Webb G R, Redman S, Gibberd R W, Sanson-Fisher R W
Faculty of Medicine, University of Newcastle, NSW, Australia.
Drug Alcohol Depend. 1991 May;27(3):223-31. doi: 10.1016/0376-8716(91)90005-j.
The study aimed to assess the test-retest reliability of two commonly used measures of alcohol consumption, the quantity-frequency (QF) method and the diary method, as well as the stability of scores on the two measures over time. Two methods of assessing reliability and stability were employed. The first was a traditional method based on calculation of correlation coefficients for agreement between scores on repeated measures over a short retest interval to yield test-retest reliability coefficients, and over a long retest interval to yield stability coefficients. The second method was that devised by Wiley and Wiley (1970) to differentiate the effects of reliability and stability on repeated measures over time. The two methods were applied to a sample of heavy drinkers and to a sample of light drinkers. The results indicated that both the QF and diary measures are reliable in measuring alcohol consumption of light drinkers. Both measures are less reliable for heavy drinkers. The results indicate, in addition, that drinking consumption levels of light drinkers demonstrate a high degree of stability. However, the consumption levels of heavy drinkers demonstrate less stability, especially over a long time period. Heavy drinkers significantly reduced reported levels of alcohol consumption on both measures after the first test, suggesting a regression to the mean effect or the possibility of unintended intervention effects due to repeated measurement of drinking behaviour.
该研究旨在评估两种常用饮酒量测量方法——量频（QF）法和日记法的重测信度，以及这两种测量方法的分数随时间的稳定性。采用了两种评估信度和稳定性的方法。第一种是传统方法，基于计算在短重测间隔内重复测量分数之间一致性的相关系数以得出重测信度系数，在长重测间隔内计算相关系数以得出稳定性系数。第二种方法是由威利和威利（1970年）设计的，用于区分信度和稳定性对随时间重复测量的影响。这两种方法应用于重度饮酒者样本和轻度饮酒者样本。结果表明，QF法和日记法在测量轻度饮酒者的饮酒量方面都是可靠的。这两种方法对重度饮酒者的可靠性较低。此外，结果表明轻度饮酒者的饮酒消费水平具有高度稳定性。然而，重度饮酒者的消费水平稳定性较低，尤其是在较长时间段内。重度饮酒者在首次测试后，两种测量方法报告的饮酒量水平均显著降低，这表明存在均值回归效应或由于对饮酒行为的重复测量而产生意外干预效应的可能性。
