健康与遗传心理学研究室知识库

简版抑郁一焦虑一压力量表(depression-anxiety-stress-21, DASS-21)是由Lovibond andLovibond (1995)所编制的一种用于评价负性情绪的量表。.早期通过成人的研究发现，该量表可区分为抑郁、焦虑和压力三个因子，且在成人中的信效度较高。近期有人将其施测于青少年，有些表明原始的三因子模型拟合得很好，而有些表明尚有更好的拟合模型存在。国内尚缺乏针对青少年的研究结果。本研究在以往研究的基础上，分别考察了该量表在我国普通青少年和地震后有情绪问题的青少年中的适用性。

(1) 研究一首先考察了量表中经典的三因子模型在我国普通青少年中的适用性。通过探索性因子分析，表明DASS-21中各条目分数与抑郁分量表得分的相关系数为0.385-0.809;与焦虑分量表得分的相关系数为0.400-780;与压力分量表得分的相关系数为0.335-0.826，均P<0.01。DASS-21抑郁、焦虑、压力各分量表内在一致性倍度系数分别为0.801, 0.823 , 0.826;总量表的内在一致性系数为0.939，DASS量表中抑郁、焦虑、压力分量表之间的Pearson相关系数为0.781-0.787，与量表总分的相关系数为0.920-0.930。但是，验证性因子分析的结果表明， x²/df=2.417， P<0.001;CFI=0.885, IFI=0.888、 TLI=0.858,、RMSEA=0.082 ,表明经典的三因子模型对我国普通青少年的数据拟合得不是很好。进一步通过探索性因素分析提取出四个因子，建立四因子模型，结果表明，四因子模型与我国的普通青少年样本拟合得很好， x²/df=2.167, P<0.001; CFI=0.907, IFI=0.909、TLI =0.883、RMSEA = 0.074。

(2)研究二考察了该量表经典的三因子模型在我国地震后有情绪问题的请少年中的实用性，结果表明，DASS-21中各条目与抑郁分量表得分的相关系数为0.306-0.724;与焦虑分量表得分的相关系数为0.2540.719; 与压力分量表得分的相关系数为0.366-0.742，均P<0.01。 DASS-21抑郁、焦虑、压力各分量表内在一致性信度系数Cronbach α系数分别为0.793 , 0.781, 0.778;总量表的内在一致性系数为0.911。DASS量表中抑郁、焦虑、压力分量表之间的Pearson相关系数为0.714-0.740，与量表总分的相关系数为0.895-0.914;以创伤后应激障碍(PTSD )得分为效标，DASS抑郁、焦虑、压力各分量表与PTSD得分的Praeson相关系数为0.626-0.741，DASS总分与PTSD得分相关系数为0.774(均P<0.01)。验证性因素分析结果表明，x²/df =4.180，P<0.001; CFI=0.904、 IFI=0.905、 TLI=0.881,RMSEA = 0.06，各指标均符合测量学标准。

(3)研究三在研究二的基础上，分别考察了心理弹性和亲子关系质量与地震后青少年焦虑、抑郁和压力情绪之间的关系。线性回归分析结果表明，父子关系质量、坚韧性、力量性和乐观性是焦虑情绪和压力情绪的显著预测，变量，且父子关系质量、坚韧性和乐观性对青少年焦虑和压力情绪的预测作用是负性的，而力量性对青少年焦虑和压力情绪的预测作用是正性的;父子关系质量、母子关系质量、坚韧性、力量性和乐观性是抑郁情绪的显著预测变量，且父子关系质量、母子关系质量、坚韧性和乐观性对青少年抑郁情绪的预测作用是负性的，而力量性对青少年抑郁情绪的预测作用是正性的。结果说明，亲子关系质量可在一定程上缓解青少年的负性情绪。而心理弹性的不同成分对负性情绪的作用是不同的。研究还发现，亲子关系质量和心理弹性对男生和女生的三种负性情绪的预测作用是不同的。对男生而言，父子关系质量、坚韧性、力量型和乐观型对焦虑情绪的预测作用是显著的，对抑郁情绪的预测作用也是显著的;但是，只有坚韧性对压力情绪的预测作用是显著的。对女生而言，母子关系质量和乐观性对焦虑情绪的预测作用是显著的;母子关系质量、坚韧性、乐观性和力量性对抑郁情绪的预测作用是显著的;母子关系质量、坚韧性和乐观性对压力情绪的预测作用是显著的。这说明，对不同性别的青少年而言，不同的亲子关系质量和心理弹性的不同成分对不同负性情绪的缓和作用是不同的。父子关系对男生应对负性情绪的影响更大，而母子关系对女生应对负性情绪的影响更大。

