智慧型電腦輔助專案式學習於量化研究之探討

Abstract

專案式學習是一種將學習圍繞在專案上的學習模型，並且已經被證明是一種有效的學習策略，然而，如何給予學生適當的輔助以幫助學生在專案中學習，是老師在進行專案式學習時常常會遇到的問題。有研究指出，電腦科技能有效幫助老師進行專案式學習，因此近年來，已經有一些電腦輔助專案式學習的研究被提出，其主要的目的就是當學生在做專案時提供適當的輔助，其中，提供鷹架幫助學生學習是電腦輔助專案式學習的研究方向之一。 社會科學研究是利用科學方法來研究人類行為和社會現象，而量化研究是社會科學研究中的一部分。要進行一個好的量化研究，除了基本量化研究相關的知識以外，經驗也是非常重要的。但是，大部分的量化研究學生缺少經驗，如此會導致他們浪費了很多不必要的時間和可能得到有問題的研究成果。而專案式學習能有助於學生學習如何應用所學到的量化研究知識來進行量化研究，許多教研究方法課程的老師，在課程的最後，也都會叫學生實際做一個量化研究的專案。由於在做專案的時候都會用到特定領域的知識，因此老師提供鷹架的方法會因為領域的不同而有很大的差異。目前較少有研究提供量化研究領域的電腦輔助。 依照我們的觀察，學生在進行量化研究專案時，會遇到下列四個問題：從已知的研究題目與假設選擇適當的研究方法、在蒐集的資料中發現可能的統計顯著性差異、選擇適當的統計方法驗證假設、及操作統計軟體。而專家的經驗對於克服這些問題是非常有幫助的。 因此，在本研究中，我們提出了一個與知識表格類似的意圖發現方法，幫助學生能夠選擇適當的研究方法；使用資料倉儲幫助學生觀察資料，並且在其上設計專門用來發現統計顯著性差異的指標，幫助學生發現統計顯著性差異；利用規則式專家系統利用專家的經驗來幫助學生找到適當的統計方法驗證假設；提出SPSS模擬器產生機制來幫助評量學生SPSS軟體的學習成效。我們進行了一些實驗來評估學生對本研究的滿意度，從這些實驗的研究結果來看，我們可以發現大部分的學生都對本研究覺得滿意。

Project-Based Learning (PBL), which is a model that organizes learning around projects, has proven to be an effective learning strategy. Support of student learning is one of the teachers’ enactment problems of PBL. Many researches have claimed that computer technology is effective for PBL. In recent years, some Computer Assisted Project-Based Learning (CAPBL) studies are proposed to provide appropriate supports for students doing the projects. Providing scaffolding for apprentices to help apprentices do the project is one of the research directions in CAPBL. Social science research is the use of scientific methods to investigate human behavior and social phenomenon, and quantitative research (QtR) is a part of social science research if the collected data are numerical. To perform a good QtR, not only basic QtR knowledge, e.g. explicit knowledge, but also experiences, e.g. tacit knowledge, are important. However, most of Quantitative Research Apprentices (QtRAs) lack the experiences of performing quantitative research, which could lead to them wasting a lot of time and getting problematic research results. PBL is helpful for QtRAs to learn the application of QtR and many teachers teaching research methods ask students to do a final project. The ways of providing scaffolding for vary domains are usually different because domain specific knowledge is used in the process of doing projects. Few studies provide computer assisted supports on QtR currently. In our observation, there exist four issues when QtRAs perform QtR projects: hard to select appropriate research methods for given research problems and hypotheses, hard to discover statistically significant differences from data sets, hard to choose appropriate statistical methods to test hypotheses, and hard to operate data analysis softwares. The experts’ experiences are useful to help QtRAs overcome these issues. In this study, an intention finding approach which is similar to repertory grid is proposed to assist QtRAs selecting appropriate research methods. Data warehousing and new indicators are used to help QtRAs explore the data sets and find the statistically significant differences. Rule-based expert system is further used to help QtRAs choose appropriate statistical methods to test their hypotheses. Furthermore, to help QtRAs assess their operating skills on statistical software, SPSS Simulator Generator Scheme is proposed to build some SPSS Simulators. To evaluate the satisfactions of this study, some experiments have been done. According to the experiment results, we can conclude that most of the QtRAs are satisfied with this study.