認知風格、人口統計屬性、任務績效及情感體驗：一項關於天文物理資料系統核心使用者的實證研究

Cognitive Styles, Demographic Attributes, Task Performance, and Affective Experiences: An Empirical Investigation into Astrophysics Data System (ADS) Core Users

美國國家航空暨太空總署（NASA）與史密松天文台（SAO）同建的天文物理資料系統（ADS）是天文物理學者重點使用的數位資料庫。ADS的視覺介面包括作者引文網路、文章網路、文字雲及引文計量。本研究由20個ADS核心使用者參與可用性測試，測探認知風格與個人背景對任務績效和情感體驗的影響。研究結果揭示參與者多屬評析及圖像思維風格、過半的參與者自評為圖像思維者。評析傾向及自認圖像思維者，對於任務完成效率有統計學意義的影響。年齡及之前使用經驗對任務執行功效有顯著影響。圖像思維指數於困惑次數有負向關係，而自認圖像思維者比非自認參與者較少體驗到沮喪和猶豫。本研究對深度探討科學家使用者的思維型態和情感體驗有啟發，而對促進ADS視覺介面的改建有幫助。

As a primary digital library portal for astrophysics researchers, SAO/NASA ADS (Astrophysics Data System) 2.0 interface features several visualization tools such as Author Network and Metrics. This research study involves 20 ADS long term users who participated in a usability and eye tracking research session. Participants first completed a cognitive test, and then performed five tasks in ADS 2.0 where they explored its multiple visualization tools. Results show that over half of the participants were Imagers and half of the participants were Analytic. Cognitive styles were found to have significant impacts on several efficiency-based measures. Analytic-oriented participants were observed to spent shorter time on web pages and apps, made fewer web page changes than less-Analytic-driving participants in performing common tasks, whereas AI (Analytic-Imagery) participants also completed their five tasks faster than non-AI participants. Meanwhile, self-identified Imagery participants were found to be more efficient in their task completion through multiple measures including total time on task, number of mouse clicks, and number of query revisions made. Imagery scores were negatively associated with frequency of confusion and the observed counts of being surprised. Compared to those who did not claimed to be a visual person, self-identified Imagery participants were observed to have significantly less frequency in frustration and hesitation during their task performance. Both demographic variables and past user experiences were found to correlate with task performance; query revision also correlated with multiple time-based measurements. Considered as an indicator of efficiency, query revisions were found to correlate negatively with the rate of complete with ease, and positively with several time-based efficiency measures, rate of complete with some difficulty, and the frequency of frustration. These results provide rich insights into the cognitive styles of ADS’ core users, the impact of such styles and demographic attributes on their task performance their affective and cognitive experiences, and their interaction behaviors while using the visualization component of ADS 2.0, and would subsequently contribute to the design of bibliographic retrieval systems for scientists.