性格評估篩檢量表─中文版信效度之建立：以台灣社區成人樣本為例

Psychometric Properties of the Chinese Version of Personality Assessment Screener in an Adult Community Sample in Taiwan

鄭筱儒(Xiao-Ru Cheng)

指導教授 ： 陳怡群

國立中正大學/社會科學院/臨床心理學研究所/碩士(2014年)

摘要

目的：本研究目的在於檢驗「性格評估篩檢量表」（Personality Assessment Screener, PAS）中文版之信效度並建立其切截分數，並將此樣本在PAS中文版上的表現，與林珂睿 (2010)的台灣大學生樣本以及Morey (1997)的美國社區和大學生樣本進行比較。方法：選取台灣北、中、南、東四大地區共327名台灣社區成人（18至59歲）來參與本研究，其填寫一包含PAS中文版並以隨機方式排序的套裝量表，並於一週後再次填寫相同排序的所有量表。結果：首先，項目分析結果發現PAS中文版中的試題多數適用於評估此量表欲測量之建構，但在敵意控制、飲酒問題、衝動行為等三次量尺中各有一題試題仍有待進一步的修正。其次，在信效度檢驗部分，本研究顯示PAS中文版全量表具有適當的再測信度（r = .81）（各次量尺介於 .57至 .81）、內部一致性則稍弱（Cronbach’s α = .66）（各次量尺介於 -.08至 .78），但若刪除不適當題項後則可提升至適當範圍 （Cronbach’s α = .72）；而在聚合效度部份，PAS中文版全量表總分與貝克焦慮量表（Beck Anxiety Inventory, BAI）（r = .44）、貝克憂鬱量表第二版（Beck Depression Inventory Second Version, BDI-II）（r = .50）及症狀檢核表修訂版（Symptom Check List-90-R, SCL-90-R）（r = .53）總分皆有顯著的正相關，且各次量尺亦與上述測驗工具中用以評估相似建構之次量尺分數有適當地正相關；於區辨效度部分，獨立樣本t檢定顯示PAS中文版的全量表總分和次量尺皆能適當區辨是否為憂鬱症狀、焦慮症狀、身心症狀的高風險族群；而預測效度方面，階層迴歸分析（hierarchical regression）也指出PAS中文版中的負向情感次量尺最能預測整體身心症狀；而ROC曲線（Receiver Operating Characteristic Analysis）顯示當PAS中文版全量表總分之切截分數訂在19分時，可對篩檢社區中高風險心理病理症狀之成人，具有良好的鑑別度（AUC = .82），敏感度為 .76，特異度為 .79。然而，本研究結果亦發現，與其他次量尺相較下，PAS中文版中敵意控制、飲酒問題、社交退縮、衝動行為等四個次量尺在信效度上的表現不佳。接著，單一樣本t檢定結果發現本研究樣本和美國社區樣本相較下，兩群體在PAS全量表總分上並無顯著差異，但在部分次量尺上有顯著差異，此差異與將台灣和美國大學生樣本相較之結果相似，即台灣樣本在社交退縮和疏遠次量尺上，皆高於美國樣本，反之，在衝動行為、敵意控制、飲酒問題上皆低於美國樣本。另外，台灣與美國的大學生樣本，兩者在全量表總分無差異，但兩者之全量表總分皆高於美國或台灣的社區樣本，而在次量尺部分，台灣與美國大學生樣本的得分雖部分有差異，但在負向情感、自殺想法、衝動行為三次量尺皆一致地高於社區樣本。結論：本研究初步檢驗結果支持PAS中文版為一具信效度、可用於篩檢本研究台灣社區成人心理病理性格特質之性格篩檢工具，唯敵意控制、飲酒問題、衝動行為三次量尺中的各一題試題及社交退縮次量尺中的兩題試題較不適用於台灣社區成人，建議在使用時對此四次量尺的解釋應予以保留。此外，本研究亦發現，PAS中的各個次量尺的模組呈現不同文化和不同年齡間的差異。
關鍵字：性格、篩檢量表、成人、心理計量特質

