I have a purpose in mind when I read. (2.73)

38. I react emotionally to information in the text. (2.58)

22. I read aloud to help me understand when texts become difficult. (2.48)

Note: The mean scores were shown in parentheses. Strategies that were found in both languages were in italics.

Second, although the rest of the two frequently used strategies in L1, to

summarize in one’s own words and to read slowly and carefully, were not found in the top-five list of L2, they were indeed the two strategies immediately following the top five list (see Table 8). Similarly, the rest of the two top-five strategies in L2, to use

background knowledge to guess the meanings of unknown words and to underline or circle information, were also found to be of rather frequent use in L1 (i.e. both items had mean scores higher than the overall mean of 3.37; see Table 7) despite that they were not among the top-five strategies presented in Table 9. These results again indicated that students preferred the use of certain reading strategies no matter whether they read in L1 or L2.

Regarding the bottom five strategies, the results turned out quite differently between the two languages. First, as shown in Table 9, the two lists had only one strategy in common, reading aloud to increase understanding. This particular strategy was found to be least often used in both languages. Second, while strategies like using word-parts, using dictionaries, and taking notes were found to be among the

bottom-five list in L1, they were found to be of rather frequent use in L2 (i.e. they had mean scores higher than the overall mean of 3.10; see Table 8). Since these three strategies appeared to have a focus on processing texts at a more local level, their different distributions in L1 and L2 seemed to indicate the processing differences between languages. Third, while looking into the remaining strategy left in the bottom-five list of L1, it was found that the strategy, to try to get back on track when losing concentration, was also not very frequently used by students when reading in L2 (with a mean score much lower than the overall mean of 3.10, see also Table 8), suggesting its low popularity in both languages. Finally, when examining the bottom-five strategies in L2, in addition to the aforementioned strategy

(reading-aloud), which was least often applied in both languages, the rest of the strategies seemed to share a focus on the metacognitive aspect of text processing.

They included: checking whether text content fits purpose, critically analyzing and evaluating information, having a purpose for reading, and reacting emotionally to information in the text. This result appeared to indicate that students tended to engage

less in metacognitive reading processes when reading in L2.

Paired-Sample T Test

As in the analysis of the two beliefs questionnaires, in order to compare students’ reported use of strategies within and across languages, a series of

paired-samples t tests were also performed on the data collected from the two RSQs.

Table 10 shows the mean scores and standard deviations for the two strategy scales and each of the two subscales

Table 10.

Mean scores and standard deviations for strategy use in L1 and L2

LI ( N = 408) L2( N = 419) Overall^a G^b Non-G^c Overall G Non-G

Mean 3.37 3.46 3.14 3.10 2.87 3.27

Std. Deviation .55 .58 .67 .63 .67 .66 Note: ^a Overall = Overall strategy use

^bG = General Reading Strategy Factor

^cNon-G = Non-General Reading Strategy Factor

As shown in Table 10, concerning the two strategy subscales within each language, there appeared to be observed differences between the mean scores of the two subscales in L1 (General = 3.46; Non-General = 3.14) and in L2 (General = 2.87;

Non-General = 3.27). As for the use of strategies across the two languages, several differences were found. First, regarding the overall use of reading strategies between the two scales, it was found that students appeared to reportedly use more reading strategies in L1 than in L2 (L1 = 3.37; L2 = 3.10). Second, there was also a difference between the means of the two general reading strategy subscales, with L1 being higher than L2 (L1 = 3.46; L2 = 2.87). Finally, when comparing the mean scores of the two non-general reading strategy subscales, the difference, however, appeared to be in the opposite direction, with L2 being higher than L1 (L1= 3.14; L2 = 3.27).

These observed differences, as revealed in Table 10, were all confirmed by the

results obtained from the paired-samples t tests. Table 11 shows the results of the subsequent paired-samples t tests.

Table 11.

Summary of t tests for the reported strategy use within and across languages

Paired-samples t tests Df t p

L1 General vs. L1 Non-General L2 General vs. L2 Non-General

L1 Overall vs. L2 Overall L1 General vs. L2 General L1 Non-General vs. L2 Non-General

Note: **Difference is significant at the 0.01 level (2-tailed).

