To find out the reasons behind the non-significance in the effects of cultural
familiarity on the participants’ incidental vocabulary learning and answer the first
research question, the researcher analyzes one step further and discusses the data
presented in section 4.1.1 from different dimensions here. The first research question
is restated as follows.
(1) When the participants’ are required to read assigned passages mainly for the
purpose of global comprehension and the gist, does less cultural familiarity in
planned reading yield significantly better results in learners’ vocabulary gain and
retention of selected twenty target words?
To start with, the participants’ lexical development among all five texts of
diverse degree of cultural familiarity ratings will be compared together. Furthermore,
87
the influence of the participants’ discrepant perception toward the cultural familiarity
of the same text in each test on their incidental vocabulary learning will be explored
by comparing the scores in vocabulary gain and retention first between the
participants who assign high rankings and those who give low rankings, and then
between those who rate the text high (4 and 5) in the Likert scale and those who rate
low (1 and 2).
The Comparison of the Participants’ Vocabulary Gain and Retention among All Texts
The non-significant differences (see Table 4-4) in vocabulary gain and retention
of five texts of different cultural familiarity ratings might be attributed to the similar
cultural familiarity ratings received by all the five texts. Students consider these five
texts to be of similar cultural familiarity because no significance is found among the
average scores of these cultural familiarity ratings measured by Likert Scale. Due to
this reason, the vocabulary gain and retention of each test (see Table 4-4) is not
significantly distinct from one another in spite of the slightly differences in ratings of
cultural familiarity. This finding is in keeping with the results shown in the translation
production task in Pulido’s study (2004). The data from translation recognition task
(Pulido, 2004) reveal slightly more gain from the more familiar text. The findings in
Pulido’s (2004) study concerning cultural familiarity and those in Hong’s study (2005)
88
reveal that subjects obtain more gains from texts with greater cultural and topic
familiarity. However, such pattern is not consistently supported in the present study
where the vocabulary learning in more culturally familiar texts can be more or less
than that in less culturally familiar texts.
Because no significant difference is detected in the participants’ ratings of
cultural familiarity among the selected five texts (see Table 4-4), the influence of
diverse extent of cultural familiarity on the participants’ lexical development could
not be readily observed and explored. Therefore, the comparison of test 3 (Puppet
Theater: less culturally familiar) and test 5 (Math Anxiety: more culturally familiar)
whose cultural familiarity ratings exhibit statistical significance (see Table 4-5) is
made to find out the answers to the first research question. According to Table 4-6, the
participants obtain a slightly higher mean score for vocabulary gain but a lower mean
score for vocabulary retention in less familiar test (test 3). Some previous researchers
(Graesser & Nakamura, 1982; Graesser, Woll, Kowalski, & Smith, 1980; Rizzella &
O’Brien, 2002; Yekovich & Walker, 1986) maintain that the lower lexical gain in the
culturally familiar text might be attributed to high level of source confusion which is
triggered by the competition between activation from background knowledge and that
from episodic memory gleaned from text messages. In other words, they might have a
hard time distinguishing between words actually embedded within the passage and
89
those closely concerned with the theme but never explicitly presented. This in turn
thwarted the participants’ attempt to notice the outward manifestation of the target
words, let alone learning their meanings from the context. To understand why the
participants’ mean scores in vocabulary gain and retention in less culturally familiar
text (test 3) were not consistently higher than those in more culturally familiar text
(test 5), the researcher further examined their performance in terms of pretest and
immediate test between test 3 and test 5, as shown in Table 4-17.
Table 4-17: The Participants' Performance in Pretest, Immediate Test, and Posttest between Test 3 and Test 5
Test3 Test5 Test
M SD M SD t
Pretest 8.69 2.03 12.40 2.79 -7.38**
Immediate test 11.46 2.65 14.90 2.94 -5.69**
Post test 9.69 2.48 14.16 3.19 -8.03**
Note. N = 34.
