Paired-Samples t-tests were conducted, testing the effects of categorical independent variables of reading proficiency and topic familiarity on the three continuous

CHAPTER FOUR

RESULTS AND DISCUSSIONS

4.1. Results and Analysis

For data analysis, the researcher used the Statistical Package for Social Science (SPSS, version 13.0) to analyze the collected data. Firstly, in order to determine the effects of the independent variables of topic familiarity and reading proficiency on the dependent variables of reading comprehension, vocabulary gain and retention,

Paired-Samples t-tests were conducted, testing the effects of categorical independent variables of reading proficiency and topic familiarity on the three continuous

variables of reading comprehension, vocabulary gain and retention. And also to compare the effect of reading proficiency across high and low groups on reading comprehension, vocabulary gain and retention, an Independent-Samples t-test was conducted. In addition, the relationships among participants’ vocabulary size, reading proficiency, reading comprehension vocabulary gain and retention was examined by the results of Pearson correlations analysis

4.1.1. Overall Results

Table 4-1 is a summary table of overall results; the following findings can be drawn from the overall results in the Table 4-1.

1. Readers who read the more familiar text (Mean = 31.06) got higher scores in

reading comprehension than the less familiar text (Mean = 24.71). And the

difference showed significance, t(169) = 5.81, p = < .01; see Table 4-3.

2. For intermediate readers, they had higher reading comprehension scores of the more familiar text (Mean = 38.24) than of the less familiar text (Mean = 33.82), though the difference of reading comprehension scores across the two texts did not reach significance, t(33) = 1.71, p = > .05; see Table 4-3.

3. For elementary readers, they had higher reading comprehension scores of the more familiar text (Mean = 20.00) than of the less familiar text (Mean = 16.76), though the mean difference of reading comprehension scores across the two texts did not reach significance, t(33) = 2.00, p = > .05; see Table 4-3.

Table 4-1 A Summary Table of Overall Results All (N =170)

(M & SD)

H (n =34) (M & SD)

L (n = 34) (M & SD)

Vocabulary Size 40.49 7.20 48.74 8.23 34.76 4.35

GEPT 57.55 20.78 87.71 8.85 29.15 6.91

FAM Vocab (TW) 18.65 4.81 24.44 5.53 15.38 2.67 Pre-test

UNF Vocab (TW) 17.71 4.37 22.21 5.55 15.12 2.88 Vocab (TW) 21.68 5.28 28.41 6.29 18.59 2.80

FAM RC 31.06 13.01 38.24 11.41 20.00 10.15

Vocab (TW) 19.82 4.53 24.68 5.50 16.85 3.10 Immediate

Test

UNF RC 24.71 12.41 33.82 12.06 16.76 10.65

FAM Vocab (TW) 20.84 5.02 27.32 6.08 17.65 2.32 Post-test

UNF Vocab (TW) 19.08 4.40 23.94 5.33 16.12 2.88

Note. FAM = the familiar text; UNF = the unfamiliar text.

Table 4-2 is a summary table of vocabulary gain and retention. The following findings can be drawn from the overall results in Table 4-2:

4. Readers who read the more familiar text got significantly higher scores in

vocabulary gain (Mean = 3.03) and retention (Mean = 2.11) than the vocabulary

gain (Mean = 2.18) and retention (Mean= 1.37) in the less familiar text, but the

scores of vocabulary retention were weaker due to the attrition in vocabulary gain

over time, t(169) = 4.48, p < .01; t(169) = 5.05, p < .01; see Table 4-3.

5. For intermediate readers, they achieved higher vocabulary gain (Mean = 3.97) and retention (Mean = 2.88) in the more familiar text than in the less familiar text (Mean of gain = 2.47; Mean of retention = 1.74). And there were significant vocabulary gain and retention for H group in both texts. The scores of gain and retention while reading the more familiar text were significantly higher than those in the unfamiliar text, which meant their gain and retention disparity across the two texts reached significance, t(33) = 2.30, p < .05; t(33) = 2.17, p < .05; see Table 4-3.

6. For elementary readers, they also achieved higher vocabulary gain (Mean = 3.21) and retention (Mean = 2.26) both in the more familiar text and in the less familiar text (Mean of gain = 1.74; Mean of retention = 1.00). And there were significant gain and retention for L group in both texts. Noticeably, L group also had significant gain and retention in the less familiar text. The scores of gain and retention in the more familiar text were significantly higher than those in the unfamiliar text, which meant their gain and retention disparity across the two texts reached significance, t(33) = 3.91, p < .05; t(33) = 4.45, p < .05; see Table 4-3.

Table 4-2 A Summary Table of Vocabulary Gain and Retention All(N=170)

(M & SD)

H (n=34) (M & SD)

L (n=34) (M & SD)

FAM 3.03 2.49 3.97 2.89 3.21 2.59

Vocabulary

Gain UNF 2.11 1.93 2.47 2.22 1.74 1.52

FAM 2.18 1.87 2.88 2.38 2.26 1.85

Vocabulary

Retention UNF 1.37 1.44 1.74 1.78 1.00 1.02

Note. FAM = the familiar text; UNF = the unfamiliar text.

