育儿评估量表的研制及应用研究

【作者】 程黎明；

【导师】 戴耀华；

【作者基本信息】 北京协和医学院 ， 儿科学， 2011， 博士

【摘要】 《育儿评估量表》研制及应用研究[目的]养育可定义为自出生至成人期,父母为个体的体格、情感、智能及社会能力的发展所提供的一切支持。养育并非是一个单一的结构,它所包含的内容非常广泛,它是父母的养育观念、养育行为及其对儿童情感表现的一种组合方式,这种组合方式是相对稳定的,不随情境的改变而改变。近年来的研究表明父母的养育状况与儿童的发展结局存在多种关联。积极的养育观念和行为可以促进儿童的发展,而消极的养育观念和行为则导致儿童的适应不良及心理行为问题。对父母养育状况进行评价有助于鼓励积极的养育观念和行为,减少消极的养育观念和行为。但目前国内尚没有可信且可靠的用于评价父母养育知识和行为的工具。因此,本研究的目的在于编制一套中国社会文化背景下的父母养育状况评价量表,对父母为儿童提供各种支持等养育环境的组成进行评价。[方法]为保证量表满足测评学指标,采用标准的量表编制程序。首先,由于养育这一研究结构包含的内容非常丰富,利用多途径收集数据信息,编制量表条目库：①充分查阅国内外有关养育及养育评价的文献,借鉴既往研究经验,获得有关养育观念、养育方式及养育行为等信息,以及当前现有的养育测评工具的信息。②专家咨询：儿童心理专家、测评学专家、养育专家共26名。③专题小组讨论：以家长和从事幼教、儿童保健的专业人员为访谈对象。包括50名专业人员及50名3-6岁儿童家长(其中3岁儿童家长15名、4岁儿童家长18名、5岁儿童家长17名),每个专题小组由10名专业人员或家长组成。综合上述三种方法所获得的信息,采用逻辑法编制量表条目库。在条目库形成后,对条目进行一般性修改,目的是确定条目的可接受性、相关性、可理解程度及完整性。①邀请来自儿童保健及测评学领域内的专家对条目库进行审阅,补充可能的遗漏,并进行必要的修改。②同时于2009年6月份在首都儿科研究所儿童保健门诊选择了30名家长,要求家长通读条目,指出相关性不强、语义难以理解或者有重复的条目,予以修改或删除。经一般性修改后,再次对条目进行筛选：①专家评分法,目的是确定条目的可接受性、相关性、可理解程度及完整性。邀请儿童保健及测评学领域内的专家,对条目库进行重要性评分,0分为不重要,1分为重要。计算每个条目的重要性得分,保留重要性得分高于所有条目平均得分的条目；②家长试测,对象为98名3-6儿童家长,通过方便抽样的方法在首都儿科研究所及湖南省妇幼保健院儿童保健及免疫接种门诊获得。利用SPSS13.00软件对试测结果进行探索性因素分析,保留负荷值>0.3的条目。综合专家评分法及因素分析对量表条目进行筛选,形成《育儿评估量表》正式版本。大规模正式测评。正式测评样本兼顾地理位置分布、大中小型城市和儿童年龄分布。于2009年9月-2011年3月,选择北京市、上海市、南京市、厦门市、长沙市、广州市、天津市、浙江湖州市、武汉市、山东济南、山东威海、山东荣成、广西南宁市、广西天等县、河北河间、山西壶关等十六个城市作为研究地点。各地分别选择3-6岁儿童及其家长约60名。其中评估人员从各研究地点的妇幼保健机构抽调。在正式调查开始之前,将各地评估人员集中进行培训,采用视频案例教学的方法,要求严格按照统一的指导语进行测评,培训合格者方可进行现场测评工作。正式测评结束后,将测评后量表集中审阅核查,建立数据库,双边录入,并进行一致性核查。经过核查后的数据方可进入分析。项目分析采用相关性分析计算各个条目与量表总分的相关系数。删除与总分相关系数<0.3的条目。然后对保留条目进行二次探索性因素分析,以探求《育儿评估量表》各变量间的潜在结构和共同因素。根据KMO值和Banlett球度检验的结果判断因素分析的适宜性；采用主成分分析法抽取共同因素,求得初始变量相关矩阵,选用最大变异法进行正交旋转,删除负荷量<0.3的条目,保留有效条目。然后进行相应的信度、效度检验、验证性因素分析和实证分析。[结果]1.在条目库生成步骤中,根据专题小组讨论、专家咨询及文献查阅的结果共编写了含有117个条目的初始条目库,经过一般性修改,保留了69个条目。2.在条目筛选步骤中,共有15名专家对69个条目进行了重要性评分、98名3-6岁儿童及家长接受试测。同时满足专家评分法及因素分析法筛选标准的条目共有50个。最终保留的50个条目经因素分析共分为8个因子,可解释总方差变异的65.54%,8个因子的命名如下：(1)育儿观念,(2)关注和接纳,(3)自我调控能力的培养,(4)养育关系与沟通,(5)学习环境,(6)语言环境,(7)玩耍及娱乐,(8)安全及居住环境。该8个因子和50个条目形成《育儿评估量表》正式版本。3.利用正式版本进行大规模正式测评,对测评结果进行项目分析、第二次探索性因素分析和信度、效度检验。项目分析结果显示各条目得分与全量表得分之间的相关系数范围：0.310-0.565.第二次探索性因素分析的结果显示：8个因子可以解释总方差变异的73.554%。信度检验的结果：①量表同质信度克伦巴赫α系数(Cronbach’s alpha coefficient)为0.86,8个因素的克伦巴赫α系数分别为0.59、0.75、0.78、0.53、0.76、0.73、0.79、0.66。②量表总体Guttman分半信度为0.811,8个因素Guttman分半信度分别为0.766、0.765。③重测信度：对60名儿童及其家长在间隔2周后进行复测,总体重测相关系数为0.84,8个因素的重测相关系数分别0.84、0.69、0.82、0.89、0.70、0.68、0.81、0.78。4、效度分析的结果：①内容效度：本研究通过文献查阅、专家咨询、家长及专业人员专题小组讨论的方法构建条目库,并通过专家审阅及专家评分法筛选条目,保证了内容效度；②结构效度：各分量表得分与全量表得分之间的相关系数范围：0.611-0.736；各分量表得分之间的相关系数范围为：0.342-0.599.均呈中到高度相关,说明各因素之间既具有一定的独立性。各条目得分与其所属分量表得分之间的相关系数范围为：0.382-0.684,呈中到高度相关,说明各条目与其所属的因素紧密相关。验证性因素分析模型拟合达到要求,八因子模型的均方根近似误差(rmsea)为0.04,表明一个紧密配合的结构与有关自由度。NFI和CFI的值分别是0.84和0.91,说明结构的紧密性。③准则关联效度：育儿评估量表总分与儿章语言能力、社会能力及适应性行为得分之间的相关系数分别为：0.37、0.39和0.48.语言环境得分与儿童语言能力得分之间存在显著的正相关,相关系数为0.60,关注和接纳得分与儿童适应性行为得分均呈正相关,相关系数为0.49.养育关系与沟通及语言环境与儿童的适应性行为之间也呈显著正相关,相关系数分别为：0.40和0.36。此外,养育关系与沟通与儿童的语言能力和社会能力得分之间也呈显著正相关，相关系数分别为:033和0.30.育儿评估量表总分与儿童发育商之间也呈显著正相关，相关系数为:0.3 19.5、进一步实证效度:①量表总分和若干分量表(育儿观念、学习环境、语言环境、玩耍与娱乐、安全和居住环境)得分在父母不同学历程度之间均表现出统计学差异(P均=0.000)，随父母文化程度的提高而增高。②分量表养育关系与沟通得分在不同父母婚姻状况之间表现出了显著性差异，婚姻状况正常父母的得分高于离异父母评分。④父亲不同工作时间分组间，量表总分、分量表养育关系与沟通、学习环境表现出了显著的组间差异;而在母亲不同工作时间分组间，量表总分、育儿观念、养育关系与沟通、学习环境、玩耍与娱乐得分表现出了显著的组间差异。【结论】1.本研究摈弃了单纯翻译引进国外量表的做法，而是基于我国社会文化背景，自主地编制了《育儿评估量表》，本量表的研制完成，首次为国内3一6岁儿童父母养育状况的评价提供了可靠的工具。2.经项目分析和因素分析，论证了该量表结构合理，量表条目在评价父母养育知识/观念、养育技能/行为方面有一定的鉴别力。3.信度分析和效度检验证明，《育儿评估量表》的同质信度、分半信度和重测信度以及内容效度、结构效度和标准相关效度均达到测评学的要求。4.父母不同文化程度、不同职业、不同工作时间及不同级别城市之间育儿评估量表得分差异显著。表明该量表均有一定的实证效应，能够在一定程度上反映父母养育的实际状况。更多还原

