Institute/Department Gender
Instructor A Real Estate Male
Instructor B Real Estate Female
Instructor C Real Estate Male
Instructor D Real Estate Male
Instructor E Marketing and Logistics Management Male
Instructor F International Trade Female
Instructor G Risk Management and Insurance Male
Instructor H Risk Management and Insurance Female
Instructor I Construction Engineering Male
Instructor J Construction Engineering Male
The researcher firstly e-mailed to the professors to request their agreements to participate in this study. However, the researcher began to make phone calls instead of e-mails to enhance the efficiency. Every instructor that accepted being surveyed was also asked to take part in the interview process. Unfortunately, due to the heavy teaching load and busy schedules, only 8 of the 18 instructors underwent the in-depth interview process after being surveyed. In addition to the 8 instructors, two more instructors who were not surveyed accepted the interview. One instructor accepted being interviewed but refused to do the survey because he believed the interview could convey enough information. The
other instructor typed out her responses to the interview questions and sent them to the researcher by e-mail, but gave no response to the requirement of survey.
Instruments
The instruments used for this research included two surveys as well as interview questions. Before collecting the data, the researcher required two English-major professors to review the Chinese surveys and one English major professor to examine the Chinese interview questions to give suggestions to establish expert validity. The instruments included two surveys in Chinese version. One was used to investigate the viewpoints of graduates and the other was used to investigate the viewpoints of instructors who taught non-English major graduates. The student and teacher surveys were similar in many ways, but not identical. Each survey included nine different categories. Some items related to metacognitive strategies were adopted from Mokhtari and Reichard’s survey
(2004).Category one was designed to explore graduate students’ personal background.
Category two was provided to explore the importance of vocabulary knowledge in academic English reading. Category three was used to investigate the significance of grammar knowledge in academic English reading. Categories four and five revolved around the subject of schema theory. Category four was designed to explore the
significance of formal schemata while Category five was set to study the subject of content schemata.
According to Mokhtari and Reichard’s classification (2004), category six to eight were designed to investigate the influence of metacognitive strategies in academic English reading. Category six was related to global reading strategies, and was oriented toward global analysis of texts in reading. Category seven involved problem solving strategies, which could be applied to solve problems when readers were having difficulties
understanding a text. Category eight concerned supporting reading strategies, which included the use of outside reference materials for reading comprehension. Mokhtari and Reichard’s research did not investigate the influence of self-evaluation strategies. However, these strategies were associated with metacognitive strategies and were demonstrated by Oxford (1990). Thus, Category nine was designed to investigate readers’ application of self-evaluation strategies.
The formal survey was revised after a pilot study. The original survey consisted of 54 items, which were also divided into nine categories. Category one included six items (1~6).
Category two also included six items (7~12) as did category three (13~18). Category four consisted of five items (19~23), and category five included six items (24~29). Category six included seven items (30~36), and category seven included other seven items (37~42).
Category eight included six items (43~48), and category nine included the final six items (49~54).
The pilot study was conducted in May, 2006. Seventeen subjects were involved. All were males and in their first year of graduate school majoring in Information Management
in Pingtung. The 17 subjects completed the surveys in Chinese version together and all 17 surveys were immediately available for analysis.
The data was analyzed by a Statistical Package for Social Science (SPSS 12.0) to check the reliability of the survey results and the classification of the nine specific categories. The Cronbach’s alpha of the survey as a whole was 0.929. The Cronbach’s alpha for category one was 0.740. The Cronbach’s alpha for category two and three were 0.542 and 0.564; for category four and five were 0.641 and 0.431. Furthermore, the Cronbach’s alpha for category six, seven, eight, and nine were 0.828, 0.828, 0.813 and 0.871, respectively. Three subjects claimed that 54 items were too many, so they felt it increasingly difficult to concentrate on study nearing the end of the survey. Thus, in order to solve this problem and raise the reliability of each category, the researcher deleted some items according to the correlation between each item and the category. The correlated item-total correlation of each deleted item was lower than the correlation of other items among the nine categories; that is, the item deleted was not highly correlated to the theme of the category.
Items in the first category were designed for the collection of personal information.
Therefore, items in that category remained the same before and after revision. In category two, items eight and nine were deleted. The correlated item-total correlation for item eight was 0.015, and for item nine it was 0.066. With the revised survey, the reliability for
category two rose from 0.542 to 0.736. In category three, item thirteen and fourteen were deleted. The correlated item-total correlation for item thirteen was 0.035, and the
correlation for item fourteen was 0.103. After revision, the reliability in category three rose from 0.564 to 0.704. In category four, the correlated item-total correlation for item nineteen was 0.178 and was also deleted. After that, the reliability improved from 0.641 to 0.666. In category five, item twenty-five and twenty-seven were deleted. The correlated item-total correlation for item twenty-five was -0.167, and the correlated item-total correlation for item twenty-seven was 0.005. After revision, the reliability rose from 0.431 to 0.686. No item in category six was deleted. In category seven, item thirty-seven and forty were deleted. The correlated item-total correlation for item thirty-seven was 0.420, and that for item forty was 0.439. The reliability rose from 0.828 toward 0.911. In category eight, only item forty-seven was deleted. The correlated item-total correlation for item forty-seven was 0.400. After revision, the reliability rose from 0.813 to 0.823. In category nine, item forty-nine was deleted as well. It had a correlated item-total correlation of 0.507. After revision, the
reliability rose from 0.871 toward 0.875. In total, eleven of the original fifty-four items were deleted. The Cronbach’s alpha of the revised survey improved to 0.929 from 0.917. The Cronbach’s alpha before and after elimination were shown in Table 3.1.
Table 3.4