Table 4-3 A Summary Table of Paired-Samples t-tests

ALL (N=170) M SD t p

Comprehension FAM 31.06 13.01 5.81 .00**

UNF 24.71 12.41

Vocabulary Gain FAM 3.03 2.49 4.48 .00**

UNF 2.11 1.93

Retention FAM 2.18 1.87 5.05 .00**

UNF 1.37 1.44

H (n=34)

Comprehension FAM 38.24 11.41 1.71 .10

UNF 33.82 12.06

Vocabulary Gain FAM 3.97 2.89 2.30 .03*

UNF 2.47 2.22

Retention FAM 2.88 2.38 2.17 .04*

UNF 1.74 1.78

L (n=34)

Comprehension FAM 20.00 10.15 2.00 .05

UNF 16.76 10.65

Vocabulary Gain FAM 3.21 2.59 3.91 .00**

UNF 1.74 1.52

Retention FAM 2.26 1.85 4.45 .00**

UNF 1.00 1.02

Note. **p < .01, *p < .05; FAM = the familiar text; UNF = the unfamiliar text.

4.1.2. Comprehension of Different Topic Familiarity

Table 4-3 showed the descriptive statistics for reading comprehension of the

more and less familiar passages. The maximum and minimum of scores showed a

considerable range in proficiency across participants. The result showed that

participants have better reading comprehension score on the more familiar text. On

average, participants answered 62% (Mean = 31.1) of comprehension questions correctness of the familiar text, whereas participants answered 50% (Mean = 24.7) of correctness of the less familiar text. The total reading comprehension showed that most participants answer 55.77% (M = 55.77) correctness across the two texts.

The result of paired-samples t-test of reading comprehension of different topic familiarity (see Table 4-3) showed statistical significance, t(169) = 5.81, p = . < .01.

This result revealed that participants had higher reading comprehension of the more familiar text than of the less familiar text

4.1.3. Vocabulary Gain and Retention of Different Topic Familiarity Table 4-3 illustrated the descriptive statistics for the continuous dependent variable of vocabulary gain and retention. The mean score (Mean = 3.03) of vocabulary gain in the more familiar text was higher than that (Mean =2.11) in the less familiar text. This showed that participants achieved higher vocabulary gain in the more familiar text than in the less familiar text. And the differences between

vocabulary gain scores in the more and less familiar texts were statistically significant, t(169) = 4.48, p <.05; see Table 4-3). Also, the mean score of vocabulary retention

(Mean = 2.18) from the more familiar text was higher than that (Mean = 1.36) from the less familiar text. This showed that participants retained higher vocabulary gain in the more familiar text than in the less familiar text. And the differences between vocabulary retention scores in the more and less familiar texts were statistically significant, t(169) = 5.05, p <.05; see Table 4-3).

In short, the results above in Table 4-3 revealed that readers who read the

familiar text had higher and significant reading comprehension than the less familiar

text. Also, the results in Table 4-3 showed that readers who read the familiar text had

higher and significant vocabulary gain and retention than the less familiar text. These

findings conformed to the researcher’s expected outcome. Also as expected, the scores of vocabulary retention in both texts were lower due to the attrition in vocabulary gain over time.

4.1.4. Grouping by Reading Proficiency

Table 4-4 provided descriptive statistics for the variable of English reading proficiency, revealing a considerable range in scores. The full mark for this GEPT reading proficiency was 120. The minimum score of participants was 11, and the maximum score was 108. Participants scoring over 80 can pass the GEPT reading proficiency test. Because the mean score was 57.6., it was clear averagely the scores of most participants are just at elementary level, in terms of reading ability.

Table 4-4 showed the grouping of high and low (H & L) groups categorized by scores of reading proficiency. The first and last 20% of participants (34 for each) were selected as high and low groups. The mean score of H was 87.71; the maximum score was 108, and the minimum was 77. The mean score of L was 29.15; the maximum score was 37, and the minimum was 11. On average, the high group passed 80 (elementary level) and were intermediate readers in this study; the low group got the score under 40, there were defined as elementary readers in this study.

Table 4-4 H and L Groups Divided by English Reading Proficiency

Group M N SD Min. Max.

H 87.71 34 8.85 77.00 108.00

L 29.15 34 6.91 11.00 37.00

All 57.55 170 20.78 11.00 108.00

Note. H Group (n = 34); L Group (n = 34); All = 170.

4.1.5. Paired-Samples t-tests on Reading Comprehension, Vocabulary Gain and Retention of Different Topic Familiarity

In order to determine the effect of the categorical independent variable of topic familiarity (familiar & unfamiliar texts) on the dependent variable of reading

comprehension, vocabulary gain and retention in the more and less familiar texts, Paired-Samples t-tests were conducted.

Table 4-3 revealed that H group achieved higher reading comprehension of the more familiar text than of the less familiar text (M

= 38.24 > M

= 33.82), though the mean difference did not reach significance, t(33) = 1.71, p >.05; see Table 4-3.

Similarly, Table 4-3 showed that L group achieved higher reading.

comprehension of the more familiar text than of the less familiar text (M

= 20.00 >

M

= 16.76), though the mean difference did not reach significance, t(33) = 2.00, p

> .05; see Table 4-3.

These results displayed that across both H & L groups, the scores of reading comprehension of the more familiar text were higher than of the less familiar text, though the difference between the two texts did not reach statistical significance.

