4.1 The CAF Constructs
4.1.2 Lexical complexity
4.1.2.2 Lexical variation
4.1.2.2.2 Comparative changes in lexical variation between the low and high-
4.1.2.2.2.1 Low-proficiency groups.
Table 18 shows ten statistically significant findings of the ANOVA test, revealing which of the low-proficiency groups’ lexical variation improved or deteriorated the most in the two composition tasks. As presented in the table, in the classification task, six metrics (NDWERZ, NDWESZ, MSTTR, Uber, VV1, and ModV) indicated statistically significant scores in the short term: NDWERZ increased for all three groups with OLC1 (M = 3.207) increasing the most, followed by BLC1 (M = 1.213) and, lastly, TLC1 (M = 0.160). The p- value was < .031. NDWESZ developed the most for OLC1 (M = 2.673), followed by BLC1 (M = 2.420), while it declined for TLC1 (M = -0.220); the p-value was < .039. In addition, MSTTR improved for OLC1 (M = 0.049), followed closely by BLC1 (M = 0.045), whereas it decreased for TLC1 (M = -0.003). The p-value was < .028. Uber increased the most for BLC1 (M = 2.235), followed by OLC1 (M = 2.176), but it deteriorated for TLC1 (M = -1.136). The p-value was < .049. VV1 developed for only BLC1 (M = 0.133); in fact, it declined in both TLC1 (M = -0.091) and OLC1 (M = -0.002). The p-value was < .009. Finally, while ModV decreased in both OLC1 (M = -0.034) and, with an insignificant difference, TLC1 (M = - 0.033), it improved for BLC1 (M = 0.013). The p-value was < .049. The above-mentioned results show that BLC1’s lexical variation increased in six metrics (three of which were the most), whereas the sub-construct of TLC1 deteriorated in five metrics. These findings reveal that in the short term the BLC was the most successful, but the TLC was the least successful, in developing lexical variation in the classification task.
In the argumentative task, four measures (NDW, VV1, SVV1, and CVV1) indicated statistically significant scores in the long term: NDW showed that while TLC1 (M = -16.333) declined the most, followed by OLC1 (M = -4.400), BLC1 improved (M = 15.133); the p- value was < .002. VV1 decreased the most for OLC1 (M =-0.065), followed by BLC1 (M = - 0.006), whereas it increased for TLC1 (M = 0.117); the p-value was < .015. The SVV1 measure deteriorated the most for OLC1 (M = -2.065), followed by TLC1 (M = -1.353), but it developed for BLC1 (M =2.338); the p-value was < .020. Similarly, CVV1 declined the most for OLC1 (M = -0.226), followed by TLC1 (M = -0.142), while it improved for BLC1 (M = 0.272); the p-value was < .017. Thus, of four lexical variation measures, BLC1 was the only group to increase in three of them, whereas OLC1 solely decreased in all of them. This means that in the argumentative task, the BLC was the most effective, but the OLC was the least
175 effective, in developing lexical variation in the long term. All of these results are graphically presented in Figure 8.
