우리나라의 경영학 관련 주요 학회지에 게재된 메타분석 연구현황

김용겸(Kim, Yong-Kyeom) 대한경영학회 2010 大韓經營學會誌 Vol.23 No.4

BHHR 메타분석과 Hunter-Schmidt 메타분석 간의 비교 분석

오순문 한국교원대학교 교육연구원 2010 敎員敎育 Vol.26 No.5

The purposes of this study are to investigate the comparativeness of two different meta-analysis, and then with two kinds of methodologies, to investigate the relationship between the moral leadership of school principal and the teachers' job satisfaction. There are several distinguishing differences in two methodologies ; Hunter-Schmidt Meta-Analysis(H-S meta-analysis) & the meta-analysis of Borenstein et al.(BHHR meta-analysis). First of all, the distinction between confidence intervals and credibility intervals is very important. While BHHR meta-analysis use the confidence intervals referring to estimates of a single value-the value of , H-S meta-analysis use the credibility intervals referring to the distribution of parameter values. Second distinction is whether or not to handle the correction of study imperfections. Unlike BHHR meta-analysis, H-S meta-analysis strives to make corrections for attenuation due to measurement error, and other artifacts. Finally, other differences are technical in nature. These include the way between-studies variance is estimated, the use of correlation coefficients rather than Fisher's Z-transformed values, etc. The effects of these choices have been extensively examined through simulations, which suggest that the differences are likely to be trivial in many cases. The analytic results for the relationship between moral leadership of school principals and the teachers' job satisfaction are as follows. According to BHHR meta-analysis, the weighted mean correlation is .617. Lower & Upper limit of the confidence intervals(95%) are .537 and .687, respectively. According to H-S meta-analysis, the is .660. Lower & Upper 10% of Credibility intervals(80%) are .526 and .793, respectively. 본 연구의 목적은 Borenstein 등이 주장하고 있는 메타분석과 최근 산업조직심리학 내지 조직행동 분야에서 지지를 얻고 있는 Hunter-Schmidt 메타분석에 대해 방법론적 측면에서 비교분석하고, 아울러 각각의 메타분석 방법론에 기초하여 학교장의 도덕적 지도성과 교사의 직무만족 간의 상관관계를 메타분석 하는 것이다. 우선 메타분석 여러 절차 중 메타통계분석 및 분석결과 제시에 초점을 맞춰 상이한 분석방법과 절차 및 결과 해석 등을 살펴보았다. 구체적인 방법론적 차이점들은 신뢰구간과 확신구간, 측정오차 등 오차의 교정, 피셔 z 전환값 사용 여부, 효과크기 비교, 동질성 검사 여부, 각각의 전용 메타분석 프로그램 등 6가지로 나누어 비교하였다. 또한 비교분석과 병행하여 학교장의 도덕적 지도성과 교사의 직무만족에 대한 상관관계를 Borenstein 등의 메타분석 방법과 Hunter-Schmidt 메타분석 방식에 기초하여 각각 CMA 프로그램과 Hunter-Schmidt Meta-Analysis Programs를 이용한 메타분석을 실시하고 그 결과를 제시하였다.

Association of Single Nucleotide Polymorphisms in Toll-like Receptor Genes With Asthma Risk: A Systematic Review and Meta-analysis

Kalthoum Tizaoui,Wajih Kaabachi,Kamel Hamzaoui,Agnès Hamzaoui 대한천식알레르기학회 2015 Allergy, Asthma & Immunology Research Vol.7 No.2