【Abstract】 【Baekground】Parenting 15 defined as the Proeess of Promoting and suPPorting thePhysical，emotional，social，and intelleetual develoPment of a ehild from infaney toadulthood.Parenting refers to the activity of raising a ehild rather than the biologiealrelationshiP.Ineluding variety of eontents，Parenting ean not be treated as a singlestructure but should be seen as a combination of Parental ideas/如owledge，behaviour/skills and styles.The eondition of Parenting won，t change over time and 15relatively stable.Parenting attitudes and skills have been linked to manydeveloPmental outeomes.For examPle，asPeets of Positive Parenting，sueh as the mother’s "sensitivity", including warmth and responsiveness, can predict the child’s academic readiness, self-control, and social competence. In contrast, parental behaviors identified as harsh, including physical and verbal discipline, have been linked to maladaptive behaviors, including internalization and inappropriate externalization of problems. These conflicting results concerning parenting styles in China suggest that these questions require an updated assessment but the main concern is that there is no reliable and valid tool for testing parenting ideas and behaviours in China.[Methods]Standardized procedure was used to develop the scale for ensuring that the scale can satisfy psychometric norms.First, information was collected from many sources for developing the item pools:①literature review:reviewing literatures regarding parental ideas/knowledge, behaviours/skills, styles and measuring tools;②experts consultation:consulting experts from psychological, psychometric and parenting fields etc;③Focus group discussion:parents, kindergareners and professionals of child health care were invited to participate into the focus group discussion. Information extracted from the above three sources was used to develop the item pool.After being established, general revision was done to the item pool aimed to confirm the acceptability, relevancy, degree of understandability and integrity of items. First, professionals of child healthcare and experts from psychometric field were invited to read through the item pools as to supplement some contents that possibly being omitted and do necessary modification. At the same time,30 parents were chosen from the clinics of child health care in Capital Institute of Pediatrics and were asked to read through the items, among which those being considered as difficult to understand, with less relevancy or duplicate were deleted or verified.Item selection was done by:①experts pointsystem, by which items were graded by professionals of child