|
The School consists of both theoretical sessions (in the form of lectures) and practical sessions. The practical sessions take place in a computer lab. Their aim is to complement the lectures with empirical applications (data and methodologies) related to the subjects analyzed in the theoretical sessions. The working language is English.
Participants will have the opportunity to replicate published studies, learn the instruments and apply econometric techniques in STATA to conduct their research on their own. They will have the chance to replicate existing studies and see the published econometric results. In this way they will obtain the background in STATA to conduct their meta-analysis in their field of interest.
Morning and afternoon lectures will take place at Iolida Beach Hotel.
PROGRAMME
Sunday 25/6/2017
- • Arrival of participants and welcome reception (at 20:00 at Iolida Hotel beach)
Monday 26/6/2016
- • Welcome: 9:00-9:30
- •Theoretical session [Tom Stanley] : 9:30-11:00
- • Coffee break: 11:00-11:30
- •Theoretical session [Tom Stanley] : 11:20-13:00
- •Lunch break: 13:00-14:00
- •Practical Session / Computer Labs: 14:00 - 17:00
Tuesday 27/6/2017
- • Theoretical session [Tom Stanley] : 9:30-11:00
- • Coffee break: 11:00 -11:30
- • Theoretical session [Tom Stanley] : 11:20-13:00
- • Lunch break: 13:00-14:00
- Practical Session / Computer Labs: 14:00 - 17:00
Wednesday 28/06/2016
- • Theoretical session [Tom Stanley] : 9:30-11:00
- • Coffee break: 11:00 -11:30
- • Theoretical session [Tom Stanley] : 11:20-13:00
- • Lunch break: 13:00-14:00
- • Practical Session / Computer Labs: 14:00 - 17:00
Thursday 29/6/2017
- • Theoretical session [Tom Stanley] : 9:30-11:00
- • Coffee break: 11:00 -11:30
- • Theoretical session [Tom Stanley] : 11:20-13:00
- • Lunch break: 13:00-14:00
- • Practical Session / Computer Labs: 14:00 - 17:00
Friday 30/6/2016
- • Theoretical session [Tom Stanley] : 9:30-11:00
- • Coffee break: 11:00 -11:30
- • Theoretical session [Tom Stanley] : 11:20-13:00
- • Lunch break: 13:00-14:00
- • Practical Session / Computer Labs: 14:00 - 17:00
Friday 30/6/2016
- • Farewell dinner (20:30) in a local tavern
COURSE DESCRIPTION
1. GENERAL OVERVIEW OF THE COURSE
The course deals with the issue of Meta-Analysis and Meta-Regression Analysis which are important issues of applied research. Meta-Analysis is a valuable statistical tool which permits to test the validity of the effect of a phenomenon on another phenomenon. In the case where the results of empirical studies are controversial, the investigation of these results is a crucial matter. When the effect varies from one study to the next, meta-analysis can be used to identify the reason for the variation. It is clear that the validity of a hypothesis cannot be based on the results of a single study. Meta-Analysis is the statistical procedure for combing data from multiple studies. It is a type of systematic review that employs the full range of statistical methods to summarize and to help researchers to understand what an entire empirical literature means. Meta-analysis can play a key role in planning new studies. It can identify which questions have already been answered and which remain to be answered. The meta-analysis is useful because it puts all available data in context and explores the validity of the existed empirical results.
Meta-regression analysis is designed to model the effects of observed econometric specifications. It is a systematic and comprehensive review of all comparable econometric findings. It models any potential bias or systematic variation, thereby explaining the excess variation always observed among reported results.
The aim of this course is to provide theoretical and empirical knowledge of meta-analysis and meta-regression. Furtherer more it can help the participants to be familiar with the meta-analysis and meta-regression techniques. The participants will learn how can they can use the econometric – statistical package STATA to apply meta-analysis and meta-regression analysis.
2. CONTENTS OF THE COURSE
The course consists of a number of lectures based on theoretical issues concerning the topic of meta-analysis and meta-regression analysis which will take place during the mornings and empirical applications of the theoretical issues taught in the morning. The empirical applications will take place during the afternoon sessions.
2.1. Theoretical lectures
The first lecture will focus on the history of meta-analysis and offers a framework that can be used to summarize and qualifying estimates of policy-relevant parameters. It offers strategies for identifying and coding empirical economics and business research. This lecture will focus on the collection of the data that defines meta-analysis and more specifically it will discuss where to collect data and what information to collect. In other words we will discuss how to search and code research for a meta-analysis.
