The study was conducted in the Palo Alto Unified School District in northern California. Due to limited time and resources, a convenient sampling strategy (Alreck & Settle, 1994) was used to select the study participants from two local high schools. The participants included 124 students enrolled in Chinese classes. Given that the main experiment lasted for only four days, the 16 students (seven students in the control group and nine students in the treatment group) who were absent on any of the four days were excluded from the analyses. T test comparisons showed that the pretest results of the absent students in the control and treatment groups were not significantly different (p > 0.05), and therefore their absence would not bias the study. The participants ranged in age from 14 to 17 years. Among the 108 participants, 70 were male (65%) and 38 were female (35%). Thirty-five students (32%) were heritage speakers, and 73 students (68%) were non-heritage speakers.
This study employed a randomized experimental design. Students were randomized into a treatment group and a control group within the same class as follows. First, the original 124 students were homogeneously grouped by pretest scores. The pretest was a recognition test administered one day before the experiment and consisted of all 32 characters to be taught in the study. During the pretest, students were asked to write down the pinyin and meaning of the characters. Test papers were collected immediately afterward, and results were never shared with students so that students did not receive any instructional exposure to those characters except in the experiment. After ranking students by their pretest performance, matched pairs with successive ranks were formed. Second, matched pairs were randomly assigned such that one student participated in the treatment group and the other in the control group. Third, a few participants within a matched pair were switched to balance the gender distribution between the treatment and control groups. Table 1 presents information about participants' group, school, and level.
Table 1. Distribution of Experimental and Control Group
| ||Experimental Group||Control Group|
|Level II (about 120 instructional hours)||11 (School A) + 11 (School B)||12 (School A) + 11(School B)|
|Level III (about 240 instructional hours)||8 (School A) + 15 (School B)||9 (School A) + 15 (School B)|
|Level IV (about 360 instructional hours)||8 (School A)||8 (School A)|
|Total||53 students||55 students|
Students at both schools used the same textbook series, Nihao, and had completed approximately the same amount of instructional time and content based on their level of Chinese instruction. A standardized Chinese assessment instrument, the Standards-based Assessment and Measurement of Proficiency (STAMP), was administered to students approximately two weeks before the study. The results indicated that the majority of Chinese II students fell into the novice-low4 to novice-mid categories, the majority of Chinese III students fell into the novice-mid to novice-high categories, and the majority of Chinese IV students fell into the novice-high to intermediate-low categories.
Both the control and treatment groups were taught eight Chinese characters daily for four consecutive days for a total of 32 different characters (see Table 2). The characters were all selected from the word lists of Lessons 4, 5, and 6 in Nihao IV. The highest-level students (Chinese IV students at School A) in the study were learning Lesson 2 in Nihao IV when the study began. Therefore, the characters used in the study had never been taught formally in class to students. Five criteria were used to select the characters for the study: human memory capacity,5 number of strokes (i.e., density), frequency of characters, transparency of phono-semantic compounds, and learning in context (see Appendix B). By following these criteria, the daily level of difficulty was controlled to enable within-subject comparisons across the four instructional days.
For both the treatment and control groups, handouts on which characters and activities would be covered were provided to students. A substitute Chinese teacher (see “Procedure” below for details) was trained in using the handouts for the study lessons. The treatment group also received an extra handout illustrating types of Chinese characters based on their origin and the MIC method (see Appendix A). In addition to the handouts, Chinese-English dictionaries were available to all participants in the study's follow-up activities.
To prevent research outcomes from being influenced by subjective bias, the participants and the experimenter were blind to who was assigned to the experimental and control groups. A credentialed substitute Mandarin teacher was employed as the experimenter to teach all participants in both groups. The teacher was trained in the steps that should be followed in delivering the study instruction to the participants.
