基于CTT和IRT的遗忘频率-10量表的汉化及信效度检验

Chinesization and reliability and validity test of Frequency of Forgetting⁃ 10 Scale based on classical test theory and item response theory

目的:以经典测量理论和项目反应理论为基础,对英文版遗忘频率-10量表(FoF-10)进行汉化,并通过传统测量方法和Rasch模型分析验证其信效度。方法:采用方便抽样法选取2021年6月—2022年3月南京市某三级甲等医院记忆门诊的280例病人作为研究对象,运用Brislin翻译模式对FoF-10进行汉化,形成中文版FoF-10(CHI-FoF-10)。运用经典测量理论分析方法和Rasch分析方法评价量表信效度。结果:CHI-FoF-10条目水平的内容效度指数为0.830~1.000,量表水平的内容效度指数为0.915;以一般自我效能感量表作为CHI-FoF-10的效标量表,其效标效度的Pearson相关系数为0.848;CHI-FoF-10预测病人记忆自我效能的受试者工作特征曲线下面积为0.869[95%CI(0.809,0.929)],当截断值为37.5分时,约登指数最大(0.721);探索性因子分析提取1个特征值>1的公因子,累积方差贡献率为65.590%,各条目在该维度上的载荷均>0.40。CHI-FoF-10重测信度为0.954;Cronbach's α系数为0.941。数据-模型拟合度分析结果显示,CHI-FoF-10的项目信度为0.99,样本信度为0.93;项目分离指数为8.90,个人分离指数为3.57;怀特图显示,被试者能力呈正态分布,分布跨度约为6个Logits。结论:CHI-FoF-10具有良好的信效度,可以为临床病人记忆自我效能感的测量提供参考。

Objective:Based on the classical test theory and item response theory,English version of Frequency of Forgetting-10 Scale(FoF-10)was translated into Chinese,and its reliability and validity were verified by traditional measurement methods and Rasch model analysis.Methods:A total of 280 patients from memory clinic of a tertiary grade A hospital in Nanjing was selected by convenient sampling method as the research subjects from June 2021 to March 2022.The FoF-10 was translated into Chinese by Brislin translation model,and Chinese Frequency of Forgetting-10 Scale(CHI-FoF-10)was formed.The reliability and validity of the scale was evaluated by classical test theory analysis methods and Rasch analysis methods.Results:The item-level content validity index of CH-FoF-10 was from 0.830 to 1.000,and scale-content validity index of CHI-FoF-10 was 0.915.Taking General Self-Efficacy Scale as benchmark for CHI-FoF-10,the Pearson correlation coefficient of its criterion validity was 0.848.The area under receiver operating characteristic curve of CHI-FoF-10 for predicting patient memory self-efficacy was 0.869[95%CI(0.809,0.929)].When the cut-off value was 37.5 points,the Youden index was the largest(0.721).Exploratory factor analysis extracted a common factor with feature value>1,with a cumulative variance contribution rate of 65.590%.The load of each item in this dimension was>0.40.The retest reliability of CHI-FoF-10 was 0.954,the Cronbach'sαcoefficient was 0.941.The data model fit analysis results showed that the item reliability of CHI-FoF-10 was 0.99,and the person reliability was 0.93.The item separation index was 8.90,and the person separation index was 3.57.Wright map showed that the ability of the subjects was normally distributed,and the distribution span was about 6 Logits.Conclusions:CHI-FoF-10 has good reliability and validity,and can provide reference for measuring memory sel-efficacy of clinical patients.