Abstract: Objective To develop Eating Behavior Scale for children with high reliability and validity,so as to provide a scientific and reliable tool for the evaluation of children′s eating behavior in China. Methods A total of 349 parents of children aged 1—3 years were investigated. Then the initial questionnaires were formed. Factor analysis,discriminant analysis,Cronbach′s α coefficient and Spearman correlation analysis were used to screen the initial items. And 1 021 parents of children aged 1—3 years were asked to fill out primary questionnaire. Finally the formal questionnaire was formed. The reliability and validity of the formal questionnaire was tested by Guttman split-half reliability analysis,retest reliability analysis,exploratory factor analysis and confirmatory factor analysis. Results 1) Finally the Eating Behavior Scale of children consisted of 39 items with 7 dimensions was formulated. The Cronbach′s α coefficient of scale was 0.889,Guttman split-half reliability was 0.922,the retest reliability of the scale was 0.735,and the correlation coefficient between each factor and the scale ranged from 0.455 to 0.817. Confirmatory factor analysis showed that the values of x²/df,GFI,CFI,RFI and RMSEA of the scale were 2.980,0.815,0.802,0.708 and 0.062,respectively. 2) Children′s weight was related to birth weight,family structure and bad eating habits(P＜0.05). The length was related to family structure and low salt diet(P＜0.05). The education level of parents had a significant effect on children′s eating behaviors(P＜0.05). Family income had a significant effect on children′s food preference(P＜0.05). The effects of different family structures on food response,satiety response and common meal opportunity were significantly different(P＜0.05). Conclusion Children′s Eating Behavior Scale is a scientific and reliable tool with good reliability and validity,which can objectively evaluate the dietary behavior problems of 1- to 3- year-old children in China.

Key words: children, Eating Behavior Scale, validity, reliability

摘要： 目的 编制一套本土化的幼儿饮食行为量表,以早期识别、干预幼儿饮食行为问题。方法 对349名幼儿行初始问卷调查。运用因子分析、区分度分析、Cronbach′s α系数及Spearman相关分析筛选条目,形成试用问卷。对1 021名幼儿行试用问卷调查,筛选条目并形成正式量表。运用折半信度、重测信度、探索性因子分析和验证性因子分析等方法评价量表的信效度。结果 1)形成包含7个维度、39个条目的幼儿饮食行为量表,Cronbach′s α系数为0.889,折半信度为0.922,重测信度0.735,与各维度的相关系数为0.455~0.817。验证性因子分析示χ²/df=2.980,GFI=0.815,CFI=0.802,RFI=0.708,RMSEA=0.062。2)幼儿体重与出生体重、家庭结构、不良饮食习惯相关(P<0.05);幼儿身长与家庭结构和低盐饮食相关(P<0.05)。父母文化水平对幼儿饮食行为的影响显著(P<0.05)。家庭收入对幼儿挑食偏食影响显著(P<0.05)。不同家庭结构对食物响应、过饱响应和共同进餐机会的影响差异有统计学意义(P<0.05)。结论 幼儿饮食行为量表具有稳定性和可靠性,可真实客观地反映幼儿的饮食行为问题,是我国本土化幼儿饮食行为评估的科学可信工具。

关键词: 幼儿, 饮食行为量表, 效度, 信度