In another aspect, Paired-Samples t-tests were adopted to examine if there were significant vocabulary gain and retention in both texts across the H and L groups. In comparing the scores of the pretest on TWs of the familiar text to the scores of the immediate test, Table 4-5 showed that H group had significant vocabulary gain in the more familiar text, t(33) = 8.02, p < .01). And when comparing the scores of the pretest on TWs of the familiar text to the scores of the posttest, H group had

significant vocabulary retention in the more familiar text, t(33) = 7.05, p < .01. In the unfamiliar text, in comparing the scores of the pretest on TWs to the scores of the immediate test, H group had significant vocabulary gain in the less familiar text, t(33)

= 6.49, p < .01. And when comparing the scores of the pretest on TWs to the scores of

the posttest, H group had significant vocabulary retention in the less familiar text, , t(33) = 5.68, p < .01.

Table 4-5 Vocabulary Gain and Retention in Both Familiar and Unfamiliar Texts across H and L Groups (Paired-Samples t-test)

Group Vocabulary Gain M SD t p

Pretest on TWs of Familiar

Text 24.44 5.53 8.02 .00**

H Immediate Test on TWs of

Familiar Text 28.41 6.29

Pretest on TWs of Familiar Text

24.44 5.53 7.05 .00**

H Posttest on TWs of Familiar Text

27.32 6.08

Pretest on TWs of Unfamiliar Text

22.21 5.55 6.49 .00**

H Immediate test on TWs of Unfamiliar Text

24.68 5.50

Pretest on TWs of Unfamiliar Text

22.21 5.55 5.68 .00**

H Posttest on TWs of Unfamiliar Text

23.94 5.33

Pretest on TWs of Familiar Text

15.38 2.67 7.21 .00**

L Immediate Test on TWs of Familiar Text

18.59 2.80

Pretest on TWs of Familiar Text

15.38 2.67 7.15 .00**

L Posttest on TWs of Familiar Text

17.65 2.32

Pretest on TWs of Unfamiliar Text

15.12 2.88 6.64 .00**

L Immediate Test on TWs of Unfamiliar Text

16.85 3.10

Pretest on TWs of Unfamiliar Text

15.12 2.88 5.75 .00**

L Posttest on TWs of Unfamiliar Text

16.12 2.88

Note. **p < .01, *p < .05; H Group (n = 34); L Group (n = 34)

The same comparison was applied to the L group. In comparing the scores of the

pretest on TWs of the familiar text to the scores of the immediate test, Table 4-5

showed that L group had significant vocabulary gain in the more familiar text, t(33) =

7.21, p < .01. And when comparing the scores of the pretest on TWs of the familiar text to the scores of the posttest, L group had significant vocabulary retention in the more familiar text, t(33) = 7.15, p < .01. In the unfamiliar text, in comparing the scores of the pretest on TWs to the scores of the immediate test, L group had

significant vocabulary gain in the less familiar text, t(33) = 6.64, p < .01. And when comparing the scores of the pretest on TWs to the scores of the posttest, L group had significant vocabulary retention in the less familiar text, t(33) = 5.75, p < .01. From the results in Table 4-5, there were significant vocabulary gain and retention in both texts across the H and L groups.

To conclude to results of Table 4-3 and 4-5, these results showed that both H &

L groups achieved higher vocabulary gain in the familiar text than in the less familiar text. For H group participants, they had higher gain in the more familiar text (M

= 3.97 > M

= 2.47; see Table 4-2). Also, for L group, they had better gain in the more familiar text than in the less familiar text (M

= 3.21 > M

= 1.74; see Table 4-2). In comparing the gain difference between the two texts for H group, Table 4-3 showed that the vocabulary gain difference across two texts reached significance, t(33) = 2.30, p < .05, which meant that H group had significantly higher vocabulary gain while

reading the more familiar text than the less familiar text. In comparing the retention difference between the two texts for H group, Table 4-3 showed that the retention difference across two texts reached significance, t(33) = 2.17, p < .05, which meant that H group had significantly higher retention two weeks after reading the more familiar text than the less familiar text. In comparing the gain difference between the two texts for L group, Table 4-3 showed that the gain difference across two texts reached significance, t(33) = 3.91, p < .01, which meant that L group had significantly higher vocabulary gain while reading the more familiar text than the less familiar text.

In comparing the retention difference between the two texts for L group, Table 4-3

showed that the retention difference across two texts reached significance, t(33) = 4.45, p < .01, which meant that L group had significantly higher retention two weeks after reading the more familiar text than the less familiar text.

In sum, the results of Paired–Samples t-tests revealed several findings:

1. For intermediate readers (H group), they had higher reading comprehension scores of the more familiar text (Mean = 38.24) than the less familiar text (Mean = 33.82), though the difference of reading comprehension scores across the two texts did not reach significance.

2. Intermediate readers achieved higher vocabulary gain (Mean = 3.97) and retention (Mean = 2.88) in the more familiar text than in the less familiar text (Mean of gain

= 2.47; Mean of retention = 1.74). There were significant vocabulary gain and retention for intermediate readers in both texts. And the scores of gain and

retention while reading the more familiar text were significantly higher than those in the unfamiliar text, which meant their gain and retention disparity reached significance.