Regarding the two strategy subscales within each language, students scored significantly higher on the general strategy subscale when reading in L1 (df = 407, t = 12.60, p = .00). However, when it comes to reading in L2, the opposite direction was found. They tended to score significantly higher on the non-general reading strategy subscale (df = 418, t = -19.57, p = .00). As for the differences between L1 and L2, first, it was found that students reportedly used a significantly higher amount of reading strategies in L1 than in L2 (df = 399, t = 8.68, p = .00). Second, concerning the two general strategy subscales between languages, students scored significantly higher in L1 than in L2 (df = 403, t = 17.11, p = .00). On the other hand, regarding the two non-general strategy subscales between languages, the score was found to be significantly higher in L2 than in L1.

The Relationship Between Readers’ Beliefs and Reported Use of Reading Strategies in L1 and L2 This section presents findings related to the investigation of relationships between student subjects’ beliefs about reading and their reported use of reading strategies in L1 and L2. As described in the previous chapter, two levels of analysis were conducted on the responses collected from the belief and strategy questionnaires

across the two languages. First, the study examined the correlations among the means of belief and strategy categories identified within each language. Data collected from the whole sample group (N = 432) were analyzed using Pearson correlations.

Second, the study examined the strategy patterns reported by students classified as having distinctive beliefs about reading in L1 and L2. Based on their responses to the finalized RBQ, a total of 68 students, out of the 432 sample group, were selected using the criterion described in Chapter 3. The 68 students were reportedly having high transaction (TA) or high transmission (TM) belief when reading across the two languages. More specifically, they represented the four types of readers with different belief profiles: Type A (L1TA-L2TA), Type B (L1TA-L2TM), Type C (L1TM-L2TM), and Type D (L1TM-L2TA). Note that such classification, as described earlier, was intended to contrast two types of reading beliefs, transaction versus transmission. The two subcategories of transaction belief, TA I- Reader Interpretation and TA II Reader Engagement, were thus combined under the category of TA during the selection. To compare the different uses of strategies reported by these four groups of readers, one-way ANOVA was performed.

Pearson Correlations The Chinese RBQ and RSQ

Students’ responses to the finalized Chinese RBQ and RSQ were analyzed to find out the relationship between learners’ beliefs about reading and their reported use of reading strategies in L1. Pearson correlations were performed to examine the correlations among the mean scores of three belief subscales (TA I – Reader

Interpretation, TA II – Reader Engagement, and TM), two strategy subscales (General and Non-General), and the overall strategy scale. The results of Pearson correlations in L1 can be found in Table 12.

Table 12.

Summary of Pearson correlations between belief and reported strategy types in L1

S-G^d S-Non-G^e S-Overall Note: **Correlation is significant at the 0.01 level (2-tailed).

a TA I = Transaction Belief I – Reader Interpretation

b TA II = Transaction Belief II – Reader Engagement

C TM = Transmission Belief

d S-G = General Reading Strategy Factor

e S-Non-G = Non-General Reading Strategy Factor

As the table shows, generally speaking, all three types of beliefs, TA I, TA II, and TM, were found to be significantly correlated with the overall strategy scales as well as the two strategy subscales, with correlation scores ranging from .30 to .67 (p<.01). These results indicated that when reading in L1, students’ beliefs about reading were positively correlated with their reported use of reading strategies. In addition, when we examined the correlation scores in Table 12, it was found that the highest correlation score was found between the Transaction II – Reader Engagement subscale and General Reading Strategy subscale (TA II - General = .67) and the lowest between the TA II – Reader Engagement subscale and Non-General Reading Strategy subscale (TA-Local = .30), suggesting a possibly stronger connection

between the beliefs about the reader engagement (i.e. a subtype of TA) and the use of general reading strategies.

The English RBQ and RSQ

In a similar vein, students’ responses to the English RBQ and RSQ were also

statistically analyzed using the Pearson correlations. Table 13 shows the results of the correlations among the belief types and strategy categories in L2.

Table 13.

Summary of Pearson correlations between belief and reported strategy types in L2