Homogeneity of the four groups
Before implementing the study, it was crucial to ensure the four classes were homogeneous so that there was no significant difference among the four groups. As shown in Table 2, there was no significant difference among the four classes (p > .05) on the test of homogeneity of variances. This indicated that the participants in the four groups were not significantly different at the beginning of the learning session.
Table 2.
Test of Homogeneity of Variances (based on the results of Test 1)
Levene Statistic df1 df2 Sig.
1.263 3 59 .295
The result of the delayed test of all target words
As mentioned previously, the actual number of the participants varied for uncontrollable reasons (e.g., absence from school). All target words were reviewed a week before the last delayed test, and the participants who did not attend the review
- 56 -
session before the test were excluded. Participants who did not concentrate on the session (e.g., doing homework or reading novels) or did not attend the whole session (e.g., going to the bathroom or fetching water) were also excluded to make sure no unrelated factors were involved. Therefore, the actual number of participants in the vocabulary test of all target words was 74.
Table 3 shows the scores on the delayed test of all target words. A 2 (encoding sequence: traditional vs. adjusted) × 2 (nature of semantic encoding cue: word vs.
picture) analysis of variance (ANOVA) showed there was a significant difference among the four groups (F(3, 70) = 4.837, p < .01). Table 4 presents the descriptive data for the delayed test of all target words. The analysis of variance (Table 5) suggests that Group D (adjusted teaching sequence with word encoding cues) statistically outperformed Group C (traditional teaching sequence with picture encoding cues) on the delayed test of all target words. The results also reveal that word encoding cues with the adjusted teaching sequence (L1 semantic encoding followed by phonological encoding cues) worked better than the picture encoding cues with the traditional teaching sequence (phonological encoding followed by picture semantic encoding cues).
In other words, it appears that the traditional teaching sequence with word encoding cues successfully facilitates long-term retrieval. This answered the second
- 57 -
research question—word encoding cues are more effective than picture encoding cues.
Table 3.
ANOVA for the Delayed Test of All Target Words
Sum of Squares df Mean Square F Sig.
Between Groups 7,429.277 3 2,476.426 4.837 .004**
Within Groups 35,835.588 70 511.937
Total 43,264.865 73
Notes: * p < .05, ** p < .01
- 58 -
Table 4.
Descriptive Statistics of the Delayed Test of All Target Words
N M SD Std. Error
95% CI
Min Max Lower Bound Upper Bound
Group A 20 43.0 22.67621 5.07055 32.3872 53.6128 10.00 80.00 Group B 20 59.5 21.69768 4.85175 49.3452 69.6548 15.00 95.00 Group C 17 39.7 18.99594 4.60719 29.9391 49.4727 5.00 75.00 Group D 17 63.2 26.63064 6.45888 49.5431 76.9275 20.00 100.00
Total 74 51.3 24.34480 2.83003 45.7111 56.9916 5.00 100.00
Notes: Group A: phonological encoding followed by L1 semantic encoding cues;
Group B: picture semantic encoding followed by phonological encoding cues; Group
C: phonological encoding followed by picture semantic encoding cues; Group D: L1
semantic encoding followed by phonological encoding cues
- 59 -
Table 5.
Multiple Comparisons for the Delayed Test of All Target Words Scheffe
Lower Bound Upper Bound Group A Group B −16.50000 7.15498 .160 −36.9970 3.9970
The results of Delayed Tests 1 and 2
The previous section indicates that Group D (adjusted teaching sequence with word encoding cues) statistically outperformed Group C (traditional teaching
- 60 -
sequence with picture encoding cues) on the last delayed test. However, a significant difference was not observed between the first two delayed tests (Tests 1 and 2). As shown in Tables 6 and 7, no significant difference was observed (p > .05). Namely, for the first two weeks after the learning sessions, no significant difference was found; the significant advantage of Group D was not observed during the first two weeks.
Table 6.
ANOVA for Delayed Test 1 Sum of
Squares df Mean Square F Sig.
Between Groups 316.913 3 105.638 .122 .947
Within Groups 51,225.944 59 868.236
Total 51,542.857 62
Table 7.
ANOVA for Delayed Test 2
Sum of Squares df Mean Square F Sig.
Between Groups 4,141.164 3 1,380.388 1.591 .199
Within Groups 60,745.322 70 867.790
Total 64,886.486 73
- 61 -
The results of Delayed Test 3
After excluding two groups of participants (those absent from or who left during the third learning session and those who did not concentrate in class), the number of participants in Delayed Test 3 was 71. Though no significant difference was found on Delayed Tests 1 and 2, participants’ performances started to become significantly different after the third learning session. As shown in Table 8, a 2 (encoding sequence: traditional vs. adjusted) × 2 (nature of semantic encoding cue:
word vs. picture) analysis of variance (ANOVA) revealed that a significant difference was found (F(3, 67) = 7.853, p < .01). Table 9 presents the descriptive data for
Delayed Test 3. The analysis of variance (Table 10) suggests that the two groups presented with word encoding cues performed statistically better. Group A (traditional teaching sequence with word encoding cues) statistically outperformed Groups B and C. Group D (adjusted teaching sequence with word encoding cues) statistically outperformed Group C, revealing that the word encoding cues became more effective after three weeks. Moreover, the traditional teaching sequence seemed to be more powerful than the adjusted teaching sequence, for Group A outperformed two picture groups (Groups B and C) whereas Group D only outperformed Group C. The result that the two word encoding cue groups statistically performed better indicates that
- 62 -
word encoding cues benefited later retrieval. The answer to the second research question is consistent with the last delayed test.
