大學生消沉傾向量表之發展研究

The Development of Self-affliction Tendency Scal

本研究主要依據Church 與Brooks（2009）有關隱微自殺（subtle suicide）概念編製「消沉傾向量 表」。消沉傾向量表共有30 題，在信效度考驗上共進行內部一致性信度、再測信度、建構效度、 複核效化及效標關聯效度等考驗。以大學生為研究對象，研究樣本來自全台10 所大專校院在學 學生，總計包含2，679 位大學生。研究結果發現，消沉傾向量表具有良好信效度，在多次內部一 致性信度考驗，全量表信度係數介於 .94- .96，各分量表介於 .76- .96。各分量表組合信度皆達.70 以上，平均變異萃取量均在 .30 以上，題項間具有適度的共同變異，SMC（Squared Multiple Correlations）佳。間隔2 週測量穩定係數為 .89，間隔3 週為 .72。信度考驗結果顯示消沉傾向 量表具備良好一致性和穩定性。以結構方程模式統計方法進行驗證性因素分析，結果支持消沉傾 向量表評量的因素結構與實徵資料間的適配度佳。複核效化分析結果，顯示建模樣本與效度樣本 的指標相似度高，支持消沉傾向量表評量的跨樣本穩定性。在消沉傾向與相關概念間關係，研究 結果均支持消沉傾向量表評量具備合理的聚斂效度與區分效度。依據研究結果，本研究提出消沉 傾向量表運用與發展之建議。

The purpose of the study was to develop the Self-affliction Tendency Scale (SATS) for college students. Based on phenomenon observed and literature reviewed， the authors proposed a framework for studying the concept of “self-affliction” as a tendency of self-destruction in various aspects ranging from cognitive， emotional， behavioral， to interpersonal relationships. The SATS is comprised of 30 items in 7-point Likert-type scale and can be completed between 5 to 10 minutes. It included four dimensions: existential alienation， dysfunction， poor self-care， and contradictory relationship. The values of reliability and validity on the SATS were calculated from a sample of 2，679， all of which were Taiwanese undergraduate college students from ten universities. Accordingly， the internal consistency reliability (N = 172~1，050)， test-retest reliability (N = 66 and 98)， construct validity (N = 470)， cross-validation (N = 519)， and criterion-related validity of the developed scale (N = 172~1，050) were evaluated. Seven tests were conducted to relate the 4-dimension model to various psychological issues in the Taiwanese society. These issues included self-esteem， mental health， internet addiction， loneliness， depression， and happiness. Descriptive statistics， including means， variances， and item-total correlations， were computed for each item. The Pearson product-moment correlation coefficient， one-factor repeated measures analysis of variance， and discriminant analysis were conducted to differentiate the similarity and difference between variables. Internal consistency of the scores on SES and its subscales was estimated by Cronbach’s coefficient alpha (full scale was .94~.96， subscale of existential alienation was .81~.86， subscale of dysfunction was .93~.96， subscale of poor self-care was .76~.84， and subscale of contradictory relationship was .82~.90). Stability coefficient was based on 2 test-retest procedures (an interval of 2 weeks was .89， and an interval of 3 weeks was .72). Validation of the SATS was based on a 4-dimension model， followed by cross-validation on the calibration sample (N = 470) and validation sample (N = 519). Confirmatory factor analysis indicated that the model was well supported by observed data and the stability of the model was confirmed. The result had supported the scale structure. These evidences thus generally supported the 4-dimension self-affliction model. Applications of the SATS and the suggestions of future study were discussed.