國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
41
CHAPTER 4 Results and Discussions
This chapter presents the results of data analysis on test performances and questionnaire responses. The first section focuses on statistical comparison and discussion of the participant’s performances on the careful reading test and the expeditious reading test. It also presents the comparison and discussion of their performances on the three major expeditious reading constructs. In addition, the reading performances by different proficiency groups are presented and discussed.
The second section proposes and discusses the frequency of reading strategies used in the expeditious reading test. The strategies used by different proficiency groups are also presented. The last section focuses on the reading difficulties in the expeditious reading test. The difficulties experienced by different proficiency groups are also discussed.
Reading Performance
This section first presents and discusses the result of the statistical comparison between the participant’s careful reading performance and expeditious reading performance. In addition, the comparison among their performances on the three major expeditious reading constructs is proposed. Last, this section presents the comparison of reading performances by different proficiency groups.
Performances on Careful Reading Test and Expeditious Reading Test
To compare the students’ careful reading and expeditious reading performances, the researcher first calculated the scores that the 45 participants obtained in the expeditious reading test and the careful reading test. Both tests contained 15 items.
Each item counted for one point, so the full mark of each test was 15. The preliminary descriptive statistics showed that the average of the scores that the participants got on the careful reading test was 11. 31, and the average on the expeditious reading test
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
42
was 11. 96. The student’s performance on the expeditious reading test was slightly higher than that on the careful reading test.
To know if the participants’ expeditious reading performance was significantly higher than the careful reading performance, the researcher used SPSS, a statistical software, to conduct a two-way ANOVA (repeated measure with one within-subject independent variable and one between-subject independent variable10) on the scores of the two reading tests. The result showed that there was a marginally significant difference between the students’ expeditious reading performance and the careful reading performance (F(1, 51) = 3.241, p = .078), and a significant difference among the performances by students of different proficiency levels (f(2, 51) = 12.777, p
< .01). Besides, there was no interaction effect between the two independent variables (F(2, 51) = .950, p = . 394). It was unexpected that the students’ expeditious reading performance was marginally higher than their careful reading performance, because the expeditious reading was less instructed and practiced in the senior high school English classes.
To provide a possible explanation, the researcher inspected if genre types had an effect on learners’ test performance, as Shohamy (1984) and Paul et al. (2001)
suggested. The texts in the careful reading were two expositions and two descriptions, but those in the expeditious reading test included an exposition, a description, a manual, and a book notice. The manual and book notice were used in the expeditious reading test because readers frequently perform skimming, search reading, and scanning on these materials in real life (Spache & Berg, 1984; Urquhart & Weir, 1998). The genre types differed significantly from the exposition and description in terms of the text structure. The difference can lead to significant variance in test
10 The within-subject independent variable referred to the type of reading test, and the between-subject independent variable was the language proficiency.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
43
scores.
To examine the effect of genre types on reading performance, the researcher divided the four texts in the expeditious reading test into two categories. The first two texts (manual and book notice) formed one and the last two (exposition and
description) the other. The participants’ scores on these two categories of expeditious reading texts were compared with the scores on two careful reading texts11
(exposition and description). The results of two-way ANOVA (repeated measure with one within-subject independent variable and one between-subject independent
variable12) showed that the participant’s performance on the manual and book notice in the expeditious reading test was significantly higher than their performance on the exposition and description in the careful reading test (f(1, 50) = 10.790, p < .01), and there was a significant difference among performances by students of different proficiency level (f(2, 50) = 19.694, p < .01). Besides, there was no interaction effect between two independent variables (f(2, 50) = .240, p = .915). In contrast, the
participant’s performance on the exposition and description in the expeditious reading test did not significantly differ from their performance on the exposition and
description in the careful reading test (f(1, 50) = .235, p = .949). In short, the reason why the students’ expeditious reading performance was marginally higher than their careful reading performance might be the good performances on the first two expeditious reading texts: the manual and book notice.
The students can quickly extract the desired information in the manual and book notice because the messages were well-categorized into distinct sections, and each section was introduced with a subtitle in boldface (De Leeuw & de Leeuw, 1965).
11 The selected careful reading passages were the first and the last passage in the careful reading test because these two passages’ items encompassed all six constructs of careful reading and thus may be representative of the careful reading test performance.