Purpose: Asthma is a complex disease, with contributions from multiple genes, various genetic backgrounds, and environmental factors. Many humanepidemiological studies have demonstrated that single nucleotide polymorphisms (SNPs) in Toll-like receptor (TLR) genes are inconsistently associatedwith asthma risk. Some have demonstrated differences concerning the study design and effect size, and conflicting results have been reported. A meta-analysis is necessary to determine the magnitude of this association. Methods: Following the Preferred Reporting Items for SystematicReviews and Meta-analyses guidelines, a systematic search and meta-analysis of the literature was conducted to estimate the associationof SNPs in TLR genes with asthma risk. We screened the medical literature based on the following keyword searches in MEDLINE and EMBASE databases:‘TLR’, ‘polymorphism’, ‘asthma’, and their combinations. Results: Meta-analysis of eight studies on TLR4 Asp299Gly showed a marginalassociation of TLR4 with asthma risk (odds ratio [OR]=0.814 [95% confidence interval [CI], 0.652-1.016; P=0.069]) in the recessive model. TLR4Thr399Ile was not associated with asthma risk under any genetic model. Meta-analysis of four studies on TLR2 Arg753Gln indicated that TLR2might be significantly associated with asthma in the dominant and codominant models (P=0.029, P=0.030, and P=0.009, respectively). TLR9 -1237was marginally associated with asthma risk (OR=0.408 [95% CI, 0.163-1.021; P=0.065]) in the codominant model. Analysis using the allele contrastmodel showed that the major TLR9 -1237 T allele tended to be a significant protective factor with OR=0.689 (95% CI, 0.471-1.007; P=0.055). Conclusions: The results showed that TLR4 Asp299Gly, TLR2 Arg753Gln, and TLR9-1237 might contribute significantly to asthma susceptibility. Futuregenetic association studies would consolidate these findings. Purpose: Asthma is a complex disease, with contributions from multiple genes, various genetic backgrounds, and environmental factors. Many human epidemiological studies have demonstrated that single nucleotide polymorphisms (SNPs) in Toll-like receptor (TLR) genes are inconsistently associated with asthma risk. Some have demonstrated differences concerning the study design and effect size, and conflicting results have been reported. A meta-analysis is necessary to determine the magnitude of this association. Methods: Following the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines, a systematic search and meta-analysis of the literature was conducted to estimate the association of SNPs in TLR genes with asthma risk. We screened the medical literature based on the following keyword searches in MEDLINE and EMBASE databases: ‘TLR’, ‘polymorphism’, ‘asthma’, and their combinations. Results: Meta-analysis of eight studies on TLR4 Asp299Gly showed a marginal association of TLR4 with asthma risk (odds ratio [OR]=0.814 [95% confidence interval [CI], 0.652-1.016; P=0.069]) in the recessive model. TLR4 Thr399Ile was not associated with asthma risk under any genetic model. Meta-analysis of four studies on TLR2 Arg753Gln indicated that TLR2 might be significantly associated with asthma in the dominant and codominant models (P=0.029, P=0.030, and P=0.009, respectively). TLR9 -1237 was marginally associated with asthma risk (OR=0.408 [95% CI, 0.163-1.021; P=0.065]) in the codominant model. Analysis using the allele contrast model showed that the major TLR9 -1237 T allele tended to be a significant protective factor with OR=0.689 (95% CI, 0.471-1.007; P=0.055). Conclusions: The results showed that TLR4 Asp299Gly, TLR2 Arg753Gln, and TLR9-1237 might contribute significantly to asthma susceptibility. Future genetic association studies would consolidate these findings.

Meta-Analysis of Chinese Coal Miners’ Mental Illness under SCL Index System

Sun Shanhui,Li Hong,Li Zhuangzhuang 보안공학연구지원센터 2016 International Journal of u- and e- Service, Scienc Vol.9 No.8

Objective: Under SCL index system, through the comparative study, this paper investigates whether there are significant differences in mental illness tendency between Chinese coal mine workers underground and on the ground. Methods: This paper uses computer to retrieve in Medline, Sciencedirect, Web of Science, the Cochrane Library, Springerlink, CNKI, Wanfang and Weipu database. The retrieval time is from the very beginning of building database to the March of 2015. After the extraction of available data, Stata 10.1 analysis software was applied to heterogeneity test on the research. Then the paper uses the random effects model to calculate and combine odds ratio as well as 95% confidence interval. Results: 15 articles were finally included by using Meta analysis. Collating data of the experimental group and control group, this paper carries on forest figure mapping, publication bias test, sensitivity analysis and cumulative Meta analysis, and one study result is found to have great deviation impact on the overall conclusion. Eliminating the above research result, the results are greatly improved. Conclusion: Through the result of the Meta analysis, this paper concludes that there are significant differences in mental illness tendency between Chinese coal mine workers underground and on the ground.