Table 8.
ANOVA for Delayed Test 3
Sum of Squares df Mean Square F Sig.
Between Groups 18,290.389 3 6,096.796 7.853 .000**
Within Groups 52,019.470 67 776.410
Total 70,309.859 70
Notes: * p < .05, ** p < .01
Table 9.
Descriptive Data for Delayed Test 3
N M SD Std. Error
95% CI
Min Max Lower Bound Upper Bound
Group A 17 81.1 26.89850 6.52385 67.3465 95.0064 20.00 100.00 Group B 18 52.2 30.78494 7.25608 36.9132 67.5312 20.00 100.00 Group C 19 41.0 21.57538 4.94973 30.6536 51.4516 20.00 100.00 Group D 17 72.9 31.57624 7.65836 56.7062 89.1762 20.00 100.00 Total 71 61.1 31.69269 3.76123 53.6252 68.6283 20.00 100.00
- 63 -
Table 10.
Multiple Comparisons for Delayed Test 3
Scheffe
Mean Difference
(I-J) Std. Error Sig.
95% Confidence Interval
Lower Bound Upper Bound Group A Group B 28.95425* 9.42364 .031* 1.9284 55.9801 Notes: Group A: phonological encoding followed by L1 semantic encoding cues;
Group B: picture semantic encoding followed by phonological encoding cues; Group C: phonological encoding followed by picture semantic encoding cues; Group D: L1 semantic encoding followed by phonological encoding cues; * p < .05, ** p < .01
- 64 -
The results of Delayed Test 4
After excluding two groups of participants (those absent from or who left during the third learning session and those who did not concentrate in class), the number of participants in Delayed Test 4 was 55.
Table 11 reveals that there was still a significant difference on Delayed Test 4.
A 2 (encoding sequence: traditional vs. adjusted) × 2 (nature of semantic encoding cue: word vs. picture) analysis of variance (ANOVA) showed that there was a significant difference among groups, F(3, 51) = 3.559, p = .02 < .05. Table 12 presents the descriptive data for Delayed Test 4. The analysis of variance (Table 13) suggested that only Group D statistically outperformed Group C.
Surprisingly, Group A (traditional teaching sequence with word encoding cues), which statistically outperformed Groups B and C, did not outperform any groups on Test 4. Nor did Group A outperform any other groups on the last delayed test. In other words, its advantage disappeared in the last two delayed tests. This echoes Baddeley’s (2012) statement that phonological encoding is easy to build and its effect is rapid, but the information is also easily forgotten. Encouraging
phonological encoding seemed to facilitate students’ vocabulary learning at first, but the effect was not sustained.
- 65 -
The advantage of Group D (adjusted teaching sequence with word encoding cues) seemed to be more stable because that group statistically outperformed Group C on Delayed Test 3, Test 4, and the delayed test of all target words (also the last
delayed test). Group D’s advantage provides answers to the two research questions.
For the first research question, the adjusted teaching sequence successfully facilitated later retrieval. For the second research question, word encoding cues seemed to be a better choice for conducting the keyword method.
Table 11.
ANOVA for Delayed Test 4
Sum of Squares df Mean Square F Sig.
Between Groups 10,091.540 3 3,363.847 3.559 .020*
Within Groups 48,206.642 51 945.228
Total 58,298.182 54
Note: * p < .05
- 66 -
Table 12.
Descriptive Statistics for Delayed Test 4
N M SD Std. Error 95% CI Min Max
Lower Bound Upper Bound
Group A 14 51.4 35.70330 9.54211 30.8141 72.0430 20 100 Group B 18 54.4 32.03348 7.55036 38.5146 70.3743 20 100 Group C 13 27.6 15.35895 4.25981 18.4110 36.9736 20 60 Group D 10 68.0 35.52777 11.23487 42.5850 93.4150 20 100
Total 55 49.8 32.85721 4.43046 40.9356 58.7007 20 100 Notes: Group A: phonological encoding followed by L1 semantic encoding cues;
Group B: picture semantic encoding followed by phonological encoding cues; Group C: phonological encoding followed by picture semantic encoding cues; Group D: L1 semantic encoding followed by phonological encoding cues
Table 13.
Multiple Comparisons for Delayed Test 4 Scheffe
Lower Bound Upper Bound Group A Group B −3.01587 10.95577 .995 −34.6905 28.6588
- 67 -