Tryon Warren W
Department of Psychology, Fordham University, New York, New York 10458-5198, USA.
Behav Res Methods. 2005 Aug;37(3):492-7. doi: 10.3758/bf03192719.
Evidence for the reliability and validity of two models of Ambulatory Monitoring, Inc. actigraphs was obtained by testing four instruments of each kind 10 times each on a precision pendulum. Correlation and coefficient of variation methods were used to analyze the data. Reliability coefficients of .98 were obtained for both models. Coefficient of variation methods yielded reliability coefficients of 92% for the MotionLogger model and 97% for the BuzzBee model. Validity coefficients of.99 were obtained for both models. However, MotionLogger means were found to differ by up to 36% from one instrument to another, whereas BuzzBee means differed by 10% at most. Issues of standardization and measurement units were discussed.
通过在精密摆锤上对每种类型的四个仪器各测试10次,获得了动态监测公司两种活动记录仪模型可靠性和有效性的证据。采用相关性和变异系数方法分析数据。两种模型的可靠性系数均为0.98。变异系数方法得出MotionLogger模型的可靠性系数为92%,BuzzBee模型为97%。两种模型的有效性系数均为0.99。然而,发现MotionLogger各仪器之间的均值差异高达36%,而BuzzBee各仪器之间的均值差异最大为10%。讨论了标准化和测量单位的问题。