From Table 4-17, the researcher found that they consistently got significantly
higher scores in the pretest, immediate test, and posttest in test 5 than in test 3, which
revealed that they knew target words better in the text of higher cultural familiarity
and they thus had less room for improvement for those target words in test 5 and
achieved less vocabulary gain. On the other hand, the participants’ higher vocabulary
retention in the more familiar test could be ascribed to their superior ability to learn
90
unknown lexical items and the assistance of the text which could help them recall the
contextual meanings of target lexical items better. This echoes Nation and Coady’s
contention (1988) that the background knowledge could be capitalized on by the good
readers to help anticipate the occurrence of certain words, which in turn assists the
participants in inferring word meanings from the context. Consequently, their scores
in the immediate test and the posttest were quite close to each other, which testified
that they still retained a similar degree of word knowledge for those target words in
test 5 after reading the text and they thus obtained higher scores in vocabulary
retention. On the contrary, the less cultural familiarity of test 3 did not help students’
retention of target words as efficiently as that of test 5. As a result, the participants’
learning of target words in the immediate test could not remain at the same level in
the posttest and faded away over time there, which rendered a lower mean score in
vocabulary retention.
Based on the results in these two tests (test 3 and test 5), different cultural
familiarity might not leave any significant impact on learners’ vocabulary gain and
retention. However, the participants may recall the target words in more familiar text
better for they could more readily and quickly associate them with the content and
thus performed better in vocabulary retention. The aforementioned discrepancy in
vocabulary acquisition in culturally familiar and unfamiliar texts principally did not
91
reveal statistical significance, which showed that the participants’ performance was
basically similar regardless of the cultural familiarity of the texts. Besides, higher
familiarity in test 5 did not really significantly facilitate their general understanding of
the main ideas, which was verified by the participants’ mean scores in the reading
comprehension section presented in Table 4-6. As a result, they did not necessarily
allot more or less attention to the target words, which in turn brought about
insignificant vocabulary learning. This was in contrast with previous researchers’
assumptions (Haastrup, 1989; Hulstijn et al. 1996; Jiang, 2000; Parry, 1997) that
higher cultural or topic familiarity would increase the ease of attaining the global
picture of the assigned passage, which made readers focus less efforts on unknown
words and consequently fail to notice them. When the researcher compares the
participants’ vocabulary gain with their vocabulary retention, the mean scores of the
latter were lower due to attrition over time, which echoes previous researchers’
findings (Pulido, 2004; Hong, 2005). Though there might be some subtle differences
between the scores of these two tests in vocabulary gain and retention, no significance
was shown, which indicated that topic familiarity did not significantly influence
learners’ performance in vocabulary learning of target words.
The researcher further explores whether those who regard the text as less
culturally familiar would get higher scores than others who perceive the text as more
92
culturally familiar in vocabulary gain and retention in each test. From the data in some
tables (Table 4-7 to Table 4-8), low-ranking givers generally obtain lower mean scores
than high-ranking learners in vocabulary gain and retention, the same phenomenon
also appears in the comparison between low-rating and high-rating the participants
(see Table 4-9 to Table 4-10). This further supports Adam’s assertion (1982) that
learners could infer the words more successfully if they are more aware of the topics.
However, the disparity is not statistically significant. This might stem from the similar
perception the participants hold toward all the five texts in terms of cultural familiarity.
In the beginning, the researcher asks them to give rankings instead of evaluating these
five texts measured by means of Likert scale. Although these texts are ranked
differently, their distinction in cultural familiarity might not be noticeable enough,
which suggests that these five texts might originally be considered culturally familiar
to a similar extent. This assumption is later confirmed by the insignificance in the
topic familiarity ratings of five texts by Likert scale in Table 4-4. In brief, these five
texts are subsumed under the issue “Taiwanese Culture” and are consequently
regarded not very different in terms of cultural familiarity. As a result, the effect of
cultural or topic familiarity might not be significantly salient. With regard to why the
vocabulary learning of the participants who perceive the text more culturally familiar
is inferior to that of their counterparts with less culturally-familiar perception, some
93
major reasons are found as follows. First, the participants who consider the text more
culturally familiar might gain less in lexical development due to their less extent of
progress, their definition which is originally appropriate in the pretest but
incompatible in the context of the assigned texts, the ceiling effect caused by their
superior prior knowledge of the target words. Besides, their incidental vocabulary
learning might be eclipsed by their counterparts because those who regard the text as
culturally unfamiliar might improve from total innocence to the impression of having
seen the word in Vocabulary Knowledge Scale (from 1 point to 2 points), while the
former group do not show any improvement (from 2 points to 2 points). These
possible explanatory comments are induced and attained from the following
discussion.