12 The within-subject independent variable referred to the type of text, and the between-subject independent variable was the language proficiency.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
44
These features were pointed out by 25 students in the questionnaire survey as the reason why the items in the manual and book notice were easy to attain (see APPEDNIX G, Quote 1). On the other hand, the exposition and description in the expeditious reading test were composed of continuous paragraphs. Although the messages addressed in each paragraph were introduced by the topic sentence, the questionnaire survey showed that the students were not frequent to utilize the topic sentences to locate the desired information. They, instead, went over the paragraphs one by one to find the answers in the exposition and description. This was similar to the way they approached the exposition and description in the careful reading test.
Therefore, the students’ reading performance on the exposition and description in the two reading tests were not significantly different.
Aside from genre types, items also had a profound effect on the test performance (Anderson & Bachman, 200). The participants’ test scores on the expeditious reading test were higher because its items required less effort or scope of reading
comprehension and thus were easier to address. Twelve students mentioned in the questionnaire survey that the item of skimming for gist can be achieved through reading only the subtitles of a text (see APPEDNIX G, Quote 2). In addition, 30 students expressed that the items of search reading and scanning can be quickly attained through matching the same or similar words in the stems with those in the texts (see APPEDNIX G, Quote 3). On the other hand, the items in the careful reading test required greater effort or scope of reading comprehension. The item of identifying the relationship between a main idea and its details can be done only by intensive reading on an entire paragraph (Urquhart & Weir, 1998). Additionally, the item of guessing the meaning of unknown words demanded readers of adequate
comprehension on 95 percent of the words in the context (Laufer, 1989; Liu & Nation, 1985). It was thus more challenging for the participants to reach better performance
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
45
on the careful reading items.
Performances on Three Major Expeditious Reading Constructs
To compare the participants’ performance on the three major expeditious reading constructs, the researcher first calculated the scores they obtained on the items of skimming, search reading, and scanning. Since there were 6 items for skimming, 5 for search reading, and 4 for scanning, the full marks on the three constructs were 6, 5, and 4 respectively. The participants’ raw full marks on skimming, search reading, and scanning were transformed onto an identical scale with a full mark of 100 for each construct. The researcher calculated the averages of the transformed scores that the participants got on each construct. The preliminary descriptive statistics showed that the participants’ scanning performance was the highest (with the average of 92.59), followed by search reading performance (84.44). The performance on skimming was the lowest (68.83).
Then the researcher used SPSS to conduct a two-way ANOVA (repeated measure with one within-subject independent variable and one between-subject independent variable13) on the transformed scores that each participant achieved on skimming, search reading, and scanning. The result showed that there was a significant
difference among the performances of the three expeditious reading constructs (F(2, 102) = 26.539, p < .01), and a significant difference among the performances by students of different proficient levels (F(2, 51) = 6.003, p < .01). Besides, there was no interaction effect between the two independent variables (F(4, 104) = .240, p
= .915). A further post hoc multiple comparison test testified that the participants’
scanning performance was significantly higher than the search reading and skimming performances (p < .01), and their search reading performance was also significantly
13 The within-subject independent variable referred to the type of construct, and the between-subject independent variable was the language proficiency.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
46
higher than the skimming performance (p < .01).
The finding can be explained with the difference in construct complexity. The performance on scanning was the highest because scanning was relatively simple. The answer to a scanning question was found in a single sentence (Davey & Lasasso, 1984) and can be located with mere matching of words or numbers, as 24 students proposed in the questionnaire responses (see APPEDNIX G, Quote 4). On the other hand, the answer to a search reading question sometimes covered more than one sentence in the text. Besides, locating the answer required extra effort of identifying synonymous words or phrases (Bensoussan et al., 1984). The performance on skimming was the lowest because skimming was the most complex construct. The main ideas needed to answer a skimming question had to be located with the use of formal knowledge and then were combined across sentences to grasp the gist or macrostructure of a text (Davey & Lasasso, 1984; Urquhart & Weir, 1998). This factor was pointed out by 11 participants in the questionnaire responses as the reason why the items of skimming were especially difficult (see APPEDNIX G, Quote 5). Moreover, Alderson and Bachman (2000) proposed that test-takers’ attention was mostly on understanding facts and details when required to read in an extrinsically motived test-taking context.