교육경제학 연구에서 메타분석법 적용가능성 탐색

이희숙(Lee, Heesook) 한국교육재정경제학회 2011 敎育財政 經濟硏究 Vol.20 No.2

전 세계적으로 교육경제학 연구 분야에서 메타분석법을 활용하여 선행 연구 결과를 체계적으로 종합하려는 시도가 적극적으로 진행되고 있음에도 불구하고, 우리나라의 경우 1980년대 기법이 소개된 이후 적용한 사례가 많지 않은 실정이다. 메타분석법은 기존 연구물들의 결과가 하나의 방향으로 수렴되지 않고, 서로 상반되어 하나로 종합되지 않을 때 기존 연구물을 연구 대상으로 하여 통계적ㆍ체계적으로 종합하는 기법이다. 이는 특히 교육정책의 효과성이나 교육의 효율성에 대한 연구 결과가 상이할 경우, 결과의 방향성과 크기를 도출할 수 있는 효과적인 기법이다. 그럼에도 이 기법을 교육경제학 연구 분야에 적용하지 못한 것은 기존의 메타분석법이 교육경제학 연구에서 활용된 분석 방법 특히 회귀분석법으로부터 효과크기를 도출하기에 적합한 형태로 발전되지 못한 데 그 원인이 있다. 그러나 최근 메타회귀분석(MRA), 부분상 관계수(semi-partial correlation) 등 새로운 메타분석 방법의 등장으로 교육경제학 연구에 메타분석법을 적용하여 선행연구 결과를 체계적으로 종합할 수 있는 가능성이 열리게 되었다. 이 연구에서는 교육경제학에서 메타분석법을 적용할 연구 과제 및 대상으로서 교육의 효율성 연구를 내적ㆍ외적 효율성 연구로 구분하여 연구 동향을 파악하고, 연구 방법으로서의 메타분석법을 고찰한다. 또한 메타분석법을 활용하여 교육의 효율성을 분석한 선행 연구물을 체계적으로 종합하고자 시도한 교육경제학 연구의 예를 탐색하며, 끝으로 최근 등장하고 있는 다양한 메타분석법 등을 소개함으로써 교육경제학 연구에서 메타분석법의 적용가능성을 탐색하였다. The purpose of this article is to explore the adaptability of meta-analysis to economics of education studies, especially to educational efficiency research. Internationally, the efforts of synthesizing previous studies using meta-analysis have been pervading in the economics of education research area. However, in Korea, the case of adapting meta-analysis to economics of education studies has been scarce since the introduction of meta-analysis in 1980s. Meta analysis is a useful tool of synthesizing results from previous studies on disputable and controversial topics. The reason why economics of education research area did not utilize this method was based on the lack of methodology in synthesizing regression coefficients from previous studies. Recently new meta-analysis methods such as Meta Regression Analysis(MRA) and Semi-Partial Correlation Coefficients and so on have been developed, and it could be feasible to combine previous studies in the field of economics of education in a synthesizing way. This study identified research trend on educational efficiency as the target of meta-analytic approach and reviewed "meta-analysis" as a methodology. In addition, previous efforts to combine results from previous studies in the filed of economics of education were introduced. And finally, it handled newly revised meta-analytic method and explored the adaptability of meta-analysis to economics of education research area.

