Meta regression application for detecting publication bias and variation of results in economic research

Abstract

This paper briefly aims to introduce the meta regression application or better known as the Meta Regression Analysis (MRA). This tool is widely used to detect publication bias in the publication of scientific articles, including in the field of economics. Publication bias occurs because researchers tend to publish their research results that support certain hypotheses (significant). Research results that are not significant have a relatively small probability of being published. Publication bias can be detected by means of a relatively simple regression test. A set of quantitative studies on the same topic is tabulated and analyzed using meta regression. Meta regression is not only used to detect publication bias. This analysis tool can also explain the source of variations in the results of several studies due to differences in location, research year, methods, gender and measurement methods used by researchers. This paper uses meta-regression to examine the variation in outcomes in an intergenerational transfer study. The results showed that location differences had a significant effect on transfers. Meanwhile, the research period and the variation of methods used in the previous study did not have a significant impact.