The second lecture discusses simple statistics and graphs that have been found useful in summarizing research. Furthermore it introduces meta-regression methods that identify and correct publication selection bias. It focuses on the description of these data, presenting alternative ways of summarizing research findings. Furthermore this lecture will also discuss the publication bias and how can we correct it.
The third lecture shows how multiple meta-regression analysis is often employed to explain economic research and its excess heterogeneity. It will discuss the heterogeneity and the existed variation in any area of economic research. In this lecture we will show how to accommodate and explain this excess research variation using “multivariate” meta-regression analysis.
The fourth lecture offers a theory of meta-regression analysis and a rigorous demonstration that study quality need not affect the findings of a MRA. It more deeply explores MRA models for within-study dependence and publication selection. The purpose of this lecture is to delve bit deeper into the statistical foundation of MRA. We will present a theory of MRA directly derived from econometric and statistical theory. WE will show how MRA results are entirely unaffected by issues of observed and unobservable study quality when properly modeled.
The fifth lecture describes alternative objectives for performance systematic reviews and how they shape the way MRAs are conduced or applied. It also considers additional complexities to the structure of empirical research and how to model them statistically. The purpose of this lecture is to explore some of the alternative applications of MRA in economics. We discuss the choices of MRA variables when there are more variables than observations. Then we discuss the use of MRA for identifying exclusion restrictions, we look at the forecasting performance of MRA in both time and space. Finally we will investigate the treatment of effect sizes that involve MRA models with interaction and non-linear terms.
The lectures will be based on the book Stanley, T.D and Doucouliagos, H. (2012). Meta-Regression Analysis in Economics and Business, London : Routledge (see more https://www.routledge.com/Meta-Regression-Analysis-in-Economics-and-Business/Stanley-Doucouliagos/p/book/9781138241145)
The econometric and statistical package STATA will be used in the practical sessions.
2.2. Empirics – Laboratories
First day
The Course will start with an introduction in the meta-analysis definition, its purpose, its use and applications to various fields of study in economics, business and other sciences. The participants will learn the historical roots of meta-analysis, its advantages in modern science and the problems that tackle regarding the publication bias, problems related to the statistical approach and problems arising from agenda-driven bias. During this first section, the participants will acquaint themselves with the first steps in meta-analysis with practical examples.
Second day
Then it will proceed to the appropriate protocols of the searching process and the identification of the meta-analysis data. The identification of the studies which constitute the meta-sample is a very important step in the meta-analysis and the participants will learn the protocols: a) of the inclusion of a study and b) the exclusion rules.Moreover, the coding of the studies, which is the most time-consuming step in the meta-analysis, will be analyzed in every detail, using EXCEL and STATA.
Third day
The third part regards the application of the statistical tools and econometric methods which are necessary when we conduct a meta-analysis and participants will learn how to use the appropriate commands in STATA.They include the FAT-PET-PEESE approach to deal with publication selection bias and meta-regression econometric models and methods (such as FE, RE, WLS and others) which are used in the meta-regression analysis.
Fourth day
The fourth part will provide various aspects of a robustness analysis as it will be demanded by the reviewers and editors in order to ensure that the central findings of a meta-analysis are robust to the model variations.Furthermore, in this session common problems in meta-analysis will be presented and in which ways we can deal with them. There will be practical examples with all the necessary tools using STATA.
Fifth day
The Course will finish by showing how the meta-analysis techniques can be used for policy analysis using existing meta-studies as examples.The participants will have the opportunity to replicate published studies, learn the instruments and apply econometric techniques in STATA to conduct their research on their own.They will have the chance to replicate existing studies and see the published econometric results. In this way they will obtain the background in STATA to conduct their meta-analysis in their field of interest.
3. BASIC OUTLINE
The course will be based on instruction, over a week. Each day there will be instruction on the main methods that we will cover for a total of 15 hours of instruction. There will also be additional 15 hours of hands on empirical applications of the methods covered using data that will be provided by the instructor.The main headings of the topics that will be covered are as follows:
- Identifying and coding meta-analysis data
- Publication Bias
- Explaining Economics Research
- Econometric Theory and meta-regression analysis
- Further topics in meta-regression analysis
The main headings of the topics which will be covered in the afternoon classes are as follows:
- Introduction and First Steps in Meta-Analysis
- The Search and Coding Strategy
- The Meta-Regression Methodology
- Robustness Analysis
- “Learning by Doing” – Replications of previous Meta-Studies
|