Thus the skimming items, which assessed the comprehension of main ideas, were performed less satisfactorily by the students.
The familiarity of the item type was another factor that caused the skimming performance to be lower. Four out of six skimming items were presented in the item type of multiple matching. According to the questionnaire responses by 10
participants, the item type caused extra difficulty for skimming because they were seldom trained to practice it and found themselves less capable of finding textual and contextual clues to match topic sentences with correspondent paragraphs (see
APPEDNIX G, Quote 6).
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
47
In addition, the genre type can be a decisive factor that led to the higher search reading and scanning performance. Most items of search reading and scanning were contained in the manual and book notice. The previous statistical analysis had shown that the participants’ reading performance on those two genres was significantly higher. It was because their clear layout or distinct sections were useful aids in extracting information, as acknowledged by the 25 students in the questionnaire responses (see APPEDNIX G, Quote 1).
Last, the strategy use can affect the students’ performance on the three
expeditious reading contracts (Alderson, 1984). According to the questionnaire survey, the students were comparatively more frequent to employ the strategies related to attaining search reading and scanning tasks, such as matching the same or similar words in the questions with those in the texts. Thus they performed better on search reading and scanning. On the other hand, the strategies that aid with achieving skimming tasks like reading the topic and concluding sentences were less frequently adopted. Hence, the students’ skimming performance was lower.
Performances by Different Proficiency Groups
To compare the expeditious reading performance by students of different proficiency levels, the researcher first calculated the average of expeditious reading scores in each proficiency group. The preliminary descriptive statistics showed that the average of the high proficiency groups’ expeditious reading performance (13.211) was higher that of the mid proficiency group (11.913). The average of the mid
proficiency group’s expeditious reading performance was also higher than that of the low proficiency group (10.75).
In order to know if there was a significant difference among the three proficiency groups’ expeditious reading performances, SPSS was used to conduct a two-way ANOVA (repeated measure with one within-subject independent variable and one
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
48
between-subject independent variable14) on the participants’ test scores. The result showed that there was a significant difference among the three proficiency groups’
reading performances (f(2, 51) = 12.777, p < .01), and a marginally significant
difference between the performances of two reading types (F(1, 51) = 3.241, p = .078).
Besides, there was no interaction effect between the two independent variables (f(2, 51) = .950, p = .394). A further post hoc multiple comparison test verified that the high proficiency group’s expeditious reading performance was significantly higher than the performances by the mid proficiency group (p <.05) and the low proficiency group (p < .01). The mid proficiency group’s expeditious reading performance was also significantly higher than the performance by the low proficiency group (p < .01).
As for the comparison of the three proficiency groups’ performances on the three major expeditious reading constructs, the researcher first calculated the averages of the transformed scores on skimming, search reading, and scanning in each proficiency group. The averages of each proficiency group’s performances on the three constructs were presented in the following Table 4. The preliminary descriptive statistics showed that the high proficiency groups’ performance on skimming, search reading, and scanning were higher than those of the mid and the low proficiency groups. The mid proficiency group’s performances on the three constructs were also higher than those of the low proficiency group.
Table 4
Different Proficiency Groups’ Averages on Expeditious Reading Constructs
Construct Group High Mid Low
Skimming 73.68 68.12 62.50
Search Reading 92.63 83.48 73.33
Scanning 98.68 92.39 83.33
14 The within-subject independent variable referred to the type of reading test, and the between-subject independent variable was the language proficiency.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
49
To know if there existed a significant difference among the three proficiency groups’ performances on the three major expeditious reading constructs, the researcher used SPSS to conduct a two-way ANOVA (repeated measure with one within-subject independent variable and one between-subject independent variable15) on the transformed scores that each participant achieved on skimming, search reading, and scanning. The result showed that there was a significant difference among the three proficiency groups’ reading performances (F(2, 51) = 6.003, p < .01), and a significant difference among performances of the three expeditious reading constructs (F(2, 102) = 26.539, p < .01). Besides, there was no interaction effect between the two independent variables (F(4, 102) = .240, p = .915). A further post hoc multiple
comparison test proved that the high proficiency groups performances on skimming, search reading, and scanning were marginally higher than those of the mid
proficiency groups (p = .066) and significantly higher than those of the low
proficiency group (p > .01). The mid proficiency groups performances on skimming, search reading, and scanning were marginally higher than those of the low proficiency groups (p = .059).