3. For elementary readers (L group), they had higher reading comprehension scores of the more familiar text (Mean = 20.00) than the less familiar text (Mean = 16.76), though the difference of reading comprehension scores across the two texts did not reach significance.

4. Similarly, elementary readers also achieved higher vocabulary gain (Mean = 3.21)

and retention (Mean = 2.26) both in the more familiar text and in the less familiar

text (Mean of gain = 1.74; Mean of retention = 1.00). And there were significant

gain and retention for elementary readers in both texts. Noticeably, elementary

readers also had significant gain and retention in the less familiar text. And the

scores of gain and retention in the more familiar text were significantly higher

than those in the unfamiliar text, which meant their disparity reached significance.

In general, the results in Table 4-3 revealed that topic familiarity facilitate vocabulary gain while reading and retention after reading.

4.1.6. Independent-Samples t-tests on Reading Comprehension, Vocabulary Gain and Retention of Different Topic Familiarity across H and L Groups In order to determine the effect of the independent variable reading proficiency (H & L groups) on the dependent variable of reading comprehension, vocabulary gain and retention in the more and less familiar texts, an Independent-Samples t-test was conducted, testing the vocabulary gain and retention between H & L groups in texts of different topic familiarity.

The results revealed that across both two texts, H group achieved better reading comprehension than L group. Table 4-6 showed that H group achieved higher

comprehension than L group of both texts (M

= 38.24 > M

= 20.00; M

= 33.82

> M

= 16.76). Table 4-6 displayed that there was significant difference, t(66) = 6.96, p < .01, between the mean scores of comprehension between the two groups. These

results revealed that readers’ different reading proficiency level significantly affected reading comprehension while reading texts of different topic familiarity.

In addition, Table 4-6 showed that in both more familiar and less familiar texts, H group achieved higher vocabulary gain than L group. The vocabulary gain of H group in the familiar text is higher than that of L group (M

= 3.97 > M

= 3.21) though the gain difference showed no significance, t(66) = 1.15, p > .05. Next, the vocabulary gain of H group in the less familiar text is higher than that of L group (M

= 2.47 > M

= 1.74) though the gain difference showed no significance, t(66)

= 1.59, p > .05. The results revealed that readers’ different reading proficiency

affected vocabulary gain while reading both more and less familiar texts.

Similarly, Table 4-6 showed that in both more familiar and less familiar texts, H group achieved higher retention than L group. The retention of H group in the familiar text is higher than that of L group (M

= 2.88 > M

= 2.26) though the gain

difference showed no significance, t(66) = 1.19, p > .05; the retention of H group in the less familiar text is higher than that of L group in the less familiar text (M

= 1.74 > M

= 1.00) and noticeably the retention difference between H & L group in the less familiar text showed significance, t(66) = 2.09, p < .05.

Table 4-6 A Summary Table of Independent-Samples t-test: Reading

Comprehension, Vocabulary Gain and Retention between H and L Groups after Reading Familiar and Unfamiliar Texts

Variables Groups M SD t p H 38.24 11.41 6.96 .00**

Familiar Reading Comprehension (FC)

L 20.00 10.15

H 33.82 12.06 6.18 .00**

Unfamiliar Reading Comprehension (UC)

L 16.76 10.65

H 3.97 2.89 1.15 .25

Vocabulary Gain of Familiar Text (GF)

L 3.21 2.59

H 2.47 2.22 1.59 .12

Vocabulary Gain of Unfamiliar Text (GU)

L 1.74 1.52

H 2.88 2.38 1.19 .24

Vocabulary Retention of Familiar Text (RF)

L 2.26 1.85

H 1.74 1.78 2.09 .04*

Vocabulary Retention of Unfamiliar Text (RU)

L 1.00 1.02

Note. **p < .01, *p <.05.; n = 34 (H group); n =34 (L group)

In sum, these results showed that H group achieved higher vocabulary gain than L group in both texts, though the gain difference did not reach significance. Next, H group maintained higher retention than L group in both texts, and the retention difference reached significance while reading the less familiar text. The results

showed that readers’ different reading proficiency affected vocabulary retention while

reading more and less familiar texts. And while reading the unfamiliar text, the retention difference between H & L groups is higher than reading the more familiar text. The result revealed that L group participants performed significantly poorer in retention while reading the less familiar text. The finding corresponded with the findings of previous studies (Chern, 1993; Pulido, 2003) and also the researcher’s expected outcome.

4.1.7. Pearson Correlations Analysis on Vocabulary Size, Reading Proficiency, Vocabulary Gain and Retention of Different Topic Familiarity

Table 4-7 presented the Pearson correlations among readers’ vocabulary size, English reading proficiency, reading comprehension, and the scores of immediate and retention VKS tests in the more familiar text. Firstly, there was moderate correlation (r＝.44) between English reading proficiency and reading comprehension of the familiar text. This correlation showed that while reading familiar texts, as reading proficiency increases, so does the ability to comprehend the text. This result echoed the previous finding (Holmes,1983) of literature review that reading proficiency affects reading comprehension. Secondly, there was moderate correlation (r = .66) between reading proficiency and the scores of immediate VKS test on TWs in the familiar text. Next, there was moderate correlations (r = .68) between reading

proficiency and the scores of retention VKS test. The results revealed that as English reading proficiency increases, so does the ability to acquire and retain vocabulary in context while reading the familiar text. Furthermore, there was low correlation (r

＝.34) between reading comprehension and the scores of immediate VKS test, and

between reading comprehension and the scores of retention VKS test (r = .39) on

TWs. This displayed that reading comprehension did correlate with vocabulary gain

and retention while reading familiar texts, though the correlation was low and there

may be other factors (e.g. reader’s vocabulary size) which affect the vocabulary gain and retention more.

