初中生吸烟相关态度量表的效度和信度评价

Assessment on a Scale for Measuring Smoking-related Attitudes in Chinese Secondary School Students

目的:研制适用于中国初中生的吸烟相关态度量表。方法:参考国内外文献设计了16个条目的吸烟相关态度量表。用因子分析评价结构效度;比较尝试吸烟者与未吸烟者、男生与女生的差异来评价区分效度;用Cronbach'α系数和折半信度来评价量表的信度;比较前后两次测量的差异与相关系数来评价重测信度。结果:因子分析提取两个主成分解释总变异的46.2%,旋转后因子1定义为对烟草及吸烟相关行为的认同,因子2定义为对控烟措施的认同;女生态度得分高于男生,未吸烟者的得分高于尝试吸烟者。两个因子和所有条目的α系数分别为0.87、0.75和0.86,折半信度分别为0.71、0.59和0.72。重测(112人)相关系数分别为0.72、0.44和0.67,关联均有显著性;两次调查结果接近,差异无显著性。结论:该量表具有较好的效度和信度,可以为国内类似调查提供参考。

Objective：To develop a scale used to measure smoking-related attitudes in Chinese secondary school students. Methods：Based on literatures published in China and abroad and suggestions by relevant specialists and teachers, we developed a self-administered scale with 16-item used to measure smoking-related attitudes in Chinese secondary school students. Factor analysis was used to assess the construct validity of the scale. The differences of the scores between trying smokers and non-smoker, and males and females were tested for assessing the discriminant validity of the scale. Cronbach＇s alpha for the total scale and the sub-scales was calculated for evaluating the internal consistency, and the split-half reliability test analysis was also conducted. Moreover, 112 subjects were re-investigated after two weeks of the first survey, and the difference and correlation of scores between the two surveys were tested for assessing the stability of the scale. Results： Factor analysis identified two potential components that could explain 46.2% of the total variance, and the first factor including 9 items was defined as “The opinions to tobacco and smoking-related behaviors”, and the second one including 7 items was defined as “The opinions to measures for preventing and controlling smoking”. The mean score of the scale in non-smokers was significantly higher than that of the trying smokers, while the mean score in females was significantly higher than that in males. The Cronbach＇α coefficient was 0.87 for Factor One, 0.75 for Factor Two, and 0.86 for the general scale, and the coefficient of the split-half reliability was 0.71 for Factor One, 0.59 for Factor Two, and 0.72 for the general scale. By analyzing the data of 112 students who participated in both surveys, we obtained test-retest reliability of factor Ⅰ as 0.72, of factor Ⅱ as 0.44 and of all items as 0.67, and all the correlation coefficients were statistically significant. There were no differences for the scores of the scale between the two surveys. Conclusion： The results indicated that the smoking-related attitude scale had reasonable validity and reliability,It could provide valuable reference for future similar surveys in China.