Hooper 視覺組織測驗於台灣樣本的題目分析

視知覺組織能力與人類日常活動息息相關，但往往因為疾病或是外傷而使此能力受損。Hooper 視覺組織測驗( Hooper Visual Organization Test; HVOT )在臨床上常被用於篩檢神經缺損患者的視知覺功能，具有良好的信效度，但目前尚未有研究探討它是否適用於台灣文化。因此，本研究主要目的為( 1 )分析 HVOT的信度；( 2 )探討HVOT與教育程度、年齡及性別的關係以及( 3 )使用試題反應理論中的羅序模式( Rasch Model )，檢驗 HVOT 的題目難度與適合度。本研究依照年齡、性別、及教育程度等三個變項，採分層隨機取樣法在南台灣地區抽樣而得，全樣本共計 360 位，年齡為20歲至79歲的健康台灣成人，以個別施測的方式評估每位受試者的視覺組織能力。結果顯示，年齡與教育程度組在 HVOT 的得分有顯著效果 ( F11, 348 = 29.53; p < 0.0001; F4, 355= 88.33; p < 0.0001 )。不過，不同性別在 HVOT 的表現並無顯著差異。HVOT 的信度方面，奇偶數折半信度為 0.81( p < 0.0001 )，Cronbach’s α 為0.87，顯示 HVOT 有不錯的內部一致性。以試題適合度來看，第 7 題（狗）被視為與其它試題有相關，考慮試題重複性可刪除，第1（魚）、3（桌子）、和19（泡茶壺）題與測量視覺組織能力的關聯性低。以試題難度來看，第 7 題（狗）與第 3 題（桌子）這兩題太簡單，其餘試題皆在標準範圍。 HVOT 的題目編排並沒有依照由簡單到困難的順序。上述研究結果可做為未來發展中文化 HVOT 之參考，也建議未來持續檢驗 HVOT 的再測信度和效度，以及建立全國性常模。

關鍵字

Hooper視覺組織測驗；羅序模式；信度；試題反應理論

並列摘要

Perceptual organization deficit is common after disease or trauma that contributes to the impaired functional performance in daily activities. As a result, an early screening and treatment is essential for an improved prognosis. The Hooper Visual Organization Test (HVOT) is a widely used measure designed to evaluate visual perceptual ability in the patients with neurological impairment. The appropriateness of the HVOT for use in other cultures has been examined, incluing Greek and German. However, its measurement properties have not yet been investigated for the Taiwanese people. Thus, the purposes of this study are threefold: (1) to determine the internal consistency of the HVOT; (2) to explore the relationships between demographic parameters (age, education, and gender) and HVOT performance; and (3) to validate the HVOT by modern psychometric analysis, sepcifically by examining fit of the scale’s data to the Rasch measurement model. A purposive sampling strategy was employed in order to ensure that cognitive performances were represented from healthy individuals of differing age, gender and education in southern Taiwan. A total of 360 people between the ages of 20 and 70 were recruited from various sources. The HVOT was administered individually to each participant. The results showed that the internal consistency for the HVOT was good (split-half reliability = 0.81, Cronbach’s α = 0.87). There were significant main effects for age (F11, 348 = 29.5257; p < 0.0001) and education (F4, 355= 88.3327; p < 0.0001). Fit statistics indictaed that 26 of 30 (86.7%) items showed good fit to the Rasch model. The HVOT items were rearranged to be in hierarchical order. Our results provided a valuable reference for the development of the Chinese version of the HVOT. Future research should aim at investigating test-retest reliability and validity, as well as establishing local norms for the HVOT.