中国科学院心理健康重点实验室知识库

目的 评价中文版多元心理健康素养量表在中国男性军人中的信度和效度，为该量表在中国的推广使用提供参考依据。方法 于2018 年11 — 12 月采用整群随机抽样方法抽取驻地在北京、天津、河北和山东的1 106 名男性军人进行中文版多元心理健康素养量表测试，间隔6 周后选取其中237 人对量表进行重测；采用项目分析、内部一致性信度、重测信度、结构效度、聚敛效度和区分效度进行量表的信度和效度评价。结果 中文版多元心理健康素养量表在原量表基础上修订了2 个条目，并删除了同质性较差的2 个条目和载荷绝对值较小的2 个条目后共包括22 个条目；项目分析结果显示，中文版多元心理健康素养量表的22 个条目得分与量表总分的相关系数为0.20～0.48（均P < 0.01），量表具有较好的项目区分度；量表总体Cronbach′s α 系数和重测信度分别为0.80 和0.64，知识、信念和资源3 个维度的Cronbach′s α 系数分别为0.76、0.71 和0.77，量表具有较好的信度；探索性因子分析结果显示，中文版多元心理健康素养量表3 个公因子的累计方差贡献率为38.59 %，各条目所属因子结构与原量表基本一致；验证性因素分析结果显示，中文版多元心理健康素养量表的拟合优度指数（ GFI）为0.91、残差均方根（RMR）为0.01、平均概似平方误根系数（RMSEA）为0.06、调整拟合优度指数（AGFI）为0.90，模型拟合较好；重测样本的多元心理健康素养总分、知识维度得分、信念维度得分、资源维度得分和心理健康知识问卷总分分别为（13.00 ±4.05）、（5.45 ± 2.49）、（4.76 ± 2.07）、（2.78 ± 1.29）和（14.82 ± 2.50）分，多元心理健康素养量表总分及各维度得分与心理健康知识问卷总分均呈正相关（均P < 0.01）；初测样本的多元心理健康素养总分、知识维度得分、信念维度得分、资源维度得分、抑郁量表得分和焦虑量表得分分别为（ 11.47 ± 4.35） 、（ 5.09 ± 2.52） 、（ 3.95 ± 2.08） 、（ 2.43 ±1.43）、（3.21 ± 3.82）和（2.65 ± 2.99）分，多元心理健康素养量表总分及信念和资源维度得分与抑郁和焦虑量表得分均呈负相关（均P < 0.01）；量表的聚敛效度和区分效度均较好。结论 中文版多元心理健康素养量表具有较好的信度和效度，可作为中国军人的心理健康素养评估工具。

Objective To evaluate the reliability and validity of the multicomponent mental health literacy measure-Chinese version in male military personnel and to provide evidences for the application of the measure in China. Methods Using random cluster sampling, we recruited 1 175 male military personnel in two cities (Beijing and Tianjin) and two provinces (Hebei and Shandong) province for a mobile phone-based self-administered survey and a resurvey in 237 persons selected from 1 106 valid respondents six weeks after the initial survey during November — December 2018. The modified multicomponent mental health literacy measure-Chinese version was adopted in the study and the reliability and validity of the instrument were assessed with item analysis, internal consistency coefficient, test-retest reliability, construct validity, convergent validity and discriminant validity. Results Two items in the original measure were revised and 4 items were deleted (2 due to poor homogeneity and 2 due to small absolute value of load) and the final measure included totally 22 items. Project analysis resulted in the correlation coefficients between the scores of each items and the total score of the measure ranging from 0.20 to 0.48 (all P < 0.01), indicating a good item differentiation of the measure. The overall Cronbach′s α and test-retest reliability of the measure was 0.80 and 0.64; and the Cronbach′s α of the measure′s three domains of knowledge, belief and resources was 0.76, 0.71 and 0.77, respectively, indicating a good reliability. Exploratory factor analysis revealed

that the cumulative variance contribution rate was 38.59% for the three domains of the measure and the item factor structure was consistent with that of the original measure. Confirmatory factor analysis showed a good fitness of the constructed model, with the goodness-of-fit index (GFI) of 0.91, root mean square residual (RMR) of 0.01, root mean square error approximation (RMSEA) of 0.06, and adjusted goodness-of-fit index (AGFI) of 0.90. For the respondents of the test-retest survey, the average scores of mental health literacy, knowledge, belief, and resources domain, and mental health knowledge questionnaire were 13.00 ± 4.05, 5.45 ± 2.49, 4.76 ± 2.07, 2.78 ± 1.29, and 14.82 ± 2.50, respectively; the overall score and scores of each domain of the mental health literacy were significantly correlated positively with mental health knowledge questionnaire scores (P < 0.01 for all). For the respondents of the initial survey, the average scores of mental health literacy, knowledge, belief, and resources domain, the Center for Epidemiological Studies Depression Scale, and Generalized Anxiety Disorder Scale was 11.47 ± 4.35, 5.09 ± 2.52, 3.95 ± 2.08, 2.43 ± 1.43, 3.21 ± 3.82, and 2.65 ± 2.99, respectively; the overall score of mental health literacy and the scores of mental health literacy domains were significantly correlated inversely with depression and anxiety scores (all P < 0.01) . The results indicated good convergent validity and discriminant validity of the measure. Conclusion The modified multicomponent mental health literacy measure-Chinese version is of good reliability and validity when administered in male military personnel and could be used as an evaluation instrument of mental health literacy for Chinese military personnel.