SEM 기반 메타분석의 이론 및 적용 - Cheung의 접근법 중심으로

이현정,손수경,장유나,홍세희 한국심리학회 2018 한국심리학회지 일반 Vol.37 No.1

Recently, applications of structural equations and meta-analysis are increasing rapidly in many areas. The two methods have been used for different purposes in different fields. That is, structural equation modeling has been used for evaluating theoretical models, while meta-analysis has been applied to integrating individual studies. However, combing the two separate methods, meta-analysis can be done using structural equation modeling. In this study, SEM based meta-analysis was introduced and applied to empirical studies to demonstrate an example of analysis. We focused on the SEM based meta-analysis, which Cheung (2008, 2015) presented, and explained the method using Mplus. Our presentation was as follows: first, advantages of SEM-based meta-analysis were explained in detail in comparison with conventional meta-analysis. Next, SEM-based meta-analysis model(fixed-effects, random-effects, mixed-effects model) and analysis method were presented. For a real-data example, the effect sizes of empirical studies on the relationship between procrastination and academic self-efficacy were integrated and the effects of research characteristics as predictors were investigated. Finally, the implications of this study and future research directions were discussed. 구조방정식모형과 메타분석은 최근 들어 급격히 증가하고 있고 행동과학을 비롯한 모든 영역에서 활발히 적용하고 있는 방법론이다. 최근까지 두 방법론은 각자의 방식대로 발전해왔으며, 각각 서로 다른 연구에 적용되어 왔다. 즉, 구조방정식모형은 연구자가 제안한 연구모형의 검증에 활용되었고 메타분석은 단일 주제 연구들을 통합하는 데 적용되었다. 그러나 메타분석은 구조방정식의 프레임웤에서 수행할 수 있으며, 이 방법은 기존의 방식보다 더 많은장점을 지닌다. 따라서 본 연구에서는 구조방정식모형을 활용한 메타분석 방법론을 소개하고이를 실증연구에 적용하여 분석한 예시를 보여주고자 하였다. 최근 Cheung(2008, 2015)이 제시한 SEM 기반 메타분석 방법론을 중심으로 설명하였으며, 국내에서 구조방정식 분석 프로그램으로 널리 사용되고 있는 Mplus를 활용하여 분석할 수 있는 방법을 기술하였다. 구체적으로, 먼저 SEM 기반 메타분석의 장점을 기존의 메타분석과 대조하여 자세히 설명하였다. 다음으로 전통적인 메타분석 모형(고정효과, 임의효과, 혼합효과)에 대해 SEM에 기반한 접근법으로 모형을 설명하고 구체적인 분석 방법에 대해서 서술하였다. 또한 이러한 방법론을 적용한예시로서, 학업지연행동과 학업적 자기효능감의 관계에 대한 실증연구의 효과크기를 통합하고 조절변수로서 연구특성 변인의 효과를 검증하고자 하였다. 마지막으로 본 연구의 의의 및 향후 연구에 대해 논의하였다.

