关系框架理论（Relational Frame Theory, RFT）从人类复杂行为的功能语境取向，提出任意适应关系反应或关系框架的流畅性和灵活性是语言和认知能力的基础。本研究聚焦于关系框架理论，评估学龄前儿童一致关系、对立关系、比较关系、层级关系、时间关系、空间关系以及隐喻理解反应的发展水平，并对学龄前儿童的关系反应水平进行训练，检验关系反应训练对提升中国儿童的推衍关系反应水平的有效性。本研究以期从行为分析的研究视角为学龄前儿童言语和认知发展提供新的证据，并为关系反应训练在早教和学前教育中应用提供依据和支持。
研究二：关系反应训练的有效性探究。通过微信在参与过评估的儿童中招募21名被试。将被试配对分为训练组和Waiting-List 组。选取学龄前儿童熟识的动物类、食物类、交通工具等图片作为训练、前后测和阶段测试的材料。用SPSS22.0 对数据进行统计分析，使用 Bootstrap 方法，对训练组和Waiting-List组的前测、阶段测试以及后测的数据进行方差分析。
研究一以1 岁为单位，将被试进行分组，结果显示：（1）一致关系、对立关系、比较关系、层级关系、时间关系、空间关系和隐喻理解7维度上，年龄的差异达到显著水平。（2）分别以年龄为自变量，关系反应水平为因变量，进行回归分析，年龄对7 种关系反应水平均有正向预测作用。（3）对不同性别被试的关系反应水平进行t 检验，不同性别在7种不同关系反应上差异均未达到显著水平。（4）对不同依恋表现（主试观察）和依恋类型（问卷调查）的被试的关系反应水平进行差异检验，安全型被试的关系反应水平显著优于不安全型被试的关系反应水平。
研究二对训练组和Waiting-List组的前后测、阶段测试和训练阶段的正确率进行差异比较，结果显示：（1）训练组和Waiting-List 组的前测正确率均显著高于后测的正确率。（2）Waiting-List组在训练组训练开始后一个月进行训练，比较延迟的一个月中进行的两次阶段测试正确率的差异，在第2 次阶段测试中，训练组的在具体的联合推衍关系反应正确率显著高于Waiting-List组的正确率。
This research focuses on Relational Frame Theory, which was derived from Skinner’s Verbal Behavior. This theory told about that any applicable relational responding or the fluency and flexibility of relational framing are foundation of the language and cognitive ability. There are two purposes that we look forward to finding the new evidences about development of preschooler’s language and cognitive ability from behavioral analytic view. The first one is evaluating the developmental tendency of preschooler’s relational responding, such as coordination, opposition, comparison, hierarchical, temporal, spatial and metaphor. The second is exploring the effect of relational training, which can improve Chinese preschooler’s derived relational responding.
There are two studies.
Study 1: evaluation of preschooler’s relational responding. There are 230 valid
participants who were collected from 554 preschoolers by Wechat, from 2 to 7 years old. According to Wechsler intelligence scale and published researches of Relational Frame Theory, we collected plenty of pictures, which were known well from preschoolers, such as animals, foods, transportations, etc. Participants were individually finished the test in a quiet room. We were using variance analysis to compare the differences between age groups, and also using regression analysis to test the tendency of development of the relational responding by SPSS version 22.
Study 2: effectiveness of relational responding training. There are 21 preschoolers who participated in study. All participants were divided into two groups: training group and waiting-list group. We selected lots of preschoolers familiar pictures, which were training, pre-test, post-test and stage-test materials, such as animals, foods, transportations, etc. We were using bootstrap and variance analysis to compare the difference between training group and waiting-list group in pre-test, post-test and stage-test by SPSS version 22.0.
In study 1, participants were divided into different group by age, we found that, (1) for different age group, the relational responding scores of participants had significant difference in coordination, opposition, comparison, hierarchical, temporal, spatial and metaphor; (2) we chose independent variable, age, and dependent variable, the score of relational responding. The result of regression analysis found that age predicted the score of relational responding positively; (3) for different gender group, the relational responding score of participants had no significant difference; (4) the relational responding score of secure participants was significantly higher than insecure participants, no matter from experimenter’s observation or from questionnaire.
In study 2, we compared the accuracy of pre-test, post-test and stage-test between training group and waiting-list group. We found that: (1) the post-test accuracy of both training group and waiting-list group was significantly higher than pre-test; (2) waiting-list group began the training one month after the training group, and the difference of the two stage-tests in this month between training group and waiting-list group was compared. The result suggested that the non-arbitrarily combinatorial entailment accuracy of training group was significantly higher than waiting-list group in the second stage-test.
Firstly，as growth of the age，the ability of preschooler’s relational responding is
improving gradually, such as coordination, opposition, comparison, hierarchical,
temporal, spatial and metaphor.
Secondly, different relational responding has different developmental tendency,
coordination, opposition and comparison develop earliest, temporal and spatial take second place, hierarchical and metaphor are the last. It is important that 4-year-old is the turning point of development of the relational responding.
In addition, attachment has influence on relational responding, which secure
participants have better performance than insecure participants.
Lastly, relational responding training is effective to improve the ability of preschooler’s relational responding.