การวิเคราะห์อภิมานและการวิเคราะห์กลุ่มแฝงเพื่อศึกษาคุณลักษณะนวัตกรรมที่พัฒนาทักษะและกระบวนการทางคณิตศาสตร์ของการศึกษาขั้นพื้นฐาน : การประยุกต์ใช้วิธีการจับคู่คะแนนโน้มเอียง

Abstract

The purpose of this research was to1) to study characteristics and assess the quality of innovative research that develops mathematical skills and processes. 2) To study the magnitude of influence of innovative research on developing mathematical skills and processes. 3) to group teaching styles that develop skills, mathematical processes that develop mathematical skills and processes, and 4) To compare influence size, mean, influence size of research on different teaching models after adjusting matching the Propensity Score From theses for Master’s degree or independent studies from state universities published Between the years 2008 – 2019 from 213 research topics. The instruments used for data collection were research feature records. And research quality assessment form, Data analysis using descriptive statistics, the research influence magnitude was calculated using meta-analysis using the method of Hedges (Card.) Information collected It is the research attribute data and research results in the form of influence size value. Tools used for data collection Is the research feature record form And research quality assessment form Data analysis using descriptive statistics Meta-regression analysis with a fixed effect and Random effect Latent Class Analysis : (LCA). And analysis of differences of mean, size, influence of each pair and take the test later (Post - Hoc Comparison). Matching and comparing the Propensity Score. The results of the research found that 1) The magnitude of the influence of research related to teaching style innovations to develop skills and mathematical process Out of 213 studies, 213 influence magnitude values (d) were calculated. An influence size value between -8.744 and 12.823 and the mean influence size (d) was 1.936, meaning that Innovation has a 1.936 times deviation in mathematical skills and processes development. Standard. Influence magnitude distribution characteristics are higher than the mean value. It indicates that the influence magnitude is inherent. The distribution is right-angled (Sk = 1.099) and the distribution of the data is higher than the normal curve (Ku = 3.296). and the research quality was averaged 2.836, meaning that the research quality was at a good level. 2) The results of the analysis of research characteristics variables affecting the influence size from 20 variables found that there were 4 variables with statistical significance at .05 level, namely 1 Field of research production (2) experimental pattern (3) data analysis (4) the other variables, there is no difference. 3) The results of the classification of research groups according to the teaching style were divided into 4 groups: Group 1 The self-knowledge-building (d= 1.825) Group 2 focuses on the problem-based process-based (d=0.203) Group 3, working group as a cooperative learning group (d= 1.550) , and group 4, the examination learning group (d=1.652). It was found that Group 1 promotes self-knowledge creation. higher than another group, group 2 the problem-based process statistically significant at level .05 4) The results of a comparative analysis of teaching styles that develop skills and educational processes Mathematics at the basic educational level before using the inclination score matching technique. It was found that the mean size The total influence was 1.936. The self-knowledge-building group had the highest mean of the influence size. (d= 1.803), followed by the examination learning group (d=1.625). The working group as a cooperative learning group (d=1.550) and the problembased process-based group had the least influence (d=0.203) after the study. Adjusted by the inclination score, it was found that the mean total influence size was 0.33, the self-knowledge-building group had the highest mean of influence size (d=0.173), followed by the examination learning group. The survey form (d=0.162), the third was the cooperative learning group (d=0.157), and the problembased process-based group had the mean least influence size (d=0.077), and there was a significant effect of influence at the .05 level.