| In the last decade,a wealth of research has demonstrated that taking a test could produce better learning outcomes relative to restudying,and this phenomenon is termed as the testing effect.Although great achievements have been made by studies on the testing effect,there remain many disputes and research gaps due to its late start.In terms of learning strategy,extant studies on the testing effect focus mainly on the advantage of testing over restudying,but whether testing could always lead to better performance than elaborative studying still demands further research.Moreover,relatively few studies have explored the effectiveness of combining testing and elaborative studying strategy together.As for the complexity of learning materials,whether the testing effect is always evident in complex learning materials remains a matter of debate.Additionally,the mechanisms underlying the testing effect are still unclear.According to the cognitive load theory,learners only have limited cognitive resources,so intrinsic cognitive load(ICL),extraneous cognitive load(ECL)and germane cognitive load(GCL)will go up and down when completing a certain task,thereby affecting the learning outcomes.Besides,different learning strategies and text complexity will lead to different types of cognitive load.Recently,some researchers have made a preliminary exploration on the testing effect based on the cognitive load theory.However,the amount of such research is extremely limited and these studies are merely confined to their first language.In order to further explore the boundary conditions and mechanisms of L2 testing effect,this study explores how learning strategy and text complexity affect L2 discourse comprehension and cognitive load as well as their correlations.The main research questions are as follows:1.How does learning strategy affect L2 discourse testing effect and cognitive load?2.How does text complexity affect the effects of learning strategy on L2 discourse testing effect and cognitive load?3.Is there any correlation between L2 discourse comprehension and cognitive load under different learning strategies with different text complexities?In this study,participants were 139 freshmen of English majors from Jiangnan University,and they were randomly divided into four groups: restudy group(RS),self-explanation group(SE),free recall group(FR)and free recall combined with self-explanation group(FR+SE).In the experiment,all participants needed to learn two texts with different complexities,and the order of text learning was disrupted.All groups first read a text of any complexity and then entered the intervening phase.The RS group needed to reread previous text.The SE group was required to explain the text with examples from their own lives,studies and works while rereading previous text.The FR group should recall previous text as much as possible in the absence of previous text.The FR+SE group also needed to recall previous text and explained their recalled information with examples from their own lives,studies and works.Meanwhile,each group had to undertake a secondary color observation task while completing previously described main task,which was used to assess the total amount of cognitive load(TCL)of participants.After this phase,all participants completed a subjective reported scale of cognitive load to measure their ICL,ECL and GCL.Then after an interference task(15min),participants of all groups learned a text with another complexity,and all procedures were the same as the first text learning.One week later,all participants returned to take the final test.The major findings were shown as follows:1.Learning strategy had a significant influence on L2 discourse comprehension.The SE group achieved a superior performance on inference questions relative to the RS group.But the two testing groups(FR group and FR+SE group)significantly outperformed the SE and RS group in both verbatim questions and inference questions,and the FR+SE group which combined testing with self-explanation strategy further enhanced performance on inference questions.In addition,learning strategy had no influence on ICL,but it had a significant influence on ECL,GCL and TCL.The RS group was the lowest no matter in ECL,GCL or TCL.The SE group reported the highest ECL and TCL.Meanwhile,ECL,GCL and TCL of the FR+SE group were all significantly higher than those of the FR group.2.Text complexity did not hinder the generation of the testing effect.The two testing groups(FR group and FR+SE group)both significantly outperformed the RS group regardless of final test format and text complexity.When learning the simple text,no significant difference was observed in the two testing groups and the SE group in their performance on verbatim questions,but the former two performed significantly better than the latter on the inference questions.When learning the complex text,the two testing groups also obtained a significantly better performance than the SE group on verbatim questions,while the FR+SE group obtained the best performance on inference questions.Under simple text learning condition,ECL and TCL of the other three groups did not differ significantly but they were all higher than these of the RS group,whereas GCL of the two testing groups were both significantly higher than the RS and SE group.Under complex text learning condition,the SE group reported the highest ECL and TCL whereas the RS group was the lowest.In terms of GCL,the FR+SE group was the highest whereas the RS group was the lowest.3.ICL was generally negatively correlated with L2 discourse comprehension especially when test format was represented by verbatim questions.ECL was found negatively correlated with L2 discourse comprehension only in the SE group.Similarly,only a negative correlation was found between TCL and L2 discourse comprehension in the FR+SE group.In contrast,it was found that GCL was positively correlated with L2 discourse comprehension.This study found that the two testing conditions were more conducive to enhancing memory and transfer performance on previously learned text,demonstrating a robust testing effect and highlighting the advantages of combining self-explanation and testing.Furthermore,text complexity did not affect the testing effect,which may be accounted by low motivation and insufficient investment of cognitive resources resulted by restudying,while self-explanation strategy caused too high ECL especially in learning complex text with high element interactivity and thus leading to insufficient GCL.In addition,the study found that too high or too low ECL and TCL both did no good to learning,whereas GCL was positively correlated with learning.Therefore,in the process of L2 text teaching and learning,it is suggested to adopt a learning strategy that is conducive to schema construction,especially when the text is relatively complex.Learners should combine their own examples to generate some explanations while recalling freely on previously learned information,which allow learners to activate those existing schemas in long-term memory and construct new schema,thus promoting a deeper understanding on texts. |