94
Different Participants’ Rankings of Cultural Familiarity and Their Vocabulary Gain and Retention in Each Text
Among the mean scores in vocabulary learning between high-ranking and
low-ranking groups, the former only performed less well than the latter in vocabulary
gain in test 4 (see Table 4-7). To clarify the cause of such phenomenon, the researcher
examined these two groups’ mean scores in pretest and immediate test presented in
Table 4-18, and their performance in each target word along with vocabulary gain in
Table 4-19.
Table 4-18: The Participants' Performance in Pretest and Immediate Test in Test 4
HRk LRk Tests
M SD M SD t Pretest 9.11 1.90 8.57 2.01 0.66 Test 4
Immediate Test 11.30 1.50 10.80 2.49 0.54
Note. HRk means participants who gave high ranking (4 and 5) to the text in test 4, and LRk means those who gave low ranking (1 and 2).
95
Table 4-19: The Participants' Vocabulary Learning for Each Target Word in Test 4
Note. 1. HRt means participants who gave high ranking (4 and 5) to the text in test 4, and LRt means those who gave low ranking (1 and 2). 2. HRt (N = 9), LRt (N = 15). 3. TW represents Target Word.
In Table 4-18, the mean scores in pretest and posttest for HRk group were higher
than those for LRk group. The vocabulary gain of the HRk group was lower than that
96
of the LRk group due to less lexical improvement. When examined in more detail in
Table 4-19, though the HRk group obtained better vocabulary gain in TW 13 and TW
16 but worse scores in TW 14 and TW 15 than LRk. Regardless of the differences in
the vocabulary gain of each target word, no significance is found.
The data are further analyzed by randomly selecting nine participants
respectively from the high-ranking and low-ranking groups mentioned above.
Likewise, those who consider topics to be more culturally familiar tend to perform
better in most tests, but not to a significant degree. The performance of the
high-ranking and low-ranking groups in vocabulary gain and retention in test 4, and in
the vocabulary gain in test 5 is illustrated in Table 4-20. The former group obtain
lower mean scores than the latter group in these aspects.
Table 4-20: The Participants' Vocabulary Learning in Test 4 and Test 5
HRk LRk
Tests M SD M SD t
Pretest 9.11 1.90 8.22 2.11 0.94*
Immediate Test 11.33 1.50 10.78 2.95 0.50 Test 4 Posttest 11.89 2.42 11.56 2.65 0.28 Pretest 14.06 2.53 11.33 3.20 2.00 Immediate Test 15.89 2.91 13.67 3.30 1.52 Test 5 Posttest 15.67 3.00 12.61 3.16 2.10
Note. 1. N = 9 for both H (high) and L (low) groups. 2. HRk means the group which gave higher rankings to texts (4 and 5), LRk is the group which assigned lower rankings to texts (1 and 2).
97
With regard to test 4, the high-ranking group’s performance in pretest, immediate
test and posttest is better than that of the low-ranking group, and its mean score in the
pretest is even significantly higher than its counterpart. Nevertheless, its vocabulary
gain and retention are inferior to those of the latter group due to less improvement in
lexical knowledge. To be more specific, the LRk group achieves higher vocabulary
gain mainly in TW15 (compound) and TW16 (extend). As for the TW15 (compound),
some participants of the LRk group made progress in their self-evaluation of
knowledge for the target word “compound” from 1 (never seen this word) to 2
(having seen this word but do not know its meaning) or from 2 to 3 (having seen the
word and probably know it meaning in the right semantic field), while the HRk group
know other definitions of this word (e.g. “compound” sentences) and could get a
higher score (3 or 4) in pretest, but such definition is infelicitous in the context of the
assigned passage (e.g. an area that contains a group of buildings and is surrounded by
a fence or wall) in test 4 and it thus gets lower grades in the immediate test. For TW16
(extend), the better vocabulary gain of the LRk group lies in the fact that some of the
participants’ scores move from 1 to 2 and thus obtaine some scores, while most HRk
group participants fail to use the word correctly in terms of the part of speech and thus
attained less scores in the immediate test.
Concerning the vocabulary retention in test 4, the differences between the HRk
98
and the LRk group consist primarily in their performance in TW15 (compound) and
TW16 (extend). In TW15, the LRk group obtains more vocabulary retention because
most HRk group learners report that they have seen this word in the pretest but do not
make any progress in posttest but three participants in the L group improves from one
point to two points, and another one who originally does not know the word is even
able to use the word semantically and syntactically appropriate in a sentence. In
TW16, the HRk group performs better because two to three participants could retain
this word better and make more improvement.