교육학 연구방법으로서 메타분석(Meta-analysis)의 발전과정 고찰

장덕호,신인수 한국교육과정평가원 2011 교육과정평가연구 Vol.14 No.3

이 연구는 교육학 연구방법론으로서 메타분석(meta-analysis)의 발전과정을 고찰함으로써, 기존 문헌연구의 한계 및 메타분석의 의의를 이해하고, 메타분석을 통한 교육학연구의 학문적 발전을 위한 시사점을 찾고자 하였다. 이를 위해 우선 국내·외 메타분석연구의 증가 빈도를 시기별로 살펴보고, 통시적으로 메타분석이 발전된 과정을 연구결과의 종합화라는 메타분석의 본질적 특성 측면에서 고찰하였다. 우선 Glass가 메타분석이라는 용어를 사용한 1976년 이전의 개념적 발전기에서 출발하여 이후 연구결과를 종합하려는 학자들의 노력들을 조망해 보고, 또한 1976년 이후 메타분석에 대한 비판과 그러한 비판을 극복하는 과정을 살펴보았다. 아울러 1980년대의 다양한 교재들의 출현, 1990년대의 증거기반운동(evidence-based movement) 및 현재의 국제적인 네트워크 형성과 주요한 방법론적인 이슈들을 탐색하였다. 이상의 역사적 고찰을 바탕으로 앞으로 교육학 연구에서 메타분석의 발전을 위한 시사점을 몇 가지 제시해 보았다. 첫째, 메타분석과 유사 연구방법과의 개념적 차별화가 필요하고, 둘째, 기존 연구를 양적으로 통합하고 질적으로 종합하는(synthesizing qualitative research) 노력을 통하여 학술적으로도 검증된 지식을 축적하는 노력이 필요하며, 셋째, 경험적 교육학에서 많이 쓰이는 회귀분석 연구들을 종합하는 문제, 내재적 자료(nested data) 구조를 가진 연구물을 종합하는 문제, 메타분석에서의 독립성 가정 위반, 구조방정식 모형을 활용한 연구물들을 종합하여 이론 모형을 형성하려는 접근(model based meta-analysis, Becker 2009) 등 충분한 방법론적 발전이 부가되어야 함을 제시하였다. This study attempted to understand the limits of existing research review and the importance of meta-analysis, and to suggest some implication for future development of the method by reviewing the historical development of meta-analysis as an educational research methodology. We compared meta-analysis with other methods of research review such as narrative review and vote counting. Glass coined the term of 'meta-analysis' in 1976 and we investigated the phenomena of synthesizing quantitative research from 1900s to present. Especially, we focused on methodological critiques and overcoming efforts made by several frontiers on the field. Lastly, we presented the current issues and future research implications in addition to the present problems of meta-analysis research in Korea. Conceptual differentiation between meta-analysis and other similar methods, diffusion of the importance and knowledge sharing on meta-analysis, and additional academic efforts to solve some disputes around the data analysis, assumption, and approaches should be made in further research and practices.

메타분석의 명과 암

이영호 ( Young Ho Lee ) 대한내과학회 2019 대한내과학회지 Vol.94 No.5

Meta-analysis is a statistical method that combines and synthesizes multiple studies and integrates their results. Meta-analysis increases the sample size, and in turn, the power to study the effects of interest by combining primary studies and providing a precise estimate of the effects. Data synthesized from meta-analyses are usually more beneficial than the results of narrative reviews. In a meta-analysis, the decisions are transparent, and statistical analysis yields an objective measure of the integrated quantitative evidence. The biases of narrative reviews can be limited or overcome by conducting a meta-analysis. The systematic approach and transparency in meta-analysis help to resolve conflicts and uncertainties between studies, while leading to significant conclusions. However, this method is controversial and may not always be the best tool. Moreover, meta-analysis has several shortcomings, and in some cases, it may not be appropriate. Although meta-analysis has been criticized due to its limitations, there are solutions to such problems. The aim of this review is to describe and discuss the strengths and weaknesses of meta-analysis. (Korean J Med 2019;94:391-395)

