| 摘要: | 電子評量可以使學習者透過即時反饋來診斷自己的問題。而在眾多的電子評量機制中，
雙層式測驗是一個有效的學習工具。這是因為雙層式測驗不僅可以用於了解學習者的學
習成效，而且還可以診斷出學習者是否真實了解答案或答案的原因是來自於猜測。雖然
雙層式測驗有上述的優點，但是它仍然有不足的地方，例如: 不能使學習者達到深度學
習。因此，無法使學習者保持長期的記憶。
為了解決這些問題，本研究提出三層式測驗。它不僅承襲雙層式測驗的答案和該題答案
的原因，而且還提供複習題幫助學習者達到反覆練習。通過這樣做，學習者可以達到有
效的學習。然而，三層式測驗仍屬於評量機制，因此，在進行測驗的過程中，學習者可
能會感到挫敗，進而缺乏學習動機。為此，本研究將遊戲式學習納入三層式測驗中。這
是因為遊戲式學習不僅可以提供愉快的體驗，還可以增加學習動機。更具體地說，在實
驗一中，開發了愉悅型三層式測驗。另一方面，個人化不僅可以提高學習者的學習動機，
還可以提升他們的參與度和學習成效。由於這些正面的影響，在實驗二中，藉著將個人
化容納進愉悅型三層式測驗，而開發了個人化愉悅型三層式測驗。
除此之外，不論是愉悅型三層式測驗或個人化愉悅型三層式測驗，仍然與大多數的電子
評量一樣會面臨許多不同的學習者。而學習者之間存在個體差異性，這可能會影響他們
的學習行為。在眾多個體差異中發現，先備知識具有影響性，因為先備知識可以反映學
習者在獲得新知識之前所具備的認知能力。有鑒於此，進行了兩個實證研究，以便能夠
ii
了解先備知識如何影響學習者使用愉悅型三層式測驗與個人化愉悅型三層式測驗，包含
測驗成效、任務成效、學習行為和遊戲行為。
實驗一為探討學習者如何藉著愉悅型三層式測驗來學習英語文法，而實驗二為探討學習
者如何藉著個人化愉悅型三層式測驗來學習英語文法。實驗一與實驗二的結果有一些相
似處，也有一些相異處。關於相似處，在測驗成效方面，高先備學習者都是優於低先備
學習者。而且，不論是高先備學習者或低先備學習者，他們的後測分數都是高於前測分
數。此外，高先備學習者與低先備學習者獲得相似的進步分數。另一方面，在學習行為
方面，學習者們都熱衷於使用各種鷹架輔助提示。而在遊戲行為方面，高先備學習者在
最後一回合時必定會獲勝，而低先備學習者在最後一回合時有可能會輸。
關於相異處，在實驗一中，高先備學習者在主要問題和主要問題的原因上表現比低先備
學習者更好。相反地，在實驗二的結果表明，高先備學習者只有在主要問題原因的表現
優於低先備學習者，但在主要問題時兩者的表現沒有差異。此外，實驗一的結果表明，
低先備學習者在複習題和複習題原因上表現比高先備學習者更好。然而，實驗二的結果
表明，低先備學習者只有在複習題的表現優於高先備，但在複習題原因時沒有差異。另
一方面，無論在實驗一或實驗二，高先備學習者和低先備學習者會使用不同的策略。更
具體地說，在實驗一使用的策略與他們的答案狀態有關，例如:答對和答錯。而在實驗二
使用的策略與他們的先備知識和題目類型有關，例如:複習題和複習題原因。除此之外，
實驗二結果表明，高先備學習者在進階題的表現是優於低先備學習者，而且他們在與個
人化愉悅型三層式測驗互動的過程中善用系統所提供的學習者半控制方法。
綜上所述，這項研究的兩個實驗有助於深入了解先備知識對使用三層式測驗的影響。 同
時，它們還指引了如何將個人化和遊戲式學習納入三層式測驗。這種貢獻的最終目標是
在進行電子評量的過程中，不同學習者的需求和偏好皆可以被滿足。;E-assessment can help learners diagnose their own problems with instant feedback during their
learning process. Among various E-assessment mechanisms, two-tier tests are effective
learning tools. This is because two-tiers can not only be used to assess learners’ learning
effectiveness, but also can be employed to diagnose whether learners really know answers or
such answers results from guessing. Although two-tier tests have the aforementioned
advantages, there are still some shortcomings. For example, it cannot make learners have a deep
impression. Therefore, learners cannot maintain long-term memory.
In order to solve these problems, this research proposes a three-tier test, which not only inherits
the answers and reasons from the two-tier test, but also provides review questions to help
learners achieve drill and practice. By doing so, the aforementioned problem can be sorted out.
However, the three-tier test still belongs to an assessment mechanism. Therefore, learners may
get frustrated and lack learning motivation during the assessment. Due to these disadvantages,
this study incorporates game-based learning into the three-tier test. This is because game-based
learning can not only provide joyful experience, but also increase learning motivation. Thus, an
entertaining three-tier test (ET3) was developed in Study One. On the other hand,
personalization can improve learners’ engagement and learning effectiveness. Due to these
positive influences, a personalized entertaining three-tier test (PET3), which combined
personalization and ET3, was developed in Study Two.
Like most E-assessment, either the ET3 or PET3 face diverse learners. This is because
individual differences exist between each learner. Among various individual differences, prior
knowledge is essential because it can reflect cognitive abilities that learners possess before
v
acquiring new knowledge. Owing to such essence, two empirical studies, i.e., Study One and
Study Two, were conducted to investigate the effects of prior knowledge in the context of the
ET3 and the PET3, including test performance, task performance, learning behavior and gaming
behavior.
Study One examined how learners learned English grammar via the ET3 while Study Two
investigated how learners learned English grammar via the PET3. Results from Study One and
those from Study Two shared some similarities: (1) High Prior Knowledge Learners (HPKL)
performed better than Low Prior Knowledge Learners (LPKL) in test performance; (2) their
post-test scores were higher than the pre-test scores, regardless of HPKL or LPKL; (3) HPKL
and LPKL obtained similar gain scores. (4);both HPKL and LPKL were keen to use various
scaffolding hints; and (5) HPKL must be a winner in the last round while LPKL could be a loser
in the last round.
There were also differences between Study One and Study Two. HPKL performed better than
LPKL in Study One, in terms of main questions and reasons of the main questions. On the
contrary, the results from Study Two showed that HPKL was better than LPKL in the reasons
of the main questions but no differences were found for the main questions. In addition, the
results from Study One showed that LPKL performed better than HPKL in review questions
and reasons of the review questions. However, the results from Study Two showed that LPKL
performed better than HPKL in review questions but no differences were found for the reasons
of the review questions. On the other hand, HPKL and LPKL used different strategies for Study
One and Study Two. More specifically, the strategies used for Study One were related to their
answer statuses, i.e., correct answers or wrong answers. On the other hand, the strategies used
for Study Two were associated with learners’ prior knowledge and the types of the questions
vi
(e.g., review questions or reasons of review questions). Furthermore, the results from Study
Two showed that HPKL performed better than LPKL in the advanced questions. Moreover, both
HPKL and LPKL appreciated the semi-control approach provided by the PET3.
In summary, these two empirical studies contribute to developing a deep understanding of the
effects of prior knowledge on the use of three-tier tests. Meanwhile, they also provide guidance
on how to incorporate personalization and game-based learning into three-tier tests. The
ultimate goal of such contributions is that the needs and preferences of diverse learners can be
accommodated when they interact with e-assessment. |
