DSSI-23量表在农村老年人中信度及效度评价

Reliability and validity of the 23-Duke Social Support Scale among rural elderly population

目的评价23条目杜克社会支持量表（DSSI-23）在农村老年人中的信度和效度。方法采用分层随机整群抽样方法在四川省自贡市和广元市农村地区抽取1021名≥60岁且无严重听力及精神障碍的老年人进行问卷调查；应用分半信度和Cronbach’s α系数评价量表的信度，应用会聚效度、区别效度、结构效度和判别效度评估量表的效度。结果DSSI-23量表的分半信度系数为0．910，量表总体Cronbach’s α系数为0．873；社会交往、主观社会支持、社会支持利用3个维度的Cronbach’s α系数分别为0．459、0．855、0．888，依次删除3个维度中某一维度条目后总的Cronbach’s α系数分别为0．891、0．803、0．808，均〉0．8，量表具有较好的信度；各条目与总分间的相关系数为0．192—0．690（均P〈0．01），会聚效度与区别效度定标实验成功率均为100％，量表会聚效度和区别效度较好；对量表进行验证性因子分析，各拟合指标为：近似误差均方根（RMSEA）为0．082、拟合优度指数（GFI）为0．837、规范拟合指数（NFI）为0．807、比较拟合指数（CFI）为0．827、增值拟合指数（IFI）为0．828；不同特征农村老年人DSSI-23量表各维度得分及量表总分比较，不同年龄农村老年人社会交往和主观社会支持维度得分均不同，不同文化程度、婚姻状况农村老年人社会交往、主观社会支持、社会支持利用维度得分和量表总分均不同，差异均有统计学意义（均P〈0．05），量表具有较好的判别效度。结论DSSI-23量表用于中国农村老年人社会支持研究的信度较好，会聚效度、区别效度和判别效度较好，但结构效度较差，提示一些条目需要进一步改进。

Objective To assess the reliability and validity of 23-Item Duke Social Support Index （ DSSI-23 ） scale among an elderly population in rural areas. Methods Stratified random cluster sampling was employed to select 1 021 elderly people aged ≥60 years and without severe hearing disorders or mental problems from rural areas in Zigong and Guangyuan municipality of Sichuan province and the participants were surveyed face-to-face with DSSI-23 scale. Reliability of DSSI-23 was evaluated by split-half reliability and Cronbach＇s α coefficient. Validity of DSSI-23 was evaluated by convergent, discriminant, and structure validity. Results The split-half reliability coefficient was 0. 910 and the overall Cronbach＇s oL coefficient was 0. 873. The Cronbach＇s α coefficient for domains of social interaction, subjective social support, utilization of social support were 0. 459,0. 855, and 0. 888, respectively;after deleting items for one of the 3 domains successively,the overall Cronbach＇s a were 0. 891,0. 803 ,and 0. 808（ all 〉0. 8）. The DSSI-23 scale has good reliability. The correlation coefficient of each item with the total scale（ except the item 3 ） was 0. 35 - 0.69 （ P 〈 0.01 ）. The DSSI-23 scale has good convergent and discriminant validity, with the success rate of calibration experiment of 100% for both convergent and discriminant validity. The structure validity was tested by confirmatory factor analysis and the fit- ness indexes were 0. 082 for root mean square error of approximation（RMSEA） ,0. 837 for goodness-of-fit index（ GFI）, 0. 807 for normed fit index （NFI） ,0. 827 for comparative fit index（ CFI）, and 0. 828 for incremental fit index （IFI）, respectively. Significant differences were observed in the scores for social interaction and subjective social support among the elderly at various ages and in the scores for social interaction, subjective social support, utilization of social support, and the whole scale among the elderly with different education level and marital status （ all P 〈 0. 05 ）, suggesting that DSSI-23 scale has good discriminant validity. Conclusion The DSSI-23 scale has good reliability, convergent validity, and discriminant validity but dissatisfactory structure validity when measuring social support among elderly population in rural area, suggesting some items of the scale need to be revised.