關鍵字

成人 ； 篩檢量表 ； 性格 ； 心理計量特質

並列摘要

Purpose：The purpose of the present study was twofold: (1) to examine the psychometric properties of Chinese version of Personality Assessment Screener (CPAS), and establishing the cutting-off score and (2) to compare the performance on PAS between adults and college students in Taiwan and in the United States. Method: A sample of 327 Taiwanese adults, ranging in age from 18 to 59, in community settings were recruited from northern, middle, southern, and eastern geographic regions of Taiwan. A series of measures including of CPAS were administered in a random order to the participants, and were administered the same package again to the participants after 7 days. One-sample t tests were conducted to examine the mean differences between the sample of American adults and the samples of American college students (Morey, 1997) and Taiwanese college students (林珂睿，2010). With regard to the psychometric properties of CAPS, analyses for internal consistency and test-retest reliabilities and for convergent and discriminant validities were computed. In addition, we also determined the cut-off scores of the CPAS for the Taiwanese adults. Results: First, the results of item analyses showed that three items which belong to Hostile Control, Alcohol Problem, and Acting Out subscale, respectively, should be revised. Other items are appropriate to assess the construct of interest for the CPAS. Second, the CPAS had an acceptable 7-days test-retest reliability for the entire scale (r = .81) and for subscales (ranged from r = .51 to r = .81). The original entire scale of CPAS showed lower internal consistency (Cronbach’s α = .66), but the Cronbach’s α values could be improved to .72 after deleting the items with poor corrected-item-total correlations. In terms of validity, the correlational analyses revealed that the entire CPAS had acceptable convergent validity with Beck Anxiety Inventory (BAI) (r = .44), Beck Depression Inventory (BDI-II) (r = .50), and Symptom Check List-90-R (SCL-90-R) (r = .53). With regard to discriminant validity, independent t test showed that the total score and subscale scores of CPAS could differentiate a low-risk group from a high-risk group with depressive, anxious, or psychosomatic symptoms. In terms of predictive validity, hierarchical regression analyses showed that Negative Affect was the most optimal predictor of the severity of psychosomatic symptoms. Furthermore, ROC curve analyses showed that a cut-off score of 19, could provide satisfactory discrimination in the prediction of psychological symptoms (as the sensitivity is .76 and the specificity is .79). However, compared to other subscales, four subscales (i.e., Hostile Control, Alcohol Problem, Social Withdrawal, and Acting Out) had poor psychometric properties. Third, one sample t tests showed that no differences were found on the total scores of the CPAS between Taiwanese and American community adults; however, some of the subscales (i.e., Social Withdrawal, Alienation, Acting Out, Hostile Control, Alcohol Problem, and Anger Control) were significantly different. The difference in the personality profile of the CPAS between Taiwanese and American adult community samples was similar to that between Taiwanese and American college samples. This is Taiwanese adult community and college samples both had higher scores on Social Withdrawal and Alienation subscales of the CPAS but lower scores on the Acting Out , Hostile Control, and Alcohol Problem subscales of the CPAS than American adult community and college samples. Discussion: The preliminary results of the present study showed that the CPAS appears to be a valid tool in the assessment of psychopathological personality in the Taiwanese adult community sample, except four CPAS subscales (i.e., Hostile Control, Alcohol Problem, Social Withdrawal, and Acting Out) of CPAS which should be interpreted with caution. In addition, the present study found cultural and age differences in the PAS personality profiles.
Keywords: personality, screening tool, adults, psychometric properties

並列關鍵字

adults ； screening tool ； personality ； psychometric properties

參考文獻

車先蕙、盧孟良、陳錫中、張尚文、李宇宙 (2006)。中文版貝克焦慮量表之信效度。台灣醫學，10(4)，447-454。

李明濱、廖士程 (2006)。自殺原因與防治策略。臺灣醫學，10(3)，367-375。

林淑惠、黃韞臻 (2009)。臺灣中部地區大學生獨處能力、生活壓力與身心健康之相關研究。新竹教育大學教育學報，26(1)，33-62。

翁嘉英、林宜美、呂碧鴻、陳秀蓉、吳英璋、鄭逸如 (2008)。「短式華人敵意量表」之發展與信效度考驗。測驗學刊，55(3)，463-487。

郭千哲、陳為堅、鄭泰安 (1999)。CAGE酗酒篩檢問卷在台灣基層醫療院所的跨文化效度測試。中華公共衛生雜誌，18(2)，87-94。