A Quantitative Literature Review Approach: Meta Analysis

Meta Analysis
Today, Big Data has brought us into a new era in which managers are able to make better decisions, politicians are able to select better campaign strategies and psychologists are able to better analyze and test behavioral theories more accurately.
Big Data consists of a large number of both variables and observations, often in the thousands or even millions. By using this data, it is possible to better select crucial variables and obtain more precise results. However, analyzing Big Data requires complex and advanced computations given the characteristics of Big Data, namely Volume, Velocity and Variety (Laney, 2001). Further, these extensive data sets are often very costly or unavailable to the non professional user.

This is where meta-analysis might come in. In comparison to using Big Data sets directly, meta-analysis uses and combines available results and thus does not suffer from the aforementioned problems. Meta-analysis has been successfully applied in several fields, such as medicine, psychology, educational research, ecology as well as empirical social sciences: e.g. marketing, economics.

What is Meta Analysis?

Meta-analysis is a quantitative literature review approach which statistically combines the results of several studies addressing a shared research hypothesis. As an illustrative, albeit over-simplified example, let us consider the user rating of the new Star Wars movie on five different user rating web pages. Meta-analysis then combines those five ratings in order to get a more precise ranking and, in addition, it might reveal systematic differences in the preferences of the users of the respective pages.

Characteristics of Meta Analysis

It is thus a study of studies. In other words, it averages results across studies in order to reduce problems associated with the sampling error at the individual study level leading to frequent Type II errors, i.e. false negatives. These false negatives could lead to incorrect decisions. For example, a doctor under a false negative of a medical screening for a disease would give a patient the incorrect assurance that he or she does not have a disease when he or she in fact does. A manager may forgo the investment opportunity because he has been told that the project is not profitable, when the project actually is profitable.

Furthermore, meta-analysis, being a quantitative literature review, allows us to measure and correct for publication bias in the respective literature (e.g. Bom and Rachinger, 2017). In this aspect, it fundamentally differs from an individual study using Big Data. Publication bias refers to a systemic bias in published research in a field due to over-reporting of positive results. In the field of medicine, publication bias often arises from the funding source of drug trials. A 2007 article compared over 500 studies and found those funded by pharmaceutical companies were 50 percent less likely to report negative effects of corticosteroid drugs than those not funded by pharmaceutical companies. Moreover, in 2010, researchers identified that out of a total of 500 trials for five major classes of drugs, 85 percent of the industry-funded studies were positive (i.e., those drugs were found to be effective), compared to 50 percent of the government-funded trials.

Overall, meta-analysis is a powerful and useful statistical tool which can complement costly and complex big data analyses and, in addition, help reconcile seemingly different research results.

By Dr. Ming-Jin Jiang, Marbella International University Centre
& Dr. Heiko Rachinger, Department of Applied Economics, Universitat de les Illes Balears

Please follow and like us:

Be the first to comment

Leave a Reply

Your email address will not be published.



Enjoy this Website? Please spread the word :)

  • Facebook
  • Twitter
  • Google+
  • LinkedIn
  • Instagram
  • YouTube