Table 18
TLC1, BLC1, and OLC1 Lexical Variation Metrics: Significant Results of the ANOVA Test
Writing Tasks Classification Task
Test Comparisons Pre to Mid
Metrics (NDWERZ) (NDWESZ)
Groups TLC1 BLC1 OLC1 Total TLC1 BLC1 OLC1 Total
Mean 0.160 1.213 3.207* 1.527 -0.220 2.420 2.673* 1.624
SD 3.105 3.099 3.027 3.267 2.872 3.473 3.548 3.495
P-Value 0.031 0.039
Metrics (MSTTR) (Uber)
Groups TLC1 BLC1 OLC1 Total TLC1 BLC1 OLC1 Total
Mean -0.003 0.045 0.049* 0.030 -1.136 2.235* 2.176 1.092
SD 0.047 0.060 0.064 0.061 4.560 3.682 4.150 4.353
P-Value 0.028 0.049
Metrics (VV1) (ModV)
Groups TLC1 BLC1 OLC1 Total TLC1 BLC1 OLC1 Total
Mean -0.091* 0.133 -0.002 0.014 -0.033 0.013 -0.034* -0.018
SD 0.123 0.215 0.214 0.207 0.066 0.047 0.058 0.060
P-Value 0.009 0.049
Writing Tasks Argumentative Task
Test Comparisons Pre to Post
Metrics (NDW) (VV1)
Groups TLC1 BLC1 OLC1 Total TLC1 BLC1 OLC1 Total
Mean -16.333* 15.133 -4.400 -1.867 0.117 -0.006 -0.065* 0.015
SD 27.699 21.387 19.29 26.079 0.145 0.195 0.152 0.179
P-Value 0.002 0.015
Metrics (SVV1) (CVV1)
Groups TLC1 BLC1 OLC1 Total TLC1 BLC1 OLC1 Total
Mean -1.353 2.338 -2.065* -0.360 -0.142 0.272 -0.226* -0.032
SD 5.337 4.524 3.119 4.741 0.584 0.517 0.325 0.525
P-Value 0.020 0.017
176 To summarize, in the classification task, the BLC was the most beneficial in enhancing the lexical variation among the low-proficiency Saudi EFL participants in the short term, while the TLC was the least beneficial. The reason is that the lexical variation of BLC learners improved in six metrics, whereas it increased in one metric and deteriorated in five others for the TLC learners. Besides, in the argumentative task, the former learning
environment (BLC) was again the most successful in promoting its participants’ lexical variation in the long term, but the OLC was the least successful. The justification is that the BLC group developed in three of four lexical variation metrics, while the OLC group declined in all of them.
4.1.2.2.2.2 High-proficiency groups.
Table 19 displays nine statistically significant findings of the ANOVA test, revealing which of the three high-proficiency group’s lexical variation improved or decreased the most in the two writing tasks. As indicated in the table, in the classification task, five measures (TTR, CTTR, RTTR, LogTTR, and Uber) showed statistically significant scores in the long term: TTR measure deteriorated the most for TLC6 (M = -0.059), then BLC6 (M = -0.005), whereas it increased for OLC6 (M = 0.034). The p-value was < .030. In terms of CTTR,
-20.000 -15.000 -10.000 -5.000 0.000 5.000 10.000 15.000 20.000 T L C 1 B L C 1 O L C 1
Classification Task (Pre to Mid: NDWERZ) Classification Task (Pre to Mid: NDWESZ)
Classification Task (Pre to Mid: MSTTR) Classification Task (Pre to Mid: UBER)
Classification Task (Pre to Mid: VV1) Classification Task (Pre to Mid: MODV)
Argumentative Task (Pre to Post: NDW) Argumentative Task (Pre to Post: VV1)
Argumentative Task (Pre to Post: SVV1) Argumentative Task (Pre to Post: CVV1)
177 TLC6 was the only group to decline (M = -0.549); BLC6 (M = 0.081) developed the most, followed by OLC6 group (M = 0.065). The p-value was < .009. In both RTTR and LogTTR, although BLC6 improved the most (M = 0.115 and 1.110, respectively), followed by OLC6 (M = 0.093 and 0.010, respectively), TLC6 decreased (M = -0.778 and -0.022, respectively). The p-values were < .009 and .011, respectively. Uber also deteriorated for TLC6 (M = - 2.812); nonetheless, it increased the most for OLC6 (M = 0.792), followed by BLC6 (M = 0.171). The p-value was < .004. In all of these results, despite BLC6 developing the most in three of five measures and declining in one measure, OLC6 improved, but TLC6 decreased, in all measures. This reveals that the OLC was the most effective, while the TLC was the least effective, in increasing lexical variation in the classification task in the long term.
