Health & Environmental Research Online (HERO)


Print Feedback Export to File
5029242 
Book/Book Chapter 
Chapter 10: Maximum likelihood approaches to meta-analysis 
Mengersen, K; Schmid, CH 
2013 
Princeton University Press 
Princeton, NJ 
Handbook of Meta-analysis in Ecology and Evolution 
English 
is part of a larger document 5029185 Handbook of Meta-analysis in Ecology and Evolution
This chapter discusses an alternative, more general approach based on maximizing the likelihood of the data; that is, for a model with unknown parameters, finding the parameter values that are “most likely” to generate the observed data set. For example, suppose that we are interested in the overall age of a population of people, and we have only two options: 30 years and 100 years. Now suppose that we observe a random sample of people with average age of 20 years. Then this sample is more likely to have been observed if the overall age of the population is 30 years as opposed to 100 years. Of course, we have made many assumptions in this example, and in practice there is more information about parameter values and the observed sample. The chapter describes this more general setup in the context of meta-analysis and then gives some worked examples.