Applying Human Language Technology in Survey Research

Abstract

The term Human Language Technology (HLT) refers to a broad class of methods for automatically analyzing speech and text. This may include analysis of the topics, sentiments, relationships, or other subjects represented in the text. While early attempts at developing this technology were crude and generally ineffective, modern computing power and new methods have brought these techniques to the brink of broad application in many areas. This chapter overviews HLT and discusses the promises and pitfalls of applying these methods to open-ended responses to questions in survey data. While challenges still remain, the automated text analysis of open-text responses provides a potentially substantial opportunity for researchers.