The common denominator of the special issue is the interest in using natural language processing (NLP) to move psychological methods forward and the use of rating scales in the assessment of personality and attitudes.
A total of ten peer-reviewed articles presents a range of applications of text algorithms that supplement, criticize or even replace the use of traditional rating scales in psychological measurement. The articles span a wide range of psychological topics from organizational behavior through clinical psychology and social media to voting behavior in national elections.
Picture: “With today's technology, we are able to use text algorithms instead of questionnaires", says Professor Jan Ketil Arnulf.
“To our knowledge, this special issue of research articles within this specific field is the first of its kind in the research literature. For the first time we have a cross-national, cross-institutional team of researchers predicting that traditional questionnaires could be replaced by digital text analysis, says Professor Jan Ketil Arnulf from BI Norwegian Business School.
The researchers present their approaches and results in a forthcoming special issue of the open-access journal “Frontiers in Psychology”. The cross-national team is situated at the University of Colorado at Boulder, the University of Texas at Tyler (USA), The University of Lund (Sweden), The University of Edinburgh, and BI Norwegian Business School (Norway), all being pioneers in developing their own distinct approaches.
“With today's technology, we are able to use text algorithms instead of questionnaires. The body of research presented here shows that freely entered texts on any topic can be used to assess psychological states and predict behavior as good as, and sometimes better than through the prevalently used scales that you often find in a survey”, says Arnulf. He adds: ”The other studies published show that traditional data from questionnaires can be culture blind, and that more information can be extracted through the use of text algorithms.”
The research has several interesting implications, including how digital language analysis has advanced as a tool for social sciences. Already back in 2014, Arnulf and his colleagues published a study showing how historic survey data can be used to predict future survey responses, even without asking real respondents first.
Reference
The research were mentioned in an article at Forskning.no (in norwegan) 08.09.21.