Data processing

In order to interpret the results of exploratory teaching, the classroom data need to be “processed”, i.e. arranged in the form that makes the data more evident and present them for easier interpretation. Data processing can be done in the graphical form. A popular graphical form pr presenting data is a table. It shows the research data in their numerical values usually per cent.

Questions	Answers %

Another popular form of representing the data graphically is a bar-chart. It helps to compare and contrast the questionnaire results and to represent them in a visual form, which easy to view and to analyze.

Strongly disagree	Disagree	Don’t care	Agree	Strongly agree

An alternative form to represent the data relationship and to show them as “part of the whole” is a “ apple pie”. An “apple pie” given below shows the ratio of “syntax errors” (word order) and “morphology errors” made in the students’ written essays. The section of the syntax errors is much smaller than that of the morphological errors, which is typical of the error distribution.

In order to represent the experimental data dynamically and to show how a certain phenomenon (errors, attention, use of communicative activities in the lesson, difficulty of the material etc) was changing in the course of research, a graph is used. The graph below shows a typical tendency in the learners’ errors while they are acquiring a new grammar item. In the beginning the number of mistakes is relatively low and the scoring is high. Later the number of mistakes goes up dramatically (the scoring goes down) and after more efforts the number of mistakes goes down again (the scoring goes up). This tendency is represented by a typical U-turn, which is known to many researchers into the process of learning.

Scoring

U

Chain of lessons

Statistical analysis

In the process of exploratory teaching it might turn out to be useful to use statistical analysis, i.e. analyzing numerical data. A teacher might analyze students’ scores on a test or examination, evaluate responses to a questionnaire, estimate the importance of certain items for the learner, investigate the proportion of themes and teaching time, study the types of responses in the lesson, measure the teacher’s waiting time when the learners find it difficult to respond etc. (After McDonough, J. and S.McDonough. 1997. Research Methods for English Language Teachers. L. P. 138). Statistical analysis at present has become available and easy because of the wide use of statistical software. A widely used way to represent the data is per cent. A basic statistical concept is the mean. It is an average of a set of scores obtained by adding individual scores together and dividing by the total number of individuals who took part in the testing procedure. Another useful tool is standard deviation that shows the extent to which certain individual scores vary in relation to the mean. Let’s consider an example. The group “mean” of a certain test is 50 points. One student scored 65. It this result exceptional or ordinary? If the standard deviation from the mean is 20 points, this signifies that most students have scored between 30 and 70 points. Then the result of 65 points is ordinary. If the standard deviation is only 5 points and the majority of the students scored between 45 and 55 points, then the result of 65 points is exceptionally good. When teachers compare the data before pedagogical intervention and after it, it is necessary to prove statistically that the changes are significant. When comparing the two sets of results (scoring on the two tests), the appropriate statistical procedure will be the t-test or the F-test (another name of this procedure is analysis of variance, (ANOVA). A more simple procedure is to calculate the “means” of the scoring in the initial and final tests. If, the means of the test scoring before and after pedagogical intervention are different, then the teacher’s efforts really resulted in a tangible change in learners. The concept that underpins much of the analysis is “correlation”. Correlation is the extent to which two sets of results are connected between themselves. Suppose that students have taken two tests. One test was in grammar knowledge and the other test was in lexical knowledge. If the scoring for grammar was higher, the scoring the lexis was equally higher. If the scoring for grammar went lower, the scoring for lexis also went down. In this case, the “variables” (rows) of the results in both tests have the same tendency and were clearly associated with each other. The correlation index according to computerized processing was +1,0.

Scoring

Дата добавления: 2015-10-24