Table 4-7 also presented the Pearson correlations among participants’ vocabulary size, English reading proficiency, reading comprehension, and the scores of

immediate and retention VKS test in the less familiar text. Firstly, there was moderate correlation (r＝.53) between English reading proficiency and reading comprehension in the less familiar text. This correlation showed that even while reading less familiar texts, as reading proficiency increases, so does the ability to comprehend texts. The result also revealed the importance of reading proficiency in reading comprehension while reading less familiar texts. Secondly, there was also moderate significant correlation (r＝.62) between reading proficiency and the scores of immediate and retention test (r = .63) on TWs in the less familiar text. And these correlations revealed that as English reading proficiency improves, so does the ability to acquire and retention vocabulary in context even while reading less familiar texts.

Furthermore, there was the low correlation (r＝.34) between reading comprehension and the scores of immediate and retention (r = .34) VKS test on TWs. This displayed that reading comprehension did correlate vocabulary gain and retention while reading less familiar texts, as well as familiar texts, though the correlation was low.

The correlations between reading comprehension and vocabulary gain were low (r＝.34) when participants read more and less familiar topics. The correlations revealed that whether participants’ topic familiarity slightly relate to their vocabulary gain. And because the reading comprehension score (Mean＝31.06) of the more familiar topic is higher than that of the less familiar topic(Mean＝24.71, see Table 4-1), it is reasonable that participants achieve higher vocabulary gain in the more familiar text (Mean＝3.03) than in the less familiar text (Mean = 2.11, see Table 4-2).

But because the correlation between reading comprehension and vocabulary gain is

low (r = .34), it can be inferred that topic familiarity and reading comprehension cannot guarantee vocabulary gain. The correlations between vocabulary gain and other two variables, such as reading proficiency and vocabulary size, also showed their important correlations.

Table 4-7 Pearson Correlations of Vocabulary Size, Reading Proficiency, Reading Comprehension, Immediate VKS Score and Posttest VKS Score after Reading Both Texts

Familiar / Unfamiliar 1 2. 3 4. 5.

1.Vocabulary Size .67** .37** .74** .78**

2. Reading Proficiency ^.67** .44** .66** .68**

3. Familiar Reading comprehension ^.37** .53** .34** .39**

4. Immediate Test on TWs ^.74** .62** ..34** .98**

5. Posttest on TWs ^.77** ..63** ..34** .99**

**. Correlation is significant at the 0.01 level (2 tailed).

4.2. Discussions

The major issues in this study are:

1. Does readers’ topic familiarity, in planned reading, facilitate their vocabulary gain and retention?

2. Does topic familiarity, in planned reading, affect the performance of vocabulary gain and retention across different proficiency groups?

In this discussion section, the major issues would be examined by dealing with each of the following questions:

1. Does topic familiarity in planned reading have a significant effect on reading comprehension, and vocabulary gain and retention?

To answer the first question, attention will be given to:

2. Do readers have higher reading comprehension, vocabulary gain and retention in

the familiar text than in unfamiliar text?

3. Do intermediate readers have significant vocabulary gain and retention in both familiar and unfamiliar texts?

4. Do elementary readers have significant vocabulary gain and retention in both familiar and unfamiliar texts?

5. Do intermediate readers have significantly higher reading comprehension, vocabulary gain and retention in familiar texts than in unfamiliar texts?

6. Do elementary readers have significantly higher reading comprehension, vocabulary gain and retention in familiar texts than in unfamiliar texts?

7. Do intermediate readers have significantly higher reading comprehension, vocabulary gain and retention than elementary readers in both familiar and unfamiliar texts?

In order to answer to first major question, the second to seventh questions will be discussed by proper order. These questions will be investigated and discussed with participants ranging from elementary readers to intermediate learners. The discussions presented as follows revealed the effect of topic familiarity in planned reading on vocabulary gain and retention.

4.2.1. Discussion on Comprehension, Vocabulary Gain and Retention of Different Topic Familiarity

The discussions in this section begin by first giving attention to the second

question: “Do readers have higher reading comprehension, vocabulary gain and

retention in the familiar text than in unfamiliar text?” The answer is yes. The results

of Table 4-3 confirmed that readers had significantly higher reading comprehension,

vocabulary gain and retention in the more familiar text than in the less familiar text.

