Sample biases in humanistic networks

Scott Weingart
Indiana University
Inforation Science / History and Philosophy of Science

The last few years have seen the use of quantitative methodologies in the humanities go from a steady trickle to a thundering flood. These are dangerous waters, however, as the impetus to use tools created for another job brings with it certain caveats that have not yet fully been explored. One such powerful tool is network analysis, which humanists have increasingly used to explore themes such as social interaction and communication, literary relatedness, and character development. Because data is drawn from humanistic sources so very differently than from the disciplines in which network analysis was developed, extra attention must be paid to sampling biases and what conclusions can be drawn from the analysis. This paper lists some of the many sampling biases that can appear when collecting humanistic data, and then explains their ramifications for drawing conclusions from network analysis.

Back to Networks and Network Analysis for the Humanities: Reunion Conference