Meta-analysis is an objective examination of published data from many studies of the same research topic identified through a literature search. Through the use of rigorous statistical methods, it can reveal patterns hidden in individual studies and can yield conclusions that have a high degree of reliability. It is a method of analysis that is especially useful for gaining an understanding of complex phenomena when independent studies have produced conflicting findings.
Meta-analysis provides much of the underpinning for evidence-based medicine. It is particularly helpful in identifying risk factors for a disorder, diagnostic criteria, and the effects of treatments on specific populations of people, as well as quantifying the size of the effects. Meta-analysis is well-suited to understanding the complexities of human behavior.
There are well-established scientific criteria for selecting studies for meta-analysis. Usually, meta-analysis is conducted on the gold standard of scientific research—randomized, controlled, double-blind trials. In addition, published guidelines not only describe standards for the inclusion of studies to be analyzed but also rank the quality of different types of studies. For example, cohort studies are likely to provide more reliable information than case reports.
Through statistical methods applied to the original data collected in the included studies, meta-analysis can account for and overcome many differences in the way the studies were conducted, such as the populations studied, how interventions were administered, and what outcomes were assessed and how. Meta-analyses, and the questions they are attempting to answer, are typically specified and registered with a scientific organization, and, with the protocols and methods openly described and reviewed independently by outside investigators, the research process is highly transparent.
Meta-analysis is often used to validate observed phenomena, determine the conditions under which effects occur, and get enough clarity in clinical decision-making to indicate a course of therapeutic action when individual studies have produced disparate findings. In reviewing the aggregate results of well-controlled studies meeting criteria for inclusion, meta-analysis can also reveal which research questions, test conditions, and research methods yield the most reliable results, not only providing findings of immediate clinical utility but furthering science.
The technique can be used to answer social and behavioral questions large and small. For example, to clarify whether or not having more options makes it harder for people to settle on any one item, a meta-analysis of over 53 conflicting studies on the phenomenon was conducted. The meta-analysis revealed that choice overload exists—but only under certain conditions. You will have difficulty selecting a TV show to watch from the massive array of possibilities, for example, if the shows differ from each other in multiple ways or if you don’t have any strong preferences when you finally get to sit down in front of the TV.
A meta-analysis conducted in 2000, for example, answered the question of whether physically attractive people have “better” personalities. Among other traits, they prove to be more extroverted and have more social skills than others. Another meta-analysis, in 2014, showed strong ties between physical attractiveness as rated by others and having good mental and physical health. The effects on such personality factors as extraversion are too small to reliably show up in individual studies but real enough to be detected in the aggregate number of study participants. Together, the studies validate hypotheses put forth by evolutionary psychologists that physical attractiveness is important in mate selection because it is a reliable cue of health and, likely, fertility.