Boyd Orr Centre for Ecosystem and Population Health, School of Veterinary Medicine, College of Medical, Veterinary and Life Sciences, University of Glasgow, G61 1QH, United Kingdom.
Prev Vet Med. 2011 Oct 1;102(1):66-74. doi: 10.1016/j.prevetmed.2011.06.010. Epub 2011 Aug 5.
The objectives of the study were to investigate whether owners were able to assign the correct body shape to their dog and to assess the dog and owner level factors associated with incorrect owner assessment of dog body shape. Six hundred and eighty questionnaires were administered to dog owners in 5 first opinion practices around Glasgow during July 2007. At the same time, the interviewer and owner assessed the body shape of each dog. The interviewer assessment of body shape was taken as the gold standard. Incorrect owner assessment of dog body shape (misperception) was divided into two groups: underestimation and overestimation. Multinomial logistic regression and classification and regression trees (CART) were used to assess risk factors associated with each type of misperception. Misperception of dog body shape was present in 44.1% of owners with underestimation (i.e. the owner considered the animal to have a leaner body shape from the gold standard assessment) being the most common form of misperception. Risk factors identified by both multinomial logistic regression and CART were gender of owner, age of the dog and dog body shape. The classification tree appeared to have improved predictive ability when compared to the multinomial model.
本研究旨在探讨犬主是否能够正确判断犬的体型，并评估与犬主错误评估犬体型相关的犬和犬主因素。2007 年 7 月，在格拉斯哥的 5 家初级保健机构中，向 680 名犬主发放了调查问卷。与此同时，评估员和犬主会对每只犬的体型进行评估。评估员的体型评估被视为金标准。犬主对犬体型的错误评估（误判）分为两组：低估和高估。使用多项逻辑回归和分类回归树（CART）来评估与每种误判类型相关的风险因素。44.1%的犬主存在犬体型误判，其中最常见的误判形式是低估（即犬主认为动物的体型比金标准评估的更瘦）。多项逻辑回归和 CART 都确定的风险因素包括犬主的性别、犬的年龄和犬的体型。与多项模型相比，分类树似乎提高了预测能力。