The finding agreed with Nunan’s (1985) study that reader-based factors such as content familiarity and background knowledge have an important effect on levels of comprehension. The finding also confirmed Diakidoky’s (1998) claim on the

significance of background knowledge in vocabulary development in reading and Pulido’s (2003) study that there was a significant effect of topic familiarity on vocabulary gain. Nevertheless, the finding did not conformed to Jacoby’s (1979) study that memory of recognition and recall was generally higher after more difficult decision on semantic judgment during text processing. An explanation for this was that when readers encounter unfamiliar topics, they may not be able to make decision in gist guessing and lexical inferencing because they fail to use the clues from context and they lack the support from their background knowledge. The finding in this study implied that all readers tend to apply background knowledge to infer the unfamiliar words while reading. Paribakht & Wesche (1999) stated that L2 readers of all levels used background knowledge to guess the meanings of unfamiliar words in texts, and readers were generally more successful in lexical inferencing when they were aware of the text topic. And Sternberg (1987) found that lexical inferencing is viewed as one of component process involved in vocabulary acquisition in reading. Although

Huckin & Bloch (1993) claimed that readers sometimes used their background knowledge improperly when they were inferring lexical meaning, the outcome of this present study revealed that background knowledge served as a strong support for lexical inferencing and vocabulary gain and retention.

4.2.2. Discussion on Comprehension, Vocabulary Gain and Retention of

Different Topic Familiarity across Different Reading Proficiency Groups

Then it was the third question: “Do intermediate readers have significant growth

in vocabulary gain and retention in both familiar and unfamiliar texts?” The answer is

yes. The results of Table 4-5 showed that intermediate readers had significant vocabulary gain and retention in both texts. Although Parry (1997) claimed that advanced learners who displayed enough background knowledge to the topic retained only low vocabulary, the finding in this study did not support Parry's finding; instead, the finding in this study conformed to the result of Chern’s (1993) study that increases in general L2 reading proficiency are also accompanied by increases in lexical

inferencing and vocabulary gain. Besides, Wu (2003) claimed that learners at the highest level of proficiency also have the best retention rate. And learners of higher proficiency level have a better vocabulary retention rate. Wu's finding as well as the finding in this study did not conform to Haastrup's (1995) and Ellis' (1989) views that topic familiarity does not necessarily favor intermediate readers. A reasonable

explanation for this finding was that for intermediate readers, they have stronger decoding skills and higher vocabulary knowledge than elementary readers. Their reading automaticity facilitates their construction and integration of ideas from context, and their use of information from long-term memory.

Next, it was the fourth question: “Do elementary readers have significant vocabulary gain and retention in both familiar and unfamiliar texts?” The answer is yes. The results of Table 4-5 showed that elementary readers had significant

vocabulary gain and retention in both texts. However, the vocabulary gain of elementary readers is much lower than intermediate readers even while reading the familiar text. This was probably because comprehension is hindered even in a familiar topic text if the text contains difficulty or unfamiliar words for elementary readers.

The finding echoed Ridgway's (1997) claim that background knowledge can overcome the weakness in readers’ vocabulary knowledge only when readers’

vocabulary knowledge level is between the upper and lower threshold. Below the

lower threshold, readers do not have sufficient L2 linguistic knowledge to infer the

meaning of the text. For less skilled readers in particular, the lack of background knowledge may diminish the chances of successful text processing, because

elementary readers may lose their allocated attention and focus during text processing, hence minimizing the possibility of vocabulary gain and retention.

Even though elementary readers, while reading the more familiar text, might activate background knowledge to make up for the weakness in reading skills and achieve vocabulary gain, the significant vocabulary gain in less familiar text was an unexpected outcome. There were two possible explanations for this. One possibility was that the text difficulty did not exceed too much over elementary readers’ reading level. Because the text difficulty was around 5-6 grade to L1 on difficulty index, elementary readers managed to gain new words while they did text proceeding.

Nagy’s (1987) study supported this result. Nagy investigated incidental learning of

word meanings from context during normal reading and found that small but reliable

vocabulary gain from the passages across all grade and ability levels (third, fifth and

seventh grades). Though seventy percent of the target words in the study occurred

only once, children at all three grade levels gained substantial knowledge about an

unfamiliar word from a single exposure. The result of Nagy’s study is similar to the

result in this study, in the way that most TWs of the two texts in this study also

occurred only once. Another explanation for the unexpected outcome of elementary

readers’ vocabulary significant gain in the less familiar text was due to the limitation

of VKS as vocabulary gain and retention measures. The VKS measures from which

the results were obtained might overestimate vocabulary gain and retention, because

elementary readers might get one-point in answering the item in the pretest “I have

never seen the word before (one point)” while answering the item in the immediate

test and posttest “I’ve seen this word, but I did not remember its meaning (two

points)”. The gain from total nonsense of the TW to the recognition of that TW is

worth 1 point (See Figure 3-1 & 3-2).

4.2.3. Discussion on Disparity of Comprehension, Vocabulary Gain and Retention of Different Topic Familiarity across Different Reading Proficiency Groups

Next the researcher discussed the effect of high and low reading proficiency on comprehension, vocabulary gain and retention. It was our fifth question: “Do

intermediate readers have significantly higher reading comprehension, vocabulary gain and retention in familiar texts than in unfamiliar texts?” The answer is positive.

