The factors that influence Chinese students’ English reading comprehension.

Research finding and results
There are some research findings and results will be presented in this chapter. This is the most important part of the dissertation.
5.1 Experiments 1 ：Prior knowledge
5.1.1 result of experiments
It was expected that prior knowledge would have a profound effect on science course performance as well as comprehension of a science text. Those students who have prior knowledge of the domain of a text should perform better on comprehension measures of that material because they have the opportunity to incorporate that prior knowledge with the text material.
Two types of instruments were used in this study. The first type included Culture text and reading comprehension tests that the researcher constructed. Culture text was chosen based on a pilot study with a representative sample of participants similar to the research participants described above. A data analysis confirmed that the test was appropriate for determining the students’ general English proficiency levels. The second type of instrument included reading comprehension tests based on computer science texts. Both the two texts were similar as far as genre, length, and linguistic difficulty are concerned. They were both extracts from the biographies of religious characters.
Thus, all the text variables except the content were controlled. Each of these tests appeared in four groups: Culture know more group, Computer know more group, Both culture and computer know more group and both culture and computer know less group.
1. Culture, Culture know more group (CU.CUM)
2. Culture, Computer know more group (CU.COM)
3. Culture, Both culture and computer know more group (CU.CUOM)
4. Culture, Both culture and computer know less group (CU.CUOL)
5. Computer, Culture know more group (CU.CUM)
6. Computer, Computer know more group (CU.COM)
7. Computer, Both culture and computer know more group (CU.CUOM)
8. Computer, Both culture and computer know less group (CU.CUOL)
For each of the texts, 14 Multiple Choice test items were developed. The tests were piloted and pre-tested with a sample group of participants.
Note that the researcher’s purpose is to investigate the effect of prior knowledge on reading comprehension. Hence, the items with high or low facility indices were not discarded, except when an item was realized to be too easy due to the directness of its answer. When the data analysis and item modification were accomplished, 14 MC items were chosen for each of the texts.
The next step was syntactic simplification. The sentences considered difficult by the participants were broken down into simple sentences. For the sake of consistency, we stuck to Richards, Platt, and Platt’s (1992) definition of a simple sentence, “a sentence which contains only one predicate” (p. 70). The sentences the learners identified as difficult were some compound (sentences with more than one predicate) and complex sentences (especially sentences with adjective clauses).
To avoid participant fatigue, the content-familiar and the content-unfamiliar tests were administered in two separate sessions with an interval of four days. The participants were asked to read the texts and answer the MC items. The time allocated to each test was 45 minutes, which was determined to be suitable in the pilot study.
Results
The data consisted of the participants’ answers to the Multiple Choice (MC) comprehension questions and the analysis of the written protocols. The descriptive statistics in Tables 5.1 provides an overall view of the participants’ performance on the reading comprehension test.

Text version and different group
N
Minimum
Maximum
M
SD
CU.CUM
6
3.00
10.00
7.44
2.05
CU.COM
6
3.00
10.00
7.00
1.84
CU.CUOM
6
3.00
10.00
7.21
1.76
CU.CUOL
6
2.00
9.00
6.88
1.99
CO.CUM
6
3.00
8.00
6.39
1.71
CO.COM
6
3.00
9.00
6.76
1.81
CO.CUOM
6
3.00
9.00
6.41
1.52
CO.CUOL
6
2.00
8.00
6.14
1.82
Table 5.1. Descriptive statistics of MC test scores

