healthcare and psychometric field as 0 score or 1 score, with the former refers to unimportant and the latter refers to important. Scores of importance for each item were calculated, items with scores higher than the average scores of all of items were retained;②field test, subjects were 98 parents with their children aged 3-6 years from the child health care clinics and immunization clinics by convenience sampling in Capital Institute of pediatrics and Maternal & Child Health Care Institute of Hunan province. SPSS 13.0 was used to do analysis with the data from field test. Exploratory factor analysis was done and factors with eigenvalues more than 1 and items with loading values more than 0.3 were remained. The results of experts pointsystem and factor analysis were integrated as the evidence to select items, with those being retained to form the formal version of Parenting Assessment Tool.After item selection, large-scale investigation was done. From September 2009 to March 2011, Beijing, Shanghai, Nanjing, Xiamen, Guangzhou, Changsha, Tianjin, Huzhou, Wuhan, Jinan, Weihai, Rongcheng, Nanning, Tiandeng, Hejian, Huguan 16 cities were made as the studying sites. About 60 dyads of parents and their children aged 3-6 years were chosen from each city. The assessors were staff from the maternal and child health care institute in each city. Before formal investigation, all assessors from each place were trained together by video case and were asked to assess parents according to the unified scale and administration manual. All the queationnaires completed were reviewed and checked twice before being entered into computer and analysed. Epidata 3.0 was used to establish database and SPSS 13.0 was used to do analysis.Pearson’s correlation coefficients between item score and total scale scores were calculated for item analysis. Items were deleted whichever correlation coefficient to be lower than 0.30 separately. Second exploratory factor analysis was then done to the items remained after the item analysis procedure for exploring the potential construct and mutual factors of the Parenting Assessment Tool. After Kaiser-Meyer-Olkin measure of sampling and Bartlett’s test of sphericity done, principal component analysis and varimax rotate were performed to extract the factors. Reliability analysis including internal consistency reliability (Cronbach’s alpha coefficients, Guttman’s split-half reliability), and test-retest reliability, validity analysis including content validity, construct validity, and criterion-related validity and confirmatory factor analysis were evaluated respectively.[Results]1、One hundred and seventeen items were generated by focus group discussion, experts consultation and literature review as the item pool, after general revision,69 items were remained.2. Fifteen experts were invited to grade the importance for the 69 items and 98 parents were administered the 69 items. Fifty items satisfied standards for remaining of both experts pointsystem and factor analysis. Eight factors were extracted from the 50 items and can explain 