The results of Table 4-3 showed that intermediate readers had significantly higher vocabulary gain and retention in familiar texts than in unfamiliar texts. This finding echoed Holmes’ (1983) claim that intermediate readers applied background

knowledge and topic familiarity much more effectively, and Chen & Donin’s (1997) view that reading comprehension is uplifted when readers possess background knowledge of the topic. These findings can be explained by Lee’s (1997) assertion that higher levels of background knowledge to the text topic and content contribute to efficiently attentional allocation during reading and avoid mess-up with unfamiliar ideas in texts, hence enabling richer textual interpretations and superior memory performance.

A similar question would be the sixth question: “Do elementary readers have

significantly higher reading comprehension, vocabulary gain and retention in familiar

texts than in unfamiliar texts?” The answer is also positive. The results of Table 4-3

showed that elementary readers had significantly higher vocabulary gain and retention

in familiar texts than in unfamiliar texts. This result can be explained by Nagy’s (1987)

findings mentioned above or by the overestimation of gain and retention caused by

VKS measure. Nevertheless, no matter which explanation was more convincing, it

was apparent that elementary readers got much lower gain and retention in the unfamiliar text than the familiar one. A possible explanation for this was that for elementary readers, because their reading skill is limited, their topic familiarity

becomes crucial in reading comprehension. This explanation advocated the facilitative effect of topic familiarity and background knowledge on vocabulary gain which many studies in the previous review supported (Adam, 1982; Arden-Close, 1993; Asher, 1980; Paribakht & Wesche, 1999; Jenkins and Dixon, 1983; Pulido, 2003).

From above, it can be drawn that for both high and elementary readers, they both had significant vocabulary gain and retention, and they had significantly higher gain and retention in the more familiar text than the unfamiliar text, and elementary readers only got little gain and retention in the unfamiliar text.

Then it came to the last question concerning the performance of different reading proficiency. Do intermediate readers have significantly higher reading comprehension, vocabulary gain and retention than elementary readers in both familiar and unfamiliar texts? The results of Table 4-6 revealed that intermediate readers did have

significantly higher reading comprehension than elementary readers of both texts.

However, they did not have significantly higher immediate vocabulary gain than elementary readers in both texts, which meant the gain disparity did not show significance in both texts. This unexpected close gain between H & L group was actually due to the limitation of VKS measure. More specifically, L group got its mean score of immediate test in the familiar text (Mean of immediate test = 18.59;

Mean of pretest = 15.38) and in the unfamiliar text (Mean of immediate test = 16.85;

Mean of pretest = 15.12). The scores were all below 20, which meant elementary

readers got only 20 points out of 50 in VKS (see Figure 3-1 & 3-2). The scores of L

group, though measured and viewed as vocabulary gain, were actually rather low

growth in vocabulary knowledge.

One noticeable finding was that although the gain disparity between H & L group in both texts did not reach significance, the retention disparity between H & L group in the less familiar text did reach significance. This implied that for elementary readers, they had much poorer retention than intermediate readers while reading unfamiliar topics. A possible explanation for this outcome was that for these

elementary readers, both their local and global comprehension is likely to be blocked while reading, therefore their scores of gain and retention were much lower. That is, when encountering unfamiliar or difficult words, though elementary readers may be able to demonstrate effective macroprocessing and comprehend the general message of a text, they will probably encounter troubles in perceiving the relationships among ideas in the text and hence be abortive at microprocessing. They experienced reading difficulty and frustration in the construction and integration of ideas from context.

This breakdown in comprehension frequently led to a short circuit in lexical

inferencing, and this weakness in comprehending the particular relationships among

ideas in a text would apparently influence vocabulary gain and retention, thus

diminishing their chances of vocabulary gain and retention in reading. A similar

finding was found in Kamil’ (1995) study that at lower levels of L2 proficiency, L1

reading strategies did not transfer to L2 reading as they did at higher levels of L2

proficiency. Likewise, Daneman (1986) claimed that lower-proficient readers often

have poorer working memory, and working memory determines how well readers can

use context to both comprehend and produce words. More specifically, when required

to comprehend an unfamiliar word, readers with small working memories were less

able to construct the meaning of that unfamiliar word from cues provided by the

context. The “working memory” theory (Jacoby et. al.,1979) proposed that readers

with small spans of working memory devote so many resources to reading processes

that they have less residual capacity for retaining the relevant contextual cues in

working memory. Even when required to access their lexical knowledge and produce a context-appropriate replacement for a familiar word, readers with smaller working memories were much slower. In short, it can be concluded that reading proficiency had a role in vocabulary gain and retention in reading. And specifically for

lower-proficient readers, McDaniel (1989) claimed that a broader and more complex factors, such text difficulty, text type and reader characteristics should be put into consideration in understanding the influence of text difficulty on lower-proficient reader’s memory of recall.

In the literature review, it has been discussed that reader-based factors, such as readers’ vocabulary size, reading proficiency and background knowledge, these factors might affect reading comprehension and vocabulary acquisition in reading.

And the previous review generally agreed that background knowledge enhanced macroprocessing, while vocabulary knowledge enhanced microprocessing, the researcher additionally explored what factors affect reading comprehension and vocabulary acquisition. Stahl & Jacobson (1986) claimed that vocabulary difficulty affected recall of microstructure, whereas background knowledge affected the macrostructure. The researcher gave the questions: “Did readers’ vocabulary size influences their vocabulary gain more than their comprehension? Did readers’ reading proficiency affect their reading comprehension more than their vocabulary gain and retention?” The Pearson Correlation Analysis was proceeded to examine the

relationships among vocabulary size, reading proficiency, reading comprehension,

vocabulary gain and retention in both familiar and unfamiliar texts.

