標題: 影音回溯資料初探─以設計思考為研究面向
Video/audio protocols for study of design thinking: A macroscopic examination
作者: 鄭乃文
Nai-Wen Cheng
劉育東
Yu-Tung Liu
應用藝術研究所
關鍵字: 設計思考;方法學;口語資料分析;影音回溯資料;知覺;design thinking;methodology;protocol analysis;video/audio protocol;perception
公開日期: 1998
摘要: 原案分析為設計思考學門最重要的研究方法。其中的資料收集階段,則以同步口語報告和回溯法最為常見。同步口語報告可以提取多量且細部的資訊,卻有影響正常設計行為、妨礙知覺動作的缺點;回溯法則相反,雖然對設計進程的影響較小,但侷限於短期記憶,僅能揭露量少而粗略的設計資訊。 影音回溯報告是針對同步報告和回溯的缺點而提出的改良式回溯法。建立在回溯的基礎上,以提供設計當時的影像資料增加記憶線索,降低選擇性回憶的出現。為探索「反映過程的設計活動」,建議口語報告的他法。 本研究以量的分析探討影音回溯報告的資料性質,並初步嘗試對圖面資料的編碼與斷句。研究問題集中於檢驗影音回溯資料收集有效性資料及擷取知覺相關資訊的能力。由四位受測者參與實證研究的結果發現,影音回溯法確實可能收集到更大量的報告,但同時也夾雜更多的非有效資料於其中。另外,影音回溯法對於「有效性資料」的擷取能力,與放聲思考法相當;對於知覺相關資訊的收取,影音回溯資料也無法反映收集資料程序中不影響知覺的優勢,結果並不特出。 研究的結論提示了我們幾個後續發展的可能;一、繼續探索影音回溯資料的「質」方面特性,例如資料的細緻程度等,補強對影音回溯資料驗證性的探討。二、在相關知覺的研究上,擴大甚或改變編碼對象,以其他非口語報告的資料作為分析主體,以彌補目前對知覺資訊收集的無力。
Protocol analysis has been a major methodology for research of design thinking. As for the methods of data-collection, both concurrent verbalization and retrospection have their own weaknesses. The demand of concurrent verbalizing would affect not only the normal design behaviors but also the perceptual interactions between designers and their own sketches; while limited by STM, retrospective condition may probably induce some selective retrievals of memory trace, resulting in a lower accuracy of reports. By providing the images/sounds of designers' exact sequence of sketching, a refined method of retrospection is taken to reduce selective recall. The visual cues are expected to help designers remember what they really thought, and video/audio retrospective reports might thus become an alternative measure of exploring the "reflective design activities". This research aims at examining the validity of video/audio retrospective (VA) protocols, based on the quantitative comparison of VA reports and the currently used method of verbal reporting- think aloud (TA) protocols. Moreover, the segmentation and a coding scheme for the drawings are primitively developed. We finally find that, VA reports can indeed retrieve a larger amount of information, but accompany with more invalid data. And, between VA and TA reports, there's no notable difference in terms of the ability of collecting valid information, even of collecting perception-related data. From the conclusions, two possibilities of future works emerge: First, the qualitative analysis of VA protocols should be conducted further. Second, for exploring the perception issue, it's necessary to expand the analyzed entities, e.g. designers' movements or the drawings, to the whole scheme, and the verbal data might be used as a secondary source or as a reference to the other.
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT870509001
http://hdl.handle.net/11536/64871
Appears in Collections:Thesis