4  Planning the Synthesis

Where in primary research, the term used for the process by which one arrives at answers to a research question based on the collected data, in a systematic review, this process is called synthesis. This nicely captures the aim to combine information from multiple sources to yield a (hopefully coherent) overview.

Synthesis of systematic review results shares many characteristics with analyses in primary research. In both cases, the process consists of dozens to thousands of subjective decisions; in both cases, the process can be challenging and complicated; and in both cases, the way the process is challenging and complicated depends on the type of data being processed. For systematic reviews, syntheses tends to be easier as sources are more similar.

For example, a relatively simple synthesis could be a meta-analysis of studies that all have the same design and used the same measures for the same variables, for example when aggregating randomized controlled trials for a specific COVID-19 vaccination. In such scenarios the studies often share the same ontological and epistemological perspectives, and a single effect size estimate can often be extracted from each source, all in the same metric. In addition, variation in extracted effect sizes most likely reflects sampling variability and contextual factors. Both can be statistically modeled (assuming those contextual factors were extracted).

An example of a relatively complicated synthesis is a systematic review that provides an integrative overview of a topic over multiple study types, for example everything that is known about why people