In the argumentative task, four metrics (ModV, NDW, SVV1, and CVV1) indicated statistically significant scores in the short term or the long term. In the short term, ModV deteriorated the most for BLC6 (M = -0.038), followed by OLC6 (M = -0.012), and
developed for TLC6 (M = 0.012). The p-value was < .041. This shows that the TLC was the most beneficial, whereas the BLC was the least beneficial, in improving lexical variation in the argumentative task in the short term. In contrast, in the long term, NDW declined the most for TLC6 (M = -18.000), followed by BLC6 (M = -3.700), but it increased for OLC6 (M = 4.200). The p-value was < .005. Similarly, SVV1 developed for OLC6 (M = 1.406), while it decreased the most for TLC6 (M = -2.982), followed by BLC6 (M = -2.870). The p-value was
< .041. Moreover, OLC6 was the only group to improve in CVV1 (M = 0.167), whereas
TLC6 (M = -0.305) and BLC6 (M = -0.296) deteriorated the most. The p-value was < .036. In all of the above findings, OLC6 was the only group to increase in three metrics of lexical variation, but both TLC6 and BLC6 declined in all of such metrics. The decreases of TLC6 were the most. This means that the OLC was the most successful, while the TLC was the least successful, in developing lexical variation in the argumentative tasks in the long term. The above-mentioned results are graphically shown in Figure 9.
Table 19
TLC6, BLC6, and OLC6 Lexical Variation Metrics: Significant Results of the ANOVA Test
Writing Tasks Classification Task
Test Comparisons Pre to Post
Metrics (TTR) (CTTR)
Groups TLC6 BLC6 OLC6 Total TLC6 BLC6 OLC6 Total
Mean -0.059* -0.005 0.034 -0.010 -0.549 0.081* 0.065 -0.134
SD 0.087 0.080 0.048 0.081 0.448 0.508 0.475 0.549
P-Value 0.030 0.009
178
Writing Tasks Classification Task
Test Comparison Pre to Post
Metrics (RTTR) (LogTTR)
Groups TLC6 BLC6 OLC6 Total TLC6 BLC6 OLC6 Total
Mean -0.778 0.115* 0.093 -0.190 -0.022 1.110* 0.010 -0.004
SD 0.633 0.719 0.671 0.777 0.025 0.023 0.019 0.026
P-Value 0.009 0.011
Writing Tasks Classification Task Argumentative Task
Test Comparisons Pre to Post Pre to Mid
Metrics (Uber) (ModV)
Groups TLC6 BLC6 OLC6 Total TLC6 BLC6 OLC6 Total
Mean -2.812 0.171 0.792* -0.616 0.012 -0.038* -0.012 -0.013
SD 3.148 2.017 1.638 2.781 0.051 0.040 0.032 0.045
P-Value 0.004 0.041
Writing Tasks Argumentative Task
Test Comparisons Pre to Post
Metrics (NDW) (SVV1)
Groups TLC6 BLC6 OLC6 Total TLC6 BLC6 OLC6 Total
Mean -18.000* -3.700 4.200 -5.833 -2.982* -2.870 1.406 -1.482
SD 11.926 16.111 13.105 16.289 3.760 4.444 4.290 4.532
P-Value 0.005 0.041
Metrics (CVV1)
Groups TLC6 BLC6 OLC6 Total
Mean -0.305* -0.296 0.167 -0.145
SD 0.382 0.447 0.487 0.481
P-Value 0.036
(*) indicates the most increase or decrease
Briefly, in the classification, the OLC was the most effective in enhancing the lexical variation of the high-proficiency Saudi EFL undergraduates in the long term, whereas the TLC
-20.000 -15.000 -10.000 -5.000 0.000 5.000 10.000 T L C 6 B L C 6 O L C 6
Classification Task (Pre to Post: TTR) Classification Task (Pre to Post: CTTR)
Classification Task (Pre to Post: RTTR) Classification Task (Pre to Post: LOGTTR)
Classification Task (Pre to Post: UBER) Argumentative Task (Pre to Mid: MODV)
Argumentative Task (Pre to Post: NDW) Argumentative Task (Pre to Post: SVV1)
Argumentative Task (Pre to Post: CVV1)
179 was the least effective. The reason is that the OLC group improved in all five lexical variation measures (one of which the most), but the TLC group deteriorated in all of them. In the argumentative task, the TLC was the most beneficial in promoting its participants’ lexical variation in the short term, while the BLC was the least beneficial. Conversely, in the long term, the OLC was the most successful in developing its learners’ lexical variation in the long term, whereas the TLC was the least successful. The justification is that the OLC group increased in three measures of lexical variation, but those in TLC and BLC declined. The decreases in such measures, however, were the most for TLC.