The experiment was conducted during the regular Chinese class periods in both schools. For the control group, the teacher used a traditional method (i.e., stroke-order rote memorization) to teach the Chinese characters. On Day 1, students followed the teacher's model, writing the character stroke by stroke (teacher-instructed traditional method). On Day 2, the teacher provided the character's stroke order on the board, and students practiced writing by themselves (teacher-cued traditional method). On Day 3, students were asked to study eight characters by themselves and practice each character's stroke order (student independent traditional method). The teacher was not in the classroom. On Day 4, the student independent traditional method from Day 3 was repeated for a new character set. This practice lasted for approximately 15 minutes on each day of the experiment. Afterward, students participated in “making up words and phrases,” an activity in which they had to look up characters in the dictionary and use those characters to make two words or phrases. This activity lasted approximately 10 minutes on each day of the experiment.
For the treatment group, on the day before the experiment, the teacher introduced the MIC method with three steps. The first step in this method was to introduce the origin and types of Chinese characters (see Appendix A, Part I). Because introducing all six linguistic breakdowns can be overwhelming for new Chinese as a foreign language learners, teachers can simplify the knowledge by leaving out phonetic loan characters and derivative cognates because they are the most difficult to deconstruct and make up less than 1% of all Chinese characters (Boltz, 1994; Wang, 1993). Therefore, in this study, only the first four types of Chinese characters (pictograms, simple ideograms, ideogrammic compounds, and phono-semantic compound characters) were introduced to students. The teacher emphasized radical knowledge when teaching the phono-semantic compound characters. For each type of character, she gave abundant examples of Chinese characters and demonstrated how they could be meaningfully interpreted.
Then, the teacher told students that phono-semantic compounds (xingsheng zi) comprised about 90% of Chinese characters and that only about 26% of the phono-semantic compounds were transparent or close to transparent. Definitions and examples of transparent and nontransparent characters were given to students. To memorize those nontransparent characters, the teacher suggested that students use bujian (chunks) for memorization. She introduced the method of bujian and indicated that only 132 of the most common bujian could be found in a list of the 3,500 most frequently used Chinese characters (see Appendix A, Part II). She also gave examples on how to chunk several Chinese characters.
In the third step, the teacher listed several characters that consisted of the same radicals or chunks; by doing so, she facilitated students' identification and association of those radicals, chunks, and characters (see Appendix A, Part III). The teacher then stated that, although the element of meaningful interpretation focused on the types of Chinese characters and radical knowledge and while the element of chunking focused on the structures of Chinese characters, students should combine both elements as there were overlaps between chunks and radicals that frequently connect the characters consisting of the same chunks or radicals. This three-step introduction lasted approximately 90 minutes and did not involve the characters selected in the study.
After students participated in the above three-step orientation to the MIC method, on Day 1 of the experiment, the teacher introduced eight characters to students by providing her own interpretation and memory tips based on the MIC method (teacher-instructed MIC). For example, when introducing the character (ren, recognize), the teacher mentioned that the left radical represented the meaning (talk, express out) and the right radical represented the sound (ren, similar pronunciation of ). When introducing the character (shi, recognize), the teacher mentioned that the left radical , again, represented the meaning (talk, express out) and the right radical represented the sound (zhi), emphasizing that the phonetic radical might not exactly match the sound of the compound character. Another example was the character of . The teacher told students that chunking could be used in memorizing this character and wrote the tip on the board: = + + + . On Day 2, the activities repeated Day 1's structure; however, the teacher changed the character set and encouraged students' own interpretation and character chunking by providing cues (teacher-cued MIC). For example, when she taught , she said, “This is a phono-semantic compound, which means one radical represents pinyin and the other represents the meaning. Who can share his or her memory tips?” When teaching , the teacher told students that they could chunk the character into familiar bujian and asked students which two bujian they could identify and in what other characters they had seen them before. On Day 3, students were asked to self-study the eight characters using the MIC method. The teacher was not present (student independent MIC). The student independent MIC method was repeated on Day 4, except that the participants studied a different character set. Similar to the control group, time was controlled for learning the eight characters each day (about 15 minutes). After learning the character set, the treatment group joined the control group for the words-and-phrases activity, which lasted approximately 10 minutes, in the same classroom.