지역아동센터 이용 아동의 자아존중감 향상 프로그램에 대한 설명적 메타분석: 단일변량 및 다중변량 메타회귀분석

서인해,공계순 한국사회복지연구회 2019 사회복지연구 Vol.50 No.4

The purpose of the study is to identify the characteristics of the effectiveness of self-esteem improvement programs for the children using Community Children Centers through meta-analysis. The research subject was 39 evaluation papers on self-esteem program in Community Children Center searched in June 2018. The variables included in the analysis were 9 variables in the areas of survey assessment, program composition and program development activities. Statistical analysis was conducted by the meta-anova, univariate meta-regression and mutivaratiate meta-regression analysis using CMA(Comprehensive Meta-Analysis) 3.0. The analysis results are as follows. The average effect size of the self-esteem improvement program at the Community Children Center is .942. Analysis of meta-anova for heterogeneity of effect size showed high effect size in pretest-posttest control group design, 11 to 20 sessions, 50 to 60 minutes per session, 10 participants or less, high degree of program theory and program development process. Univariate and multivatate meta-regression analysis of the causality of effect size heterogeneity showed that the number of program goal, number of participant, level of program theory and program development activity were statistically significant; the fewer program participants, the fewer program goals, the higher the program theory and program development process, the greater the effect. The program development process is the most important variate for the incremental change in the predictive model analysis. The academic meanings of the study results was discussed, and the future directions of practices and explanatory meta-analysis was suggested. 본 연구의 목적은 메타분석을 통하여 지역아동센터 이용 아동의 자아존중감 향상 프로그램의 효과성에 관한 특징을 규명하는 데 있다. 연구대상은 2018년 6월 현재까지 검색된 39편의 지역아동센터의 자아존중감 관련 프로그램 평가논문이며, 분석에 포함된 변수는 조사평가, 프로그램 구성 및 프로그램 개발 활동 영역의 9개 변수이다. 통계분석은 CMA (Comprehensive Meta-Analysis) 3.0을 사용하여 메타아노바와 설명적 메타분석으로서 단일변량 및 다중변량 메타회귀분석을 적용하여 분석하였다, 분석 결과, 지역아동센터의 자아존중감 향상 프로그램의 평균 효과크기는 .942로 나타났다. 효과크기의 이질성에 대한 메타아노바 분석 결과, 총 회기 11~20회, 1회기 당 50분~60분, 프로그램 참여자 10인 이하, 자아존중감 단일목표, 프로그램 개발 활동과 관련된 프로그램이론 및 프로그램 개발과정 정도가 높은 경우에 효과크기가 가장 높게 나타났다. 효과크기 이질성의 인과성에 대한 단일변량 및 다중변량 메타회귀분석 결과, 프로그램 목표 수, 프로그램 참여자 수, 프로그램 이론, 프로그램 개발과정의 정도가 통계적으로 유의미하며, 프로그램 목표가 자아존중감 단일목표일수록, 프로그램 참여자 수가 적을수록, 그리고 프로그램 개발이론과 프로그램 개발과정 정도가 높을수록 효과크기가 커지는 것으로 나타났다. 자아존중감 효과크기의 예측모형에서는 4개 변수 중에서 프로그램 개발과정 정도가 모형변화량에서 가장 큰 것으로 나타났다. 끝으로 메타분석 연구결과의 내용에 대한 학문적인 논의와 지역아동센터에서의 자아존중감 향상 프로그램 및 설명적 메타분석을 위한 제언을 하였다.

Robustness of meta-analysis results in Cochrane systematic reviews: A case for acupuncture trials

Jiyoon Won 한국한의학연구원 2022 Integrative Medicine Research Vol.11 No.4

Introduction Fragility index (FI) refers to the smallest value that change statistical significance of study results. Meta-analyses of Cochrane systematic reviews are considered as the best evidence for stake holders because they enable effect size estimation that cannot be derived by individual studies, particularly in the field of complementary and integrative medicine (CIM). Thus, this study aimed to evaluate robustness of meta-analysis in Cochrane systematic reviews of acupuncture, one of the most used CIM treatment, using FI of meta-analysis. Methods Meta-analyses of acupuncture Cochrane systematic reviews with binary benefit outcome measures were searched in PubMed, Embase, and CENTRAL and subject to analysis. Randomized clinical trials (RCTs) or quasi RCTs adopted penetrating needles as treatment and compared with controls such as sham acupuncture, usual care, and active control were included. FI of meta-analyses was calculated in web (https://clinicalepidemio.fr/fragility_ma/); and fragility quotient (FQ) was calculated by dividing FI by total sample size of meta-analysis. Results Of 248 retrieved studies, 12 Cochrane systematic reviews with 48 meta-analyses were analyzed. The median FI for statistically significant and non-significant meta-analyses was 7 [interquartile range: 3 - 12] and 6 [interquartile range: 3 - 10]. FQ was similar irrespective of statistical significance. Among controls, meta-analyses adopting sham acupuncture displayed the widest range of FI and FQ. Conclusion Robustness of meta-analyses in Cochrane systematic reviews for acupuncture was similar irrespective of statistical significance. Impact of control intervention cannot be determined in this study. As FI enables intuitive interpretation, applications for CIM field can be useful.