The short version of Depression Anxiety Stress Scales(DASS-21) is a widely used measure of negative affect compiled by Lovibond and Lovibond(1995). Early studies in adult samples have found that there could be extracted three factors from this measurement,i.e. depression, anxiety and stress. And the reliability and validity was high. Some research tested its structure in adolescent sample, but the results were inconsistent. While some research found the three-factor model fit the sample well, other studies indicated there existed better models. Such study involved adolescents was scarce in China. This study investigated the applicability of DASS-21 in Chinese normal adolescents and those adolescents with emotional problems after earthquake.

(1) The first study investigated the fitness of the classic three-factor model in Chinese normal adolescents. Using exploratory factor analysis, we found that the correlation coefficient of each item score with the depression sub-scale score was 0.385-0.809， with the anxiety sub-scale score was 0.400-0.780, with the stress sub-scale score was 0.385-0.809. All of these correlation coefficients were significant， P<0.01. The inner reliability coefficients for three sub-scales （depression/anxiety/stress) were 0.801，0.823 and 0.826 respectively. And the reliability coefficient for the total scale was 0.939. The Pearson correlation coefficient of the three sub一scale scores was 0.781-0.787 and the correlation coefficient of the three sub-scale scores with the total score was 0.920-0.930. However, the confirmatory factor analysis revealed that the fitness of the classic three-factor model with our sample was not so good, x²/df一2.417, P<0.001; CFI = 0.885, IFI=0.888, TLI=0.858 and RMSEA = 0.082. We extracted four factors through further exploratory factor analysis. And confirmatory factor analysis revealed that this four-factor modal fitted our sample vary well, x²/df=2.167, P<0.001, CFI= 0.907, IFI=0.909, TLI=0.883 and RMSEA=0.074.

(2) The second study tested the applicability of DASS-21 in Chinese adolescents with emotional problems after earthquake. We found that the correlation coefficient of each item score with the depression sub-scale score was 0.306-0.724， with the anxiety sub-scale score was 0.254-0.719， with the stress sub-scale score was 0.366-0.742. All of these correlation coefficients were significant9 P<0.01. The inner reliability coefficients for three sub-scales(depression/anxiety/stress) were 0.793，0.781 and 0.778 respectively. And the reliability coefficient for the total scale was 0.911.The Pearson correlation coefficient of the three sub-scale scores was 0.714-0.740, and the correlation coefficient of the three sub-scale scores with the total score was 0.895-0.914. And confirmatory factor analysis revealed that the three-factor model fitted the sample well, x²/df=4.180, P<0.001;CFI=0.904, IFI=0.905, TLI=0.881 and RMSEA=0.06.

(3)The third study measured adolescents' anxiety, depression and stress emotion with DASS-21。and investigated the relationship of resilience, quality of parent-child relationship and these three negative emotions. By linear regression analysis, we found that the quality of father-child relationship, robustness, power and optimism were significant predicting variables for emotions of anxiety and stress, and the predicting effects of quality of father-child relationship, robustness, and optimism to these two emotions were negative, while the predicting effect of power to these two emotions was positive. The quality of father-child relationship, mother-child relationship, robustness, power and optimism were significant predicting variables for depression emotion, and the predicting effects of quality of father-child relationship, mother-child relationship, robustness, and optimism to depression were negative, while the predicting effect of power to depression was positive. These results indicated that the quality of parent-child relationship could relieve adolescents' negative emotion to some extent, and different components of resilience play different role for negative emotions.