The effectiveness of the multimedia interaction was measured by comparing vocabulary recognition and comprehension scores on two texts: one glossed, the other not. To create an effective device, the author made every effort to match the characteristics of the learners to the content of the lesson and its presentation.

This study, therefore, looked at the general characteristics of the study group, their age, grade level and cultural factors. These factors helped determine the level of the lesson and enabled the selection of text examples that were meaningful to the group. In terms of specific competencies and content-related qualities that more directly affected decisions about media, methods, technical vocabulary and technical skills, some were taught while others were already within the studentsí repertoires. These are explained in the Practice Sessions section below. In the final analysis, the author endeavored to ensure that there were no biases or misconceptions about the subject.

The story-line was on the topic of ìLíAmitiéî which is a theme ordinarily taught by Grade Four teachers in the fall of the year. The first text is full of emotive phrasing and apprehension on the part of the narrator, Amélie, as she talks about her situation in relation to her family. We learn that Amélie has no mother. Her father, baby sister and ìbonneî have just moved to a new town. The second excerpt continues in a similar vein. Amélie is starting school tomorrow but she doesnít know anyone. Worse still, she has to bring her baby sister, whose constantly runny nose, caused by allergies, is a source of embarrassment for her. (Appendices 1 and 2)

To determine pre-test, treatment and annotational items, input from potential users has been found to be crucial in developing appropriate software. (Davis, Lyman-Hager and Hayden, 1992) Therefore, during the first week of September, 1999, four Grade Four French Immersion teachers from different schools within the province of Newfoundland and Labrador assessed the two selected texts.

Each text, approximately 250 words in length and similar in difficulty level, was an excerpt from the first chapter of ìAmélie et le mystère de l'homme Xî, part of the recommended Language Arts resource material for this grade level. Teachers used a Likert scale to rate the difficulty level of the document from 1 to 5, where 1 indicated very easy to comprehend and 5 indicated very difficult. Three of the teachers positioned their particular text at a level 4 and one placed it at a level 5 degree of difficulty. This was deemed to satisfy Krashenís (1983) comprehensible input theory, ìi+1î, where i represents the current comprehension level; the learning process is assumed to be triggered when there is a gap between a structure in the input and the learnerís current level of competence.

For appropriate annotational items, the teachers randomly selected students to perform one of three activities with a text. Some individuals underlined unknown words and expressions, others wrote definitions of certain words already underlined in the text (these underlined words were chosen by the teacher on the assumption they were unknown to the student) and, for a third group, each student was asked to write down, in French, everything they recalled in the text. (written recall protocol developed from Davis, Lyman-Hager and Hayden, 1992)

A problem arose when data from the students who underlined unknown words and expressions was not consistent with that of the group who wrote definitions. In the first instance, a large portion of the students indicated that they knew most of the words in the particular text and several indicated that they knew all the words. However, from those who wrote definitions, it was obvious that the majority of the words were misunderstood since they were incorrectly defined. Similar findings were also noted in the Davis, et. al (1992) study.

Student recall protocol results were correspondingly affected in that the majority of the propositional units (idea units or information units as per Chun and Plass, 1996) identified by teachers were not identified by students. To override these problems, vocabulary test items, propositional units as well as words and expressions to be annotated were chosen from those identified and listed by participating teachers. Ultimately, if a word or expression was identified by 2 or more teachers as being unknown to the students, it was glossed.

For the vocabulary pre-test and treatment, 10 words were randomly selected from the annotated items and scored dichotomously: a subject was given a one for every correctly identified word and a zero for those not identified. (Chun and Plass, 1996) The same scoring scheme applied for comprehension. Cooperating teachers identified 10 propositional units in the pre-test excerpt and 8 units in the treatment text. While it is not feasible to analyze every psychological or educational trait of each participant, several factors came into play about young learners in general that were critical for making good media and methods decisions.

Return to home page