Researchers are often confronted with numerous results that do not always provide them with conclusive answers. By synthesising results of smaller studies that have inconclusive outcomes, it is more likely to produce more robust and accurate estimates and stronger inferences.

The term ‘meta-analysis’ was coined in 1976 by statistician Gene Glass to refer to the integration of research findings from a collection of analytic results elicited from individual studies.

Meta-analysis grew out of necessity when the amount of research studies in various fields was growing so fast that researchers became conscious that it was necessary to come up with a technique that combines various study results.

Thus, a number of researchers began to develop a number of distinct methods in order to combine these related study results.

The birth of meta-analysis was marked when Glass delivered a speech on his finding of a new quantitative scientific analysis.

His seminal contribution was considered as a breakthrough in all sciences and since then, meta-analysis has grown at a rapid pace, and its application has spread from one discipline to another.

The birth of meta-analysis was marked when Glass delivered a speech on his finding of a new quantitative scientific analysis

Glass used standard statistical procedures, mainly analysis of variance and regression analysis, but he applied these techniques to summary statistics, rather than to raw data. Other meta-analysts contributed to develop their own statistical methods to improve and simplify the statistical procedures.

The work of Glass was extended by John Hunter and Frank Schmidt, who developed a set of general procedures for test validity studies.

Their research on validity generalisation shared many similarities with Glass’s research on meta-analytic methodology.

Since 1976, the meta-analytic methodology has continued to grow. Several publications outlining meta-analysis approaches were published in the early 1980s.  Some publications include those of Glass, McGraw and Smith in 1981; Hunter, Schmidt and Jackson in 1982; and Rosenthal in 1984. In 1985, Hedges and Olkin published a book on meta-analysis based on classical statistics.

This was crucial in instilling formality and statistical validity in the approach, as well as serving as a springboard for further meta-analytic developments. 

In the last 25 years, meta-analysis has evolved from a small number of contributions by statisticians to synthesise study findings, to a comprehensive statistical technique that provides more precise estimates of effect sizes and increase the generalisability of individual study results. Its significance in the social sciences and education is negligible in comparison to its impact in medicine, where hundreds of meta-analyses have been published. All in all, meta-analysis can be seen as the way of the future when it comes to synthesising study findings.

Liberato Camilleri is a statistics professor at the University of Malta.

Sound Bites

•        The author of this page (Liberato Camilleri), together with other researchers, applied meta-analysis to combine six studies to investigate the impact on the resilience, prosocial behaviour and social emotional learning of a group of students who received training and promotion of protective factors in mental health. The study concluded that the training had a significant positive impact on the students’ prosocial behaviour and social emotional learning, but had a smaller impact on their resilience. Moreover, meta-analysis revealed that resilience does not mediate the relationship between the two other variables, which implies that prosocial behaviour improves with better social emotional learning skills, but resilience has negligible impact on this positive relationship. The results are in the paper: ‘Meta-Analysis to Study the Promotion Effect of Protective Factors in Mental Health, Proceedings of the 2023 ISC Conference’, p. 72-76.

For more science news, listen to Radio Mocha on www.fb.com/RadioMochaMalta/.

DID YOU KNOW?

•        Between 2009 and 2022, the percentage of Maltese students watching television and reading the newspaper (including online versions) to inform themselves about national and international news decreased by 20% and 9% respectively.

•        Between 2009 and 2022, the percentage of Maltese students who participate in voluntary groups to help the local community increased by 10%, while the percentage of students who participate in religious groups or organisations decreased by 5%.

•        Between 2016 and 2022, the percentage of Maltese students who use the internet to find information about political or social issues increased by 6%, while the percentage of students who intend to vote in future elections decreased by 13%.

•        Between 2016 and 2022, the percentage of Maltese students who trust the government, parliament, courts of justice and traditional media decreased by 16%, 15%, 12% and 11% respectively.

For more trivia, see: www.um.edu.mt/think.