Genres in academic writing: Research report methods

Use Of A Writing Web-Site By Pre-Masters Students On An English for Academic Purposes Course.

A. J. Gillett, University of Hertfordshire

Method

¹The aim of the research was to compare the success of students using on-line teaching compared to students taught in the traditional classroom context. ²This experimental approach adopted a positivist position by collecting primary quantitative data from the students’ performance on an objective test taken after the teaching process has ended. ³It then compare the observed results and came to conclusions regarding the effectiveness of on-line teaching.

⁴Two groups of international students on a one-year Pre-Masters English for Academic Purposes course, each comprising 50 students were taught academic writing by different methods and compared. ⁵In each group there were 50 students from five different academic departments - computer science, business, engineering, life sciences and law. ⁶The subjects were selected from the second semester - Semester B - of the University of Hertfordshire International Bridging Programme in the 2004-2005 academic year. ⁷This programme accepts only students from a narrow English Language Proficiency band (IELTS 5.00 - 5.5). ⁸Thus, comparable language level among the test subjects was insured.

⁹The subjects were selected from the 250 students on the International Bridging Programme on the basis of performance at a satisfactory level in the Semester A examination. ¹⁰Students who had performed below the minimum level on the semester A examination were excluded. ¹¹This criterion was employed to ensure competent understanding of the tasks and adequate motivation.

¹²One group - Group A - studied English writing in the traditional way in a class with a teacher. ¹³This class met for 2 hours each week in a classroom for 12 weeks and was supplemented with written homework assignments given by the teacher each week. ¹⁴The second group - Group B - met together in a class with a teacher for one hour per week for 12 weeks and were assigned a homework task of spending one hour per week doing exercises from the UEfAP web-site (Gillett, 2005).

¹⁵The test instrument employed in this study was a revised version of the University of Hertfordshire English Language Writing Test (Roberts, 1997), which permits the assessment of academic written language performance. ¹⁶It consists of an academic reading text and comprehension questions, followed by a discursive essay on the subject of the reading text.

¹⁷Both groups A and B were given the same written examination at the end of the semester. ¹⁸The students took the examination under standard university examination conditions as part of their end of semester examination. ¹⁹The tests were marked using the following categories: task achievement; communicative quality; organisation; ideas, content and relevance; and grammar and vocabulary, by two experienced writing examiners and moderated in the standard way to ensure reliability. ²⁰In this way it was possible to see the relationship between the students' main academic subjects, and the improvement in their writing ability depending on the teaching method.

²¹A 3 x 5 analysis of variance was used to test for academic department, method of teaching and language achievement differences.

In early 1982 telephone interviews were conducted with a statewide probability sample of 2,083 registered voters in a major southwestern state. The interviews were conducted for a state agency and addressed various voting-related attitudes and opinions.

Within this context, a split ballot (experimental) design was employed whereby approximately each quarter of the sample was asked age utilizing a different question format. Three open-end and one closed-end question formats were investigated:

• How old are you?
• What is your age?
• In what year were you born?
• Are you 18-24 years of age, 25-34. 35-49, 50- 64, 65 or older?

Each question format was drawn from previous research and was selected to be illustrative of one approach to asking age. The particular question format used when asking an individual study participant his or her age was randomly determined prior to the interview. Interviewers made no determination as to what age question format was employed for a specific study participant.

All interviews were conducted from a centralized, supervised interviewing location and began with an interviewer asking to speak to a prespecified individual. The interviewer then introduced himself/herself and stated who was conducting the study and asked for the potential study participant's cooperation. The questionnaire consisted of 20 questions, of which the age question was number 15.

Actual age data were available from the state agency for 1,324 of the individuals interviewed. Therefore, following the completion of an interview it was possible to compare an individual's reported age with his or her actual age. This in turn permitted inferences as to which question format produced the most accurate age data as well as which format resulted in the lowest refusal rate or nonresponse rate.

Telephone interviews were conducted.

The interviews were conducted for a state agency.

A split ballot design was employed .

Each quarter of the sample was asked.

Three open-end and one closed-end question formats were investigated.

Each question format was drawn from previous research and was selected to be illustrative of one approach to asking age.

The particular question format ...  was randomly determined prior to the interview.

All interviews were conducted from a centralized location.

The interviewer then introduced himself/herself and stated who was conducting the study and asked for the potential study participant's cooperation.

The questionnaire consisted of 20 questions, of which the age question was number 15.

Actual age data were available from the state agency for 1,324 of the individuals interviewed.

Therefore, following the completion of an interview it was possible to compare an individual's reported age with his or her actual age.

This in turn permitted inferences as to which question format produced the most accurate age data

as well as which format resulted in the lowest refusal rate or nonresponse rate.