The above findings of the comparison of reading performances by different proficiency groups were not surprising because the test performance was profoundly influenced by test-takers’ linguistic knowledge (Alderson & Bachman, 2000). The knowledge of vocabulary was especially critical because it is a major indicator of the reading performance (Chall, 1958). According to the questionnaire survey, the higher proficiency level students had less difficulty with understanding the words used in the expeditious reading texts. Thus their overall expeditious reading performance and the performance on skimming, search reading, and scanning were higher than those by
15 The within-subject independent variable referred to the type of construct, and the between-subject independent variable was the language proficiency.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
50
lower proficiency level students.
In addition, the test performance was influenced by test-takers’ strategic knowledge (Block, 1992). The questionnaire survey showed that the high achievers employed expeditious reading strategies more frequently than the intermediate level students. The low achievers had the least frequency of use on expeditious reading strategies. Since strategy use was crucial for achieving reading tasks (Chiu, 1998; Lee, 2003), more frequent use of the expeditious reading strategies by higher proficiency level students resulted in higher expeditious reading performance.
Reading Strategy
This section presents and discusses the analysis on the questionnaire responses in the reading strategy section. The frequency of reading strategies the participants adopted is presented to find out their strategy use in the expeditious reading test and what expeditious reading strategies were more or less frequently applied. The
student’s written responses to the open-ended item are also added to the discussion. In addition, the frequency of reading strategies in each proficiency group is proposed and discussed to see if students of different proficiency levels preferred to employ different strategies or had different rates of strategy use.
Strategy Use in Expeditious Reading Test
To know the participants’ strategy use in the expeditious reading test, the researcher calculated the average frequency of each reading strategy. The frequency scale in the questionnaire had five levels: nearly never, seldom, sometimes, usually, and always. If a participant checked “nearly never” on the frequency scale of a strategy, that strategy was given 1 point. The level “seldom” counted for 2 points,
“sometimes” for 3, “usually” for 4, and “always” for 5 points. The total points of a strategy were summed up based on all the participants’ responses and then were divided by the total number of the participants (54) to generate the average frequency
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
51
of a reading strategy. The higher the average, the more frequently a strategy was applied in the expeditious reading test. Table 5 ranks the fifteen reading strategies in the questionnaire according to their average frequency. The word in the parentheses following each strategy item indicates the category of the strategy. “(E)” is for expeditious reading strategies and “(C)” for careful reading strategies. There are 11 expeditious reading strategies, and 4 careful reading strategies.
‧
Average Frequency of Reading Strategies in the Expeditious Reading Test
Rank Strategy Frequency
1 I used typographical aids like boldface to identify key
information. (E) 4.35
1 I carefully read the sentences where the target information
was located. (E) 4.35
3 I searched for the same or similar words in the text that match the words in the question. (E)
4.24
4 I read questions first to set goals for finding the gist, a term,
or a number. (E) 4.04
5 I glanced over the text or some paragraphs. (E) 4
6 I read the title or subtitles first. (E) 3.82
7 I noted the genre, length, and theme of a text first. (E) 3.46 7 I tried to figure out the meaning of each unknown word from
the clues of the context. (C) 3.46
9 I read the first sentence of a paragraph to decide if the
desired information is in that paragraph. (E) 3.22 10 I read only the parts that contain the information relevant to
the questions (E). 3.19
10 I went back and forth between the lines to make inference
from details. (C) 3.19
12 I read the first paragraph and the last paragraph to grasp the
main ideas of a text. (E) 3.15
13 I read the first sentence and the last sentence of a paragraph
to grasp its main idea. (E) 2.96
14 I read the entire text carefully. (C) 2.91
15 I tried to understand each sentence by analyzing its
grammatical structure. (C) 2.48
Overall 3.52
11 Expeditious Reading Strategies 3.71
4 Careful Reading Strategies 3.01
The result showed that the overall average frequency of fifteen reading strategies
The result showed that the overall average frequency of fifteen reading strategies