The current study involved a repeated measures design, which in addition to a pre-test, post-test and delayed post-test, also included several periodic testing sessions throughout the 8-week course of the treatment. The purpose of the repeated measures design was to reveal a clearer picture of the participants’ performance on past tense over time. For example, the researcher wanted to know when participants begin to improve on their use of past tense, how much their performance of past tense improved or declined during the 8 week treatment session, and how robust their performance of past tense was after the treatment had ended. As mentioned above, an oral elicitation test was used to determine whether participants were making gains in their spontaneous, productive use of past tense.
The Oral Elicitation test consisted of a total of nine measures, including one pre-test, six weekly tests, one post-pre-test, and one delayed post-test.
On the Oral Elicitation test, performance on the target structure (English past tense) was scored as the number of test items on which a participant correctly produced past-tense. Each test item consisted of one sentence and each sentence
contained one past-tense verb. If a participant produced a past-tense sentence using the correct form of the past-tense verb, the participant received 1 point. If a participant produced a past-tense sentence without the correct form of past-tense, he received 0 points. The highest score possible was the same as the number of selected, past-tense items on the test. There were 30 past-tense items on the test so the highest possible score was 30. The lowest possible score was 0. (See table 2 for a summary of the description, purpose, and scoring of the test.)
Table 2
Description, Purpose and Scoring of Testing Measure
_____________________________________________________________________
Testing Measure Description Scoring Purpose
_____________________________________________________________________
Oral Elicitation The participant To test participants Every time a Test hears a sentence, on spontaneous, participant repeats
says if he oral production of a past tense understands it, says past-tense verbs. sentence without whether he agrees To have a producing past or disagrees, and controlled and tense, he receives then repeats the consistent 0 points. Every sentence dimension to the time the
immediately assessment participant repeats without too much method. a past tense
deliberation. sentence using
past tense correctly, he receives 1 point.
The maximum
score is 30; the minimum score 0.
____________________________________________________________________
Acquisition of the target structure (past-tense) was operationalized quantitatively according to two criteria: the first being the difference between the participants’ pre-test and end-pre-test scores, and the second being the robustness of the participant’s improvement on the delayed post-test. For example, if a participant’s test score at the end of treatment was statistically higher than his or her pre-test score, the participant would be considered to have improved on his or her acquisition of past tense. And if a participant’s delayed post-test score was statistically higher than his or her pre-test score, the participant’s improvement in the acquisition of the target structure would be considered to have reached some level of robustness.
Acquisition of the past-tense was operationalized as statistically significant improvement in the productive use of past-tense verbs. Acquisition, as operationalized in this study, did not require one hundred percent accurate production of all past-tense English verbs in every instance. As long as participants showed statistically significant improvement on the frequency of the correct, productive use of past-tense verbs, they were considered to have improved on their acquisition of past tense.
Chapter Four: Results
As can be seen in Table 3 below, the amount of corrective feedback received by participants in each of the three recast groups was very similar. During the eight-week data collection period all research participants received a total of 8 eight-weekly, 30-minute sessions of conversation practice. Other than the control participant, which received no recasts, the total number of recasts administered during conversation practice to each of the three recast treatment groups was very similar with the
paraphrased, elaborated, and standard-recast participants each receiving 107, 105, and 116 recasts, respectively. A one-way ANOVA showed no statistical difference between the number of recasts received by each type of recast participant and it can therefore be concluded that the amount of corrective feedback (recasts) received by the research participants was probably not a factor in their comparative performance on English past tense.
Table 3
Number of Recasts Received by Each Treatment Type
____________________________________________________________________
Session Date Treatment type
____________________________________________
Paraphrased Elaborated Standard Control Recasts Recasts Recasts No recasts ____________________________________________________________________
1 11/15 25 19 26 0
2 11/22 15 21 27 0
3 11/29 14 12 18 0
4 12/06 03 16 13 0
5 12/13 17 15 20 0
6 12/20 11 06 06 0
7 12/27 10 08 04 0
8 01/03 12 08 02 0
____________________________________________________________________
Total 107 105 116 0
____________________________________________________________________
In response to research question number one, the results confirmed the researcher‘s first hypothesis that all groups which received recast treatments would
outperform the control group which received no recasts. Test scores of all recast groups proved to outperform those of the control, and the scores of the three recasts groups revealed an overall upward trend of improvement over time compared to the control group, which showed no overall improvement (see table 4).
Table 4
Weekly Test Scores and Mean Test Scores for Each Treatment Type
____________________________________________________________________
Session Date Treatment type
____________________________________________
Paraphrased Elaborated Standard Control Recasts Recasts Recasts No recasts ____________________________________________________________________
However, the researcher was unable to prove any statistically significant differences in performance between any of the three recast groups and the control group. A series of paired Wilcoxin-signed-rank tests also showed no statistically significant within-group differences for the three recast groups tested together as a
whole.
As for research question two, the results did not confirm the researcher‘s
hypothesis that elaborated recasts would outperform the paraphrased and standard recasts and that standard recasts would outperform paraphrased recasts. On the contrary, the standard recast group slightly outperformed the elaborated and
paraphrased recast group and the paraphrased recast group slightly outperformed the elaborated recast group. All recast groups performed about the same as can be seen from the overall mean test score gains in table 4 above, and the researcher was unable to prove any statistically significant differences in the past-tense performance between the three recast groups.
Regarding research question number three, the results were affirmative; the performance of language learners on repeated measures of past-tense (after receiving recast treatment) did reveal a pattern over time? As can be seen from the figure below, there were four patterns that appeared over the 9 weeks of treatment and testing.
These four patterns are as follows: (1) All recast groups exhibited a non-linear
improvement over time on their past-tense performance; (2) All recasts groups exhibited periods of fluctuation in past-tense performance prior to continued growth;
(3) the elaborated recast group continued an upward trend of improvement on the delayed post-test while the other two recast groups did not; and (4) fluctuation in past-tense performance occurred first with elaborated recasts, second with paraphrased recasts and third with standard recasts.
Figure 1
The Performance Trends of the Three Recast Types Over 9 Weeks
In addition to being tested on their accuracy of past tense, all participants were also given an exit questionnaire (see appendices D-1 and D-2 for details) in an attempt to find out how focused they were on form versus meaning and how conscious they were of the corrective feedback. Although, the researcher did not inform any of the research participants that they would be receiving recasts on past tense, the results of
0
Recast Types Performance on Past Tense
Paraphrased Elaborated Standard
the exit questionnaire revealed that all three participants (who had received recasts) had noticed that they were being given corrections on English past-tense during the
conversation sessions. We can therefore conclude that major differences between participants‘ noticing of the corrective recasts was probably not a confounding
variable in the participants‘ past-tense performance.
The other items on the exit questionnaire showed that there were only minimal differences between each participant‘s overall focus on form and meaning. On items
1 through 6e, the standard recast participant scored 2.9 on the questionnaire, showing that he was slightly more focused on form than the other three participants who all scored 3.5. However, in general, it seems that all four participants were similarly
focused on form and meaning during the conversation sessions.
Perhaps of more significance is that on the exit questionnaire‘s open-ended
questions, only the standard recast participant stated that he was aware that the oral elicitation tests were testing grammar accuracy of past-tense. The other research participants stated that the purpose of the tests were to test such things as fluency, listening comprehension, and overall English proficiency. (SeeAappendix D-2.) Considering that all the research participants were told that the purpose of the tests was listening comprehension, not past tense accuracy, the standard recasts knowledge
of the purpose of the test may have given him a distinct advantage over his counterparts.