In the following section, the first research question is addressed and the effects of the two different kinds of teaching sequence are discussed based on the results of the delayed tests and the overall observations mentioned in the previous sections. First, the effects of the different teaching sequence are compared. Second, the comparison of the two encoding cues is analyzed.
The two kinds of teaching sequence are the traditional teaching sequence and the adjusted teaching sequence. For the traditional teaching sequence, a target word was linked to its keyword, followed by its definition. The traditional teaching sequence is in accordance with the sequence proposed by Atkinson (1975). For the adjusted teaching sequence, a target word was linked to its definition, followed by its keyword. The adjustment was made to avoid the negative phonological similarity effect. The comparison concerns whether the avoidance of the phonological similarity effect facilitates the participants’ performance.
- 75 -
The adjusted teaching sequence
Both the adjusted teaching groups (Groups B and D) performed better than the traditional teaching groups at the end of the session. Though Group B did not
statistically perform better, the participants’ performance continued to improve throughout the last three weeks. The results indicate that Group D (adjusted teaching sequence with word encoding cues) performed significantly better throughout the last
three delayed tests. In other words, the avoidance of the phonological similarity effect successfully facilitated the participants’ performance three weeks after the first
learning session. This is in accordance with Campoy and Baddeley’s (2008) statement that by encouraging semantic encoding prior to phonological encoding, the
information could be more robust. This also accords with Savill et al.’s (2016) finding that newly acquired words could be remembered longer when the semantic link was built prior to the phonological link. The advantage of Group D supported the
conclusion that by encouraging semantic encoding prior to phonological encoding, the information could be maintained longer.
However, the other group taught under the adjusted sequence, Group B, did not perform significantly better. The difference between Groups B and C was that Group B received picture encoding cues. Though the participants in Group B
performed worse than those in Group A on Test 3, the disadvantage was not observed
- 76 -
with the following tests. Based on the interview, the inferior performance might be due to the participants’ recoding skill, as mentioned above. In other words, the reason that the participants in Group B did not perform significantly better might be because of their recoding ability.
Nevertheless, none of the participants in Group B stated they had difficulty connecting the keywords to the L1 definitions. During the interview, Student M2 stated as long as he could link the target word to its keyword, he could successfully retrieve the definition. Every time he failed to recall the definition correctly, it was because he forgot the keywords. Student L2 also agreed that the keywords helped him recall the meaning. Their feedback contradicts the claim of Wang and Thomas (1992) that the keywords provided by the researcher would become a hindrance over time to recalling the definitions.
The participants in Group D also expressed the view that the keywords helped them retrieve the meanings. Student H4 stated that she never saw any images in her mind when taking the tests but simply linked the target word to its keyword,
whereupon the definition would appear in her mind. Hence, she thought forming mental images unnecessary since she could retrieve the definitions simply through the acoustic link. From her feedback, it could be observed that the acoustically similar keywords did not cause any hindrance but rather helped her retrieve the definitions,
- 77 -
perhaps because the semantic encoding was encouraged prior to the phonological encoding. Therefore, the phonological similarity effect was inhibited, which helped the participants in the adjusted sequence groups recall the definitions successfully.
In sum, the participants taught under the adjusted sequence did not support the claim that keywords would become a hindrance. On the contrary, they believed the keywords helped them remember the definitions. Wang and Thomas’ (1992) claim was refuted in the current study when the participants were taught under the adjusted sequence. The hindrance was not caused by the keywords but the phonological similarity effect, which can be prevented by the adjusted teaching sequence. Hence, the adjusted teaching sequence facilitates later retrieval of vocabulary.
The traditional teaching sequence
The other two groups, Groups A and C, were taught under the traditional sequence. Though with Group A (traditional teaching sequence with word encoding cues) it seemed to be more powerful on Test 3, its effect vanished rapidly: The participants in Group A did not outperform the other participants on Test 4 nor on the test of all target words. In other words, linking the target word to its keyword
(phonological link) seemed to be more powerful at first, but this effect did not persist.
This echoes what Baddeley (2012) claimed: phonological coding is easy to build and the effect is rapid, but phonological coding (in short-term memory) is also readily
- 78 -
forgotten. In other words, though encouraging phonological encoding prior to semantic encoding may be effective at first, its effect does not persist.
The other group also taught in the traditional sequence was Group C. Group C’s inferior performance revealed that the picture encoding with traditional teaching
sequence was the least effective. The participants in Group C started to perform significantly worse on Test 3, and this disadvantage persisted throughout the last three delayed tests. This result contradicts Jarrold and Citroën’s (2013) conclusion that the phonological similarity effect was prevented by presenting the phonologically similar words with pictures. However, they also indicated that though it seemed the
phonological similarity effect could be prevented, this was in fact due to the floor effect.
In the current study, the participants’ performance was not facilitated by presenting acoustically similar words with pictures. That is to say, the results of this study contradict the results of Dolean’s (2014) study that the effectiveness of presenting keywords with pictures persisted for one month. In addition, the inferior performance was not due to the encoding cue because Group B, which also received the picture encoding cue but was taught in the adjusted sequence, did not perform significantly worse than Group D. In other words, the inferior performance was
- 79 -
mainly due to the negative phonological similarity effect, which was not inhibited by the keyword pictures.
The phonological similarity effect proposed by Conrad and Hull (1964) was observed with Groups A and C. During the interview, Students M1 and L1 stated that they considered the keywords a hindrance. Both of them had trouble connecting the keywords to the L1 definitions. Student M3 also complained that sometimes he just could not link the keywords to the L1 definitions. In their feedback the negative
phonological similarity effect can be clearly observed.
The participants’ performance reveals that the phonological similarity effect
can only be prevented when the teaching sequence is adjusted. Moreover, the adjusted teaching sequence with word encoding cues may facilitate vocabulary learning over time.