As for the interactions related to the MC reading comprehension tests, Apparently, according to From the preceding tables the researcher noticed that the average for culture know more group in reading culture text and computer know more group in reading computer are tremendous better than the other groups, and it the group have little prior knowledge both on culture and computer, their performance are quite bourgeois. These results indicate that prior knowledge is very important on reading comprehension.
Discussion
The researcher found prior knowledge to have a greater effect than lexical or syntactic simplification on EFL reading comprehension. This is compatible with the results of studies by Carrell (1987), Johnson (1982), Kang (1992), Oh (2001), Steffensen and Joag-Dev (1984), and Steffensen, Joag-Dev, and Anderson (1979). Of course, these studies used different kinds of texts.
As stated above, the researcher found that the content and linguistic simplification interacted. The lexical simplification had a significantly facilitative effect on reading comprehension of the content-unfamiliar texts, but it had an impeding effect on the participants’ comprehension of the content-familiar text, although the effect was not significant. This may indicate that when the content is familiar, readers can guess the meanings of unknown words, and linguistic simplification does not improve the readers’ comprehension. However, when the content is unfamiliar, readers cannot guess the meanings of new words. Hence, new lexical items will add to the readers’ problems, and lexical simplification will enhance their comprehension.
Note that the impeding effect of the lexical simplification on the participants’ comprehension of the content-familiar texts was not significant. As mentioned above, in the process of the lexical simplification, some unknown words did not lend themselves to being substituted with synonyms. In such cases, the definitions of the words were given next to them in parentheses. The hindering effect of the lexical simplification on the comprehension of the content-familiar text may be attributed to the parenthetical definitions of the words. That is to say that these definitions distracted the participants and decreased their comprehension. It is difficult to explain why the same procedure of lexical simplification did not impede the comprehension of the content-unfamiliar text but facilitated it significantly. One possible interpretation is that the readers may have used a conceptual rather than a syntactic strategy requiring more vocabulary, content, and world knowledge.
3.2 Experiments 2： Inference strategy
3.2.1 procedure and materials
In experiment 2, the materials used for the data collections were two articles. Both of them are from EFL textbooks. One article dealing with unfamiliar topic comes from a specific subject area that is not the specialist area of the readers: computer. Another article that deals with familiar topic is a description of the Silk Road in China. The reason for choosing this article is that the Silk Road is a super important and famous historical trade route and its influences on the culture of China were significant. Furthermore, all Chinese students have to learn the history and geography of China in high school, so I think they would be familiar with it. The texts were similar in length (Text A, 451 words; Text B, 392 words).
A set of comprehension and inference questions were also adopted and developed for each article. The questions were based upon information in the articles and cannot be answered correctly without having reading and understood the articles.
Questions for The Silk Road
1. Where is the Silk Road?
2. Why the locals have a very great respect for this ‘Land of Death ‘?
3. Where does the oasis exist in the Taklimakan desert?
4. What kind of the land surface may produce dangerous sandstorms?
5. Which desert has better natural environment? Taklimakan or Gobi? Why?
6. What does ‘most difficult ‘ (line 35) refer to?
7. Why the route along the ‘Gansu corridor’ is the least difficult entry into this region from east?
Questions for Hollerith and the Punched-Card Revolution
1. What does ‘unlikely quarters’ (line 1) refer to?
2. What inspired Babbage’s analytical engine?
3. What do you think the sentence ‘the Babbage/Jacquard idea… …’ (line 7) means?
4. Why did the Census Bureau foster the creation of new counting machines?
5. Which of the new proposals was accepted? Why?
6. Can we consider Hollerith’s model more advanced than Babbage’s? If so, in what ways?
7. What is the relationship between Hollerith and IBM?
3.2.2 research finding
These inference questions associated with generating inferences from the texts. The nature of the inference depends on the question. The inferences could be bridging, explanatory, predictive, or elaborative. Some of these questions may assess inferences that are considered to be normally generated while reading the texts. As such, the comprehension product associated with these questions corresponds to the situation model that proficient readers are considered to form in the normal course of reading. The first part inference questions above were classified as such because it involves inferring the affective response of characters in a short narrative. Other inference questions may require readers to reason beyond the texts and generate inferences that are specific to the content of the questions. The second part inference questions are examples of this type of question because they require the reader to reason beyond the text. It is important to note that not all inference questions in these tests reflect inferences that readers would normally construct when reading the texts. Inference questions that tap situation model inferences would need to be carefully constructed in light of discourse comprehension theory.
The next step in analyzing the data was to find out to what extent the participants’ inferences were correct and if this was related to participants’ level of proficiency. Thus, the frequency and percentage of the number of times the readers in each level had made correct and incorrect inferences were calculated and are shown in table 5.2
As mentioned in the methodology chapter, To test the useful of inference strategy of this study, the researcher used proficiency level readers conducted the experiment first. The result will be shown on table 5.2
PROJECT
READING
Silk road
Average
26.74
Standard deviation
4.51
Computer
Average
26.43
Standard deviation
3.27





Table 5.2 the reading score for proficiency level group
On the third step, the researcher recommended some inference strategies to the medium proficiency level readers participants. The strategies including:
a) Three steps to inferring to understanding text
Read some text.
Write what we know.
Write what we infer or interpret.
Facts (Something We Can See and Observe)
Inferences (Interpretation)
b) Inferring and questioning
Inferring and questioning go hand in hand to build understanding. As they read students asked questions to clarify their uncertainty and made inferences about the thoughts and actions. When students say things like, “Well, maybe it means…” or “I wonder…” we are quick to label these comments as inference, but even when the a question is generated, an inference is rarely far behind. Inferring and questioning are next of kin.

