In this chapter, a forward-looking guide is provided for conducting meta-analysis (MA) using observational data, illustrated by using two cross-disciplinary case studies: (a) Counting Species Living on Coral Reefs; and (b) Big Data and Firm Performance. As shown in this chapter, literature reviews can be structured as multi-method and/or mixed method: selecting studies by means of scoping, then systematic review; clarifying eligibility by means of qualitative, narrative, or model-centric review; and ending with realist review. Expert knowledge might also help in interpreting and adjusting review findings. For MA in emerging fields, dominated by observational studies, this more nuanced form of review provides a foundation for collating and preparing information prior to the quantitative component of MA. In the guide presented in this chapter, the MA appraisal tool, AMSTAR2, is adopted to group criteria into seven core domains, which apply flexibly to our non-randomised case studies: (a) an MA, using Bayesian and risk assessment analysis, of a census of studies accumulated in a global database and supplemented with expert assessments, and (b) a classical random-effects MA of observational studies that each use Structural Equation Models. Viewing MA as a mixed (quantitative and qualitative) method provides a wider array of options, compared with the standard – a systematic review of randomised studies, which is typical in medicine. It is suggested this mixed-methodsapproach is more suitable in many fields, especially multi-disciplinary fields such as built environment.