Representative Literature Reviews Using Constrained Snowball Sampling and Citation Network Analysis

Abstract

We present here a new method for literature reviews that utilizes citation network analysis from academic databases. Special attention is paid to the construction of the citation network in order to ensure that the sample contains key publications in the field. Keyword searches in databases can be misleading because keywords are not used consistently across publications. Snowball sampling is onerous because the sample grows at an exponential rate. Constrained snowball sampling (limiting the percentage of articles collected at each level) is presented as a feasible and effective alternative that allows for robust citation analysis using a small fraction of the data. The methodology is outlined here. Free software for the R environment is also available.

Keywords: citation analysis, literature review, snowball sampling