usefulness of the GO map instruction to the posttest retelling, and the reason behind
their responses were solicited. Besides, the researcher also interviewed the
participants whose responses to the open-ended questions in the posttest questionnaire
needed to be clarified. The interviews, therefore, served to elicit more detailed
thoughts from the participants toward the GO map instruction.
3.4 Data Analysis
The results of the story retelling performances on the pretest and posttest were
analyzed from five aspects: (1) the story length as indicated by the number of words
contained in the participants’ retelling performances, (2) the fluency in terms of the
original and pruned speech rate, (3) the story element count score and story content
score, (4) the level of the overall retelling performances and (5) the participants’
responses to the GO map instruction. In terms of the calculation of the story length,
the participants’ retold stories were first transcribed. The original word count was
calculated by the number of words in the retelling transcriptions and the pruned word
count was gained with the number of words of false starts and repetitions deducted
from the original word count.
The fluency of the participants’ retelling performances in both the pretest and
posttest was determined by using the original speech rate and the pruned speech rate.
The original speech rate is calculated by dividing the number of words by the total
35
time (words per minute, i.e., “w.p.m.”) (Lennon, 1990). Pruned speech rate, which
is based on Lennon’s (1990) concept of “pruned speech,” is obtained by calculating
speech rate exclusive of false starts and repetitions. The reason for adoption of the
pruned speech rate is that with the false starts and repetitions excluded, the pruned
speech rate substantially reflects the efficient and meaningful content the participants
produce per minute. That is, the pruned speech rate combines speed of speech with
its efficiency in terms of how much repair is required; hence, it serves as an
appropriate measure of fluency.
The story elements in the participants’ retold stories were evaluated in two ways:
the story element count score and story content score. The story element count score
was calculated by how many story element items were included. When an element
was mentioned, a point was awarded. The highest score is seven for there are seven
story elements. The story content score in the participants’ retelling performances
were analyzed and graded according to the grading checklist (see Appendix N). The
scoring criteria on the checklist are adjusted from Morrow (1986). The GO map
checklists (see Appendices E-2 & L-2) for the two retold stories helped provide a
reference for the researcher to judge how many points to assign regarding the story
element content and sequence. The grading checklist were used as criteria for
assigning points to the story elements included and the sequence in which the story
36
elements were organized in their retelling performances. The weight for each story
element varies. For each story element and the sequence, the degree of
completeness and detailedness determines the score the participants got. The
elements on the checklist include “Story Title”, “Setting”, “Characters”,
“Problem/Goal”, “Events”, “End” and “Sequence” with different maximum points,
ranging from the highest 4 to the lowest 0.5. The maximum point is determined by
the proportion of the particular element in stories. Specifically, since the element,
“Events”, comprises the major part of stories, the participants could get as many as
four points if they successfully mentioned eight or more events. “Setting,” including
when and where, has a maximum point of 1.5. For “Characters,” “Problem/Goal,”
“Sequence,” and “End”, the maximum point is 1. As for “Story Title”, since its
proportion is relatively small, its maximum point is 0.5. The primary principle for
assigning points depends on how the participants’ retelling content fitted the
description on the GO map checklist. Specifically, if the retold story content
corresponded with the description of an element, the participants could get the
maximum point. If the content of an element was incomplete or partially wrong,
they could still get some points for the part they recounted correctly. However, if the
content of an element was outright missing or completely twisted, they got no points.
As for “Sequence,” if the participants described the events according to the sequence
37
in which the events take place without making mistakes, a point was rewarded to
them; if there were a few mistakes but the story order was still understandable, they
got 0.5 point. If the participant recalled the story with little or no sequence, they got
no point. The points for each element were accumulated to gain the final score.
The highest total score was ten and the lowest zero. The teacher-researcher and her
colleague, who is also an English teacher, conducted all the ratings for the story
element content and holistic level of the participants’ retellings. They analyzed the
participants’ performances together and settled the differences of their assessment by
discussion. A score was final once they had reached consensus.
The holistic evaluation of the participants’ performances was carried out based
on the holistic evaluative criteria developed by the researcher (see Appendix O); it’s a
five-level evaluation form using such criteria as (1) pronunciation, intonation, and
fluency, (2) organization, completeness, sequence, coherence, cohesion and
supporting details, (3) grammatical and syntactical structure and lexical use and (4)
paraphrase. The participants’ retelling performances in the pretest and posttest were
evaluated and then assigned to a level in a holistic fashion.
All the quantitative data obtained from the pretest and posttest was processed via
t-tests. The independent t-test was conducted to investigate whether there was a
significant difference between the EG and CG regarding a particular aspect of their
38
retelling performances. The paired-samples t-test was conducted to explore whether
there was a significant difference regarding a particular aspect of their retelling
performances between the pretest and posttest within the EG or the CG. In this case,
all the quantitative data gathered in the pretest were first processed via the
independent t-test to see if there was a significant difference between the EG and CG
in the pretest on a particular aspect of their story retelling performances. The fact
that no significant between-group difference was found in the pretest performances
indicated that the EG & CG were similar in their retelling ability before the study.
Under this condition, a paired-samples t-test was then conducted to see whether there
was significant difference between the pretest and posttest in each group. A
significant difference indicated that the participants in that group progressed
significantly from the pretest to the posttest on that aspect. When both groups
displayed a significant within-group difference, another independent t-test was
conducted on the posttest to see whether there was any significant difference in the
posttest. If there was a significant between-group difference in the posttest, it
suggested that the group with higher mean value in the posttest made more progress
than the other group. If there was none, it indicated that both groups made equally
significant progress.
Finally, the responses of the participants in the EG to the multiple-choice
39
questions in the posttest questionnaire were presented with descriptive statistics.
Their responses to the open-ended questions in the questionnaire and to the probing
questions in the interviews were categorized, synthesized and analyzed.
40
CHAPTER FOUR RESULTS AND DISCUSSION
In this chapter, the effects of the GO map instruction on students’ story retelling
performances are presented and discussed. The effects of the GO map instruction on
the word count are examined first. Next, the effects on the fluency are shown.
Third, the effects on the number and content of the story elements in the participants’
story retelling are addressed. Fourth, the effects concerning the holistic story
retelling performance are discussed. Finally, the results of the participants’
perceptions of the GO map instruction are delineated.