Ronis D L, Antonakos C L, Lang W P
Institute for Social Research, University of Michigan, Ann Arbor 48106-1248, USA.
Health Educ Q. 1996 Nov;23(4):512-27. doi: 10.1177/109019819602300410.
Canonical correlation analyses of a previous dental survey suggested that dental checkups, flossing, and tooth brushing could all be predicted from a single equation. Most theories and research about the influence of beliefs on behavior, however, suggest different behaviors will be best predicted by different behavior-specific measures. The current survey investigated influences on brushing, flossing, and dental checkups in a probability sample of adults in the Detroit tri-county area. Both behavior-specific variables, such as perceived benefits and costs of flossing, and general variables, such as gender, were included as predictors. Canonical correlation analysis indicated three equations were needed to predict the three oral health behaviors. Flossing frequency, for example, was best predicted by confidence in flossing ability and beliefs about the benefits of and barriers to flossing. The results suggest that even these closely related behaviors are best predicted using separate equations that include mostly behavior-specific predictors.
先前一项牙科调查的典型相关分析表明，牙齿检查、使用牙线和刷牙都可以通过一个单一方程进行预测。然而，大多数关于信念对行为影响的理论和研究表明，不同的行为将通过不同的特定行为测量方法得到最佳预测。当前的调查在底特律三县地区的成年人概率样本中，研究了对刷牙、使用牙线和牙齿检查的影响。特定行为变量，如使用牙线的感知益处和成本，以及一般变量，如性别，都被作为预测因素纳入。典型相关分析表明，需要三个方程来预测这三种口腔健康行为。例如，使用牙线的频率最好通过对使用牙线能力的信心以及对使用牙线的益处和障碍的信念来预测。结果表明，即使是这些密切相关的行为，使用包含主要特定行为预测因素的单独方程来进行预测也是最佳的。
