关键词: 孤独, 老年人, 农村人口, De Jong Gierveld孤独量表, 信度, 效度

Abstract: Background　Loneliness has become a common psychological problem in the elderly in China，a population group increasing year by year.Early screening of loneliness can improve the level of mental health and quality of life in the elderly.The UCLA Loneliness Scale is a widely used measurement tool，but its items are many，and the measurement process is time-consuming.In contrast，De Jong Gierveld Loneliness Scale（DJGLS），a scale specially designed for the elderly，has less entries，is time-saving and efficient.Objective　To test the reliability and validity of the Chinese version of DJGLS among rural elderly people in China.Methods　The DJGLS was Chinesized following Brislin's classic model for translation and validation of instruments guidelines for cross-cultural research，including translation，and back-translation，done independently by two translators.Then it was tested in 515 rural elderly people selected from 6 villages（Hongqi，Shangliyuan，Xialiyuan，Chagou，Lifang，Huangjia），Anshan，Liaoning Province between March and June，2018 by use of convenience sampling.General personal characteristics were collected by a self-designed Demographic Characteristic Questionnaire at the same time.This round of survey achieved a response rate of 97.1%（500/515）．Two weeks later，the Chinese version of DJGLS was retested in a convenience sample of 50 out of the 500 respondents.The reliability and validity of the Chinesized DJGLS were analyzed after the retest.Results　The Cronbach' α for the Chinese version of DJGLS，and the two subscales of social loneliness，emotional loneliness was 0.820，0.792，0.737，respectively.The scale's Guttman split-half reliability coefficient was 0.806，test-retest reliability coefficient was 0.889.The content validity index（S-CVI/UA）of scale level was 0.909，the content validity index（I-CVI）of item level ranged from 0.833 to 1.000，and the（S-CVI/Ave）was 0.985.The total scores of the scale were sorted from high to low，and 27%（155 cases，critical value was 7）of the top ranked and 27%（191 cases，critical value was 3）of the bottom ranked were assigned to high，and low score groups，respectively.The average score of each item achieved by the high score group was higher than that obtained by the low score group（P<0.05）．Item-total correlation coefficients ranged from 0.439 to 0.746（P<0.05）．Correlation coefficients of social loneliness，emotional loneliness with the total score were 0.913，0.772，respectively（P<0.05）．By exploratory factor analysis using principal component analysis with varimax rotation，two common factors with eigenvalue >1.000 were obtained，explaining 50.87% of the total variance.The loading values of items were from 0.405 to 0.832.Confirmatory factor analysis showed that，the fit indices were χ2/df=2.479，GFI=0.966，AGFI=0.945，IFI=0.967，TLI=0.956，CFI=0.967，PGFI=0.600，RMSEA=0.054，RMR=0.011.The CAIC for the default model，independence model，and saturated model was 281.993，1 976.182，and 476.164，respectively.The amended standardized regression coefficient ranged from 0.24 to 0.89.Conclusion　The Chinese version of DJGLS has good reliability and validity.It could be used to assess the level of loneliness among China's rural elderly.

Key words: Loneliness, Aged, Rural population, De Jong Gierveld Loneliness Scale, Reliability, Validity