65.54% of the total variance. Names of the 8 factors were as follows:①Parenting ideas;②Attention and Acceptance;③Self-regulation;④Nurture and Communication;⑤Studying Stimulation;⑥Language Stimulation;⑦Play and Enjoyment;⑧Safety and Living environment.3. Reliability anslysis:①The Cronbach’s alph coefficient of the whole scale was 0.86, and for eight subscales, the Cronbach’s alph coefficients were 0.59,0.75,0.78,0.53, 0.76,0.73,0.79,0.66 respectively;②The Guttman’s split-half coefficient of the whole scale was 0.811, for eight factors the guttman’s split-half coefficients were 0.766,0.765 respectively;③Test-retest reliability of the whole scale was 0.84, and for eight factors, the correlation coefficients were 0.84.0.69.0.82.0.89.0.70.0.68. 0.81.0.78 respectively.4. Validity analysis:①Content validity:this study developed the item pool by literature review, experts consultation and focus group discussion, and selected the items by experts pointsystem and factor analysis, so the content validity of the scale was ensured;②Construct validity:correlation coefficients between total scale and factors ranged from 0.611 to 0.736. between eight factors ranged from 0.342-0.599. between items and factors they belong to ranged from 0.382-0.684. Confirmatory factor analysis showed the model fitting was good.③Criterion-related validity: correlation coefficient between total score and child development quotient (DQ) was 0.319、between total score and children’s language ability, sociability and adaptive ability were 0.37,0.39 and 0.48 respectively. Correlation coefficient between Language Stimulation and children’s language ability was 0.60, between Attention & Acceptance and adaptive ability was 0.49. Correlation coefficients between Nurture & Communication、Language Stimulation and adaptive ability were 0.40 and 0.36 respectively. Correlation coefficients between Nurture & Communication and language ability and sociability were 0.33,0.30 respectively.5、Significant differences were found between parenting scale scores and different parental educational levels, marital status, working times, and careers, which showed partial empirical validity of Parenting Assessment Tool.[Conclusions]1、This study developed and tested the psychometric indicators of Parenting Assessment Tool independently within Chinese social and cultural context. It is the first time to establish such a measurement for assessing the parenting conditions in China.2、After item analysis and factor analysis, it is confirmed that the Parenting Assessment Tool is a reliable and relatively valid tool. It has an acceptable structure and can distinguish parenting ideas/knowledge, behaviours/skills and styles to some extent.3、The reliability and validity analysis showed that of Parenting Assessment Tool can satisfy the psychometric standards.4、There were significant differences existed between different parental educational level, work times, occupations. These findings indicated that Parenting Assessment Tool had some empirical validity and can reflected the actual conditions of parenting.更多还原