabulary Learning TW H L
(aboriginal) 0.41 1.45 0.29 1.98 0.16 Vocabulary gain
(aboriginal) 0.41 1.45 0.71 2.21 -0.40 Vocabulary retention
TW4
(tribe) 1.00 1.32 1.57 0.98 -1.03
Note. 1. H represents participants with higher reading comprehension, L stands for those with lower reading comprehension. 2. H (N = 16), L (N = 7). 3. TW means Target Word.
In test 1, the differences between the H and L groups in vocabulary gain arise
mainly from TW1 and TW4 (see Table 4-27), where the former obtains less
vocabulary gain than the latter. Both groups have similar performance in TW1 and
TW4 in the pretest and immediate test (see Table 4-26), and their differences do not
show any significance. So the inferior vocabulary gain of the H group in test1 is
attributed to its less improvement. As for vocabulary retention (see Table 4-27), the
122
mean scores of both H and L groups in each target word are close to each other.
Besides, these two groups have similar scores in both pretest and posttest because no
significance is shown (see Table 4-26). Therefore, the L group outperforms the H
group in the mean score for the overall vocabulary retention because of its more gain
in TW4. On the whole, the vocabulary gain and retention in test 1 are proven to bear
much resemblance between the H and the L groups.
As for test 2, the discrepancy of mean scores of all the target words in the pretest,
immediate test and the posttest is presented in Table 4-28. Their vocabulary gain and
retention for each target word are summarized in Table 4-29.
123
Table 4-28: Scores of Target Words in Test 2 in Pretest, Immediate Test and Posttest by Participants of Different Reading Comprehension
Tests TW H L
Tests TW H L