Questions
Inferences


c) Inferring and finding information
Choose a book that provides a lot of questions, most of them requiring readers to infer answers by using clues in the text and their prior knowledge. Choose content that is difficult for students to read independently.
Read a few pages and begin recording some questions that pop into your head. After a few pages invite the children to share their questions and add them to the list the teacher has already started. By the end of the story, there should be a long list of questions.
Read the text again to figure out the answers. Teacher should model aloud using her schema, the pictures, and the words in the text to help me infer answers; gradually I invite children to do the same, recording our thinking on the chart.
d) recording the response
Writing their initial responses in a journal or literature log after reading. An important point to discuss when debriefing this strategy is how the initial vocabulary words were selected. Sharing the thinking behind how a teacher determines challenging vocabulary that is vital to comprehending a text helps to elevate this strategy to a level of independent use for students.
Facts (Something We Can See and Observe)
Inferences (Interpretation)


Questions
Inferences


Word
What we infer it means
What helped us?



After practicing these inference strategies, the medium proficiency level readers continue to the same test. The result will be shown on table 5.3.
PROJECT
READING
Silk road
Average
28.81
Standard deviation
3.69
Computer
Average
26.94
Standard deviation
3.45







Apparently, according to From the preceding tables the researcher noticed that the average for proficiency level group = 26.58, compared with the average for medium proficiency level readers = 27.87. On the other hand, when the medium proficiency level group was conducted, the result indicated showed that the average had tremendous better than the proficiency level group, although at first the ability of proficiency level group are better based on their IELT score. These results indicate that inference skills are very useful on reading comprehension.
















Chapter 6. Conclusion
Background knowledge has an effect on comprehension. The idea that this knowledge is codified into mental schemata (as proposed by e.g. Carrell 1983) lacks empirical confirmation. Nonetheless, 'it has been shown that even across passages on the same general theme, which had identical structure and syntax and very similar vocabulary, the more familiar version is better recalled' (Alderson 2000:43). This is true a fortiori when the text is situated in an unfamiliar culture: Steffensen et al. (1979) gave texts about weddings to L1 readers from India and North America and observed that cultural familiarity or the lack thereof led the readers to make numerous inferences about the events and situations in the text. Giving L2 readers access to information about the L2 culture can be an important way of helping them with reading comprehension.
The interaction between the content and linguistic simplification indicates that the readability of a text should be considered as the interaction of a multitude of factors. The results show that the linguistic simplification facilitated the comprehension and recall of the content-unfamiliar text, whereas it had an impeding effect on the comprehension and recall of the content-familiar text. This finding, along with the previous literature on text selection, may shed new light on more valid approaches to reading materials for EFL students with different proficiency levels and schematic backgrounds. The findings confirm the previous research indicating that input modification (simplification in this case) does not necessarily result in greater readability for EFL learners. However, the findings of this study need to be verified against future research involving more representative samples comprising both males and females. Similarly, the small size of the reading test batteries might have distorted the pattern of findings. Therefore, further research needs to be conducted before solid generalizations can be made.
High prior knowledge readers performed better on the comprehension questions than low knowledge readers. This advantage also did not depend on the quality or quantity of the self-explanations while reading the text.
When skillful readers happen upon information or words they have not learned, they activate their background knowledge to relate the new information or words to previously acquired information or word knowledge. Even when they simply read passages, they can bring out connected information as much as possible from their prior knowledge. For example, while reading a text about fish, they tap their schemata to gather information regarding fish, such as types of fish, places they live, and their characteristics. Making associations with the text information helps them understand the text. On the other hand, struggling readers do not connect new information or words to old by making use of their background knowledge. This isolates struggling readers and causes increased difficulty.
Readers have to read between the lines and relate the text information that has a less explicit explanation to their existing knowledge for better comprehension. The use of background knowledge assists readers to grasp the content of expository texts, as they can refer to their prior knowledge in order to fill in gaps between old and new information. On the other hand, struggling readers have no idea what to do when they read a text because their critical thinking does not get involved in their reading process. They either keep reading without understanding much of what is written, or stop reading when they think they cannot understand any further.
This paper has examined expository texts in terms of elements of expository texts, characteristics of skillful readers, and useful approaches. High-performing ESL/EFL readers enjoy reading because they have the ability to manipulate their background knowledge in order to comprehend unfamiliar information, words, or concepts. They make inferences so that they can fill in gaps between new information and existing knowledge.