A quiz was administered in the same classroom each day to both groups after the words-and-phrases activity (see Appendix C). In the first section of the quiz, after the teacher pronounced the Chinese characters one by one, students wrote the characters and their meanings in English. In the second section, students completed a character-recognition task in which they saw the printed characters and were asked to provide their pinyin spelling and meanings in English.
On Day 5, a week later, a retention test was administered to both the treatment and control groups. Items on this cumulative test were presented in the same format as the previous quizzes (administered on Day 1 through Day 4) and included characters taught on Day 1 through Day 4.
Finally, to examine whether the MIC method enhanced long-term student learning, an application test was given two months after the experiment. Rather than assess students on the Chinese characters they learned in the experiment, the application test followed the same format as the retention test but examined students on the characters they learned from their classroom teachers in the two-month period after the experiment finished. The intention of this test was to determine whether students could apply the MIC method to their own Chinese character learning. The experimental design is shown in Table 3.
Table 3. Design of the Study
| ||Day 1||Day 2||Days 3 and 4||One week later||Two months later|
|Content||Character set 1||Character set 2||Character sets 3 and 4||Retention test (Quiz 1 + 2 + 3 + 4)||Application test on Chinese characters taught in between|
|Control group||Teacher-instructed traditional method||Teacher-cued traditional method||Students' independent traditional method|| || |
|Treatment group||Teacher-instructed MIC||Teacher-cued MIC||Students' independent MIC|| || |
|Step 2 (Both)||Related activity||Related activity||Related activity|| || |
|Step 3 (Both)||Quiz 1||Quiz 2||Quiz 3, 4|| || |
The immediate, retention, and application tests applied the same four types of measurement to test students' memory of characters: (1) reading the character and writing down its pinyin, (2) reading the character and writing down its meaning, (3) listening to the character and writing down the character, and (4) listening to the character and writing down its meaning. For the two reading tasks, characters were presented on a handout and students were required to write down their pinyin spelling and meanings. For the two listening tasks, characters were dictated and students were required to write down the characters and their meanings.
The number and percentage of correct pinyin spellings, written characters, and character meanings were measured. Students earned one point if they wrote the pinyin correctly (including tones), one point if they wrote the character correctly, and one point if they wrote the correct meaning. When students provided the right pronunciation with an incorrect tone, they received a score of 0.75. When students provided an intelligible pronunciation that was almost correct, they received a score of 0.5. When students made a small mistake in the character writing, such as missing or adding a stroke, but the character was still recognizable, they received a score of 0.5. If the meaning provided was close, but not exactly correct, they also received a score of 0.5.
First, to make the analysis more concise and efficient, the two most relevant measurements—reading the character and writing down its pinyin spelling and listening to the pinyin spelling and writing down the character—were combined and renamed as the perception component. The other two tasks—reading the character and writing down its meaning and listening to the character and writing down its meaning—were also combined and renamed as the meaning component.
Multi-factor ANOVA was performed to analyze three main variables in the study: instructional differences as indicated by the variable day (four levels6), the treatment or control condition as indicated by the variable group (two levels), and different measurements as indicated by the variable component (two levels). To investigate the main effects of the MIC method and different instructional strategies on the test components as well as their interactions between each other, a 2 (group) × 4 (day) × 2 (component) multi-factor ANOVA was conducted.
In addition, two additional variables, language level as indicated by the variable level (three levels) and heritage or non-heritage students as indicated by the variable heritage (two levels) were also of high interest and were included in the results. To analyze the impact of the MIC method including the level and heritage variables, a 2 (group) × 3 (level) / 2 (heritage) × 4 (day) × 2 (component) multi-factor ANOVA was performed on the dependent variables.