Speaking of vocabulary gain in test 5, the HRk group slightly falls behind the
LRk group. As shown in Table 4-20, though it has higher mean score than the LRk
group in pretest and immediate test, its vocabulary gain does not surpass the LRk
group due to its less amount of progress. This phenomenon arises mainly from the
differences between these two groups in TW17 (abstract) and TW18 (anxiety). In
TW17, the HRk group does better because some of the participants learn more about
the target word (from 2 or 3 to 5), while some learners in the LRk group only gets one
point higher than their scores in the pretest. In TW18, the LRk group performs better
because almost all the members show increase in their scores and could use the target
word felicitously. However, some learners in the HRk group who originally knew
more about the target word does not use it properly and one of them even forgets what
99
it means in the immediate test. That widens the gap in the vocabulary gain between
these two groups in TW18.
Different Participants’ Ratings of Cultural Familiarity and Their Vocabulary Gain and Retention in Each Text
The participants’ incidental vocabulary learning is further probed into by
examining the data collected when the cultural familiarity of each text is determined
by Likert Scale. Those who give higher ratings to topics tend to, though not
significantly, perform better in most tests. However, as shown in Table 4-9, the
high-rating group obtains lower mean scores in vocabulary gain and retention in test 2,
and in the vocabulary gain in test 3. The reasons behind such phenomena were
analyzed and discussed in the following section.
For more detailed analysis of the participants’ lexical development in test 2 and
test3, the results of their mean scores in the pretest, immediate test and posttest were
shown in Table 4-21.
Table 4-21: The Participants' Vocabulary Learning in Test 2 and Test 3
HRt LRt
Note. 1. HRt means the group which gave higher ratings to texts (4 and 5), LRt is the group which assigned lower ratings to texts (1 and 2). 2. Test 2: HRt (N = 10), LRt (N = 10); Test 3: HRt (N = 8), LRt (N = 12)
100
In test 2 (see Table 4-21), the high-rating group’s slightly less vocabulary gain is
attributed to less extent of improvement despite its higher mean scores in the pretest
and the immediate test. In more detail, the participants’ performance in TW6 (altar),
TW7 (unique) and TW8 (preserve) contributes to the difference in vocabulary gain
between these two groups. In TW6, most learners in the HRt group do not produce
any gain and remain at the same level of knowledge (the HRt group gets 2 points in
the Vocabulary Knowledge Scale), while those in the LRt group improve from one to
two points and one participant could even use the word appropriately. In TW7, almost
all HRt group members obtain a score of five in both pretest and posttest, which
means that they know the usage of the target word quite well and demonstrate the
ceiling effect in the VKS form; but some students in the LRt group do not possess
such full understanding of this word in the pretest, which later leads to better scores in
vocabulary gain for the LRt group. In TW8, more participants in the HRt group
improve from two to five points than those in the LRt group, in which one member
even gets a lower score in the posttest because he or she mistakes “preserve” for
“reserve.” As a consequence, the LRt group is inferior in scores of vocabulary gain in
TW8. In addition to vocabulary gain, the HRt group also achieves less vocabulary
retention than the LRt group. It could be accounted for by the fact that the HRt group
101
has more prior word knowledge but lower scores in the posttest. In more detail, the
difference in lexical retention between the HRt group and the LRt group mainly stems
from their distinct performance in TW6 and TW7. In TW6, some participants in the
HRt group self-evaluate from two points in the pretest to one point in the posttest, but
such case does not appear in the L group, where most students report from one point
in the pretest to two points in the posttest. In TW7, the HRt group attains a lower
mean score because some of them erroneously consider “special” to be the synonym
for “unique.” However, most of LRt group’s members could make an appropriate
sentence with this word and their knowledge of it remains at the same level in the
posttest.
Compared with the LRt group, the HRt group’s lower mean score in vocabulary
retention in test 3 results from its less improvement caused by their higher mean score
in the pretest but lower mean score in the posttest, as shown in Table 4-21. The
participants’ mean scores of each target word in test 3 are illustrated in Table 4-22 for
further analysis of the aforementioned phenomenon.