人機互動中的異步式心理評估

人機互動中，使用者的內在心理變項（例如：認知能力、人格特質、情緒狀態，或是興趣偏好）能幫助社會機器人提供更為適切的服務。但心理變項常內隱於使用者的外在行為中（例如：面部表情、聲音語調），僅能從第三方觀察者的角度推論得知其內部意涵；然而，對於不具推論能力的機器人來說，心理變項的真實值卻仍是未知之數。早期研究中，亦有嘗試使機器人執行心理評估，蒐集使用者第一人稱的回答以推論內在心理變項，但生硬呈現完整問卷內容，會令使用者產生防衛與感到疲倦。為使具有良好信度與效度的心理測驗能應用於人機互動中，方便機器人取得內在心理變項，本研究提出異步測驗問題法（Asynchronous Test Questions, ATQs）。將原紙本問卷題目，盡可能參照指導語，與依照人機互動情境進行編寫，以求維持原心理測驗的信度與效度，再將ATQs分散至不同的人機互動環節中。此外，為保ATQs可應用至多數使用者，本研究同時招募年長者族群與年輕者族群作為受試者。再者，受試者ATQs口語回答形式眾多，因此會經由兩名資料編錄者的編錄與協調，將口語回答轉換為李克氏五點量尺（Krippendorff‘s alpha：年輕者組 = 0.995；年長者 = 0.87），助於後續資料分析。本研究選用大五人格短版問卷作為概念性驗證的目標，採用相依t檢定比較受試者大五人格的ATQs口語回答與紙本問卷回答之間差異；並透過皮爾森相關驗證兩方式的一致姓。結果顯示，兩方法在年輕者族群，無論是大五人格的各子構念又或是單題表現，皆具有高度正相關（介於0.77-0.88之間）且其回答統計上無差異；但在年長者族群中，各子構念大多僅有中度正相關（介於.61-.77之間），其中開放性（Openness to experience）的相關更僅有0.18；單題表現上，「緊張的、不安的」該題回答在統計上具有顯著差異，細節將在文中討論。總體來說，本研究驗證ATQs確能與原紙本問卷有相似效果，並提出實際應用於人機互動的方式，例如：情境式改寫；與應用至不同使用者群體之注意事項，例如：題目長度須考量年長者認知處理能力。

關鍵字

以人為本的機器人；人機互動；心理評估

並列摘要

Social robots, interactive agents that keep company with people, currently lack appropriate tools to collect data with ease to predict human behaviors. Human latent psychological variables are valuable information that can effectively predict human behaviors and be evaluated by well-established psychological tests. However, explicit and excessive testing might make users feel tedious and defensive. Previous studies have attempted to employ robot-administered psychological tests during human-robot interaction (HRI), most of which implemented tests consecutively and without appropriate adaptation based on the context of conversation. Suppose we could embed psychological tests into human-robot conversations naturally through various social tactics and the long-term company of social robots. In that case, psychological tests could be administered in a more user-friendly and more unconscious way. The present study proposed Asynchronous Test Questions (ATQs), which were adapted from psychological tests to fit conversation scenarios and would be temporally distributed into HRI. The Ten Items Personality Inventory was used to examine the possibility of ATQs by comparing verbal responses to ATQs with writing responses to the paper-based questionnaire. The comparison shows no statistically significant difference between the two approaches and had strong positive correlations between them in young adults. In contrast, in older adults, there were merely moderate positive correlations between the two approaches, and the inconsistency was shown in terms of Openness and Emotion Stability. In conclusion, the present study successfully examined the feasibility of integrating the asynchronous method into psychological tests. Some guidelines and caveats are provided when this approach is applied to older adults or certain populations.

並列關鍵字

Human-centered robotics ； Human-Robot Interaction ； psychological evaluation

參考文獻

Abdi, J., Al-Hindawi, A., Ng, T., Vizcaychipi, M. P. (2018). Scoping review on the use of socially assistive robot technology in elderly care. BMJ Open, 8. https://doi.org/10.1136/bmjopen-2017-018815

Google Scholar

Ahmad, M., Mubin, O., Orlando, J. (2017). A systematic review of adaptivity in human–robot interaction. Multimodal Technologies and Interaction, 1. https://doi.org/10.3390/mti1030014

Google Scholar

Armstrong, B., Gallant, S. N., Li, L., Patel, K., Wong, B. I. (2017). Stereotype threat effects on older adults’ episodic and working memory: A meta-analysis. Gerontologist, 57. https://doi.org/10.1093/geront/gnx056

Google Scholar

Bolarinwa, O. A. (2015). Principles and methods of validity and reliability testing of questionnaires used in social and health science researches. Nigerian Postgraduate Medical Journal, 22. https://doi.org/10.4103/1117-1936.173959

Google Scholar

Bopp, K. L., Verhaeghen, P. (2005). Aging and verbal memory span: A meta-analysis. The Journals of Gerontology: Series B, 60. https://doi.org/10.1093/geronb/60.5.p223

Google Scholar

國際替代計量

人機互動中的異步式心理評估

全文下載

主題瀏覽