4.2.4. Discussion on Pearson Correlations of Vocabulary Size, Reading

Proficiency, Reading Comprehension, Vocabulary Gain and Retention of Different Topic Familiarity

Previous review (Chern, 1993; Holmes, 1983; Pulido, 2003) seemed to support the outcome of the Pearson Correlation Analysis in this study. The results of Pearson Correlation Analysis were listed as follows:

1. There were moderate correlations (r = .44, p < .01; r = .53, p < .01) between English reading proficiency and reading comprehension with both familiar and unfamiliar texts.

2. There were moderate correlations (r = .66, p < .01; r = .62, p < .01) between reading proficiency and vocabulary gain with both familiar and unfamiliar texts.

3. There were moderate correlations (r = .68, p < .01; r = .64, p < .01) between reading proficiency and vocabulary retention with both familiar and unfamiliar texts.

These findings corresponded with Holmes (1983) study that higher-proficient readers applied background knowledge and topic familiarity more effectively than lower-proficient readers. The finding also echoed the other two studies that increases in general L2 reading proficiency are also accompanied by increases in lexical inferencing (Chern, 1993) and vocabulary gain (Pulido, 2003). In this study, English reading proficiency correlated with English vocabulary gain and retention for

generally all participants, including elementary readers and intermediate readers.

4. There were only low correlations (r = .34, p < .01; r = .34, p < .01) between

reading comprehension and vocabulary gain in both familiar and unfamiliar text.

5. There were only low correlations (r = .39, p < .01; r = .34, p < .01) between reading comprehension and vocabulary retention in both familiar and unfamiliar texts.

The outcome revealed that reading comprehension only lowly correlated with vocabulary gain and retention. A possible explanation was that reading

comprehension was based on macroprocessing, while vocabulary gain and retention might rely on more than microprocessing. This led to the discussion of another more microprocessing-based factor, readers’ vocabulary knowledge, and its correlation with other four variables, which were reading proficiency, reading comprehension,

vocabulary gain and retention.

6. While reading the more familiar text, readers’ vocabulary size has the high correlation with vocabulary gain and retention (r = .74, p < .01; r = .77, p < .01;

see Table 4-7). And readers’ reading proficiency has moderate correlation with gain and retention. (r = .62 , p < .01 r = .63, p < .01; see Table 4-7).

7. Similarly, while reading the less familiar text, readers’ vocabulary size has the high correlation with vocabulary gain and retention (r = .74, p < .01; r = .78, p

< .01; see Table 4-7). And readers’ reading proficiency has moderate correlation with gain and retention. (r = .66 , p < .01 r = .68, p < .01; see Table 4-7).

This outcome was supported with the findings of other studies (Haynes, 1993; Na

& Nation, 1985; Stahl et al., 1989) in the review. Stahl et al. (1989) claimed in their

study that background knowledge and vocabulary difficulty both appear to affect

recall from text, but these effects are independent, not interactive. This outcome has

strongly indicated that readers do not use background knowledge to compensate for

vocabulary difficulty.

In short, the outcome showed that vocabulary size has higher with correlation with vocabulary gain, than background knowledge. And the outcome implied that vocabulary knowledge helps to overcome vocabulary difficulty and to achieve

vocabulary gain and retention. This outcome was supported by other studies (Haynes, 1993; Na & Nation, 1985). Haynes (1993) stated that readers with higher vocabulary knowledge caught the vocabulary in the context and inferred the target words more successfully, while readers with poor vocabulary knowledge were trapped and experienced a short circuit in lexical inferencing, because they did not have enough knowledge of the vocabulary in the context. Na & Nation (1985) supported that vocabulary knowledge is one crucial determinant of vocabulary gain in reading. This is because that when encountering unfamiliar or unfamiliar words, learners may be able to demonstrate effective macroprocessing and comprehend the gist of a text, but they will encounter troubles in perceiving the relationships among ideas in the text and hence be abortive at microprocessing, which diminishes vocabulary gain. This weakness in comprehending the particular relationships among ideas in a text would apparently influence vocabulary gain in reading.

However, the effect of vocabulary size on vocabulary gain and retention did not imply that vocabulary size was not important for reading comprehension (r = .37; r

= .37, p < .01). Laufer (1997) claimed that a 3000-word level would predict 56% of a reading comprehension score and an increase in 1,000 word families would cause a 7% increase in a reading comprehension score. Similarly, Coady (1993) stated that vocabulary size is one of the decisive indicators of reading success. Furthermore, in Huang’s (2000) study, Taiwanese students’ vocabulary knowledge accounted for 68%

of their reading comprehension scores on the English text. Such an outcome explained

that vocabulary knowledge also played one of a critical role in English readers’

reading comprehension. And it can be drawn from this point that topic familiarity and

background knowledge can not guarantee a successful reading comprehension if the

text contains too many difficulty words for the reader. Vocabulary knowledge has

high correlation with vocabulary gain, whereas reading proficiency has moderate

correlation for reading comprehension. In addition, vocabulary knowledge also has a

role in reading comprehension.