Data and Code

Motivation:

A common set of questions when studying norms are, "how do you know a norm exists?" and "how do you know that you are considering all the relevant norms?". The proposed method will offer a more complete and rigorous way to surface norms from text.

The secondary motivation for this work is to improve the tools researchers have to study organizations. Organizations, and corporations in particular, are very challenging to study because access is restricted as a result of confidentiality and competition. This method uses publicly available text data to offer insight into organizational norms and attitudes. Better understanding in this domain can improve future experimental designs and studies on organizational behavior and change.

Purpose:

Build an algorithm that will identify text (sentences) that contain norms and values in natural language.

Evaluate multiple modeling methods to develop the most accurate algorithm. The two methods that I will test are: rule-based modeling and machine learning algorithms (using the bag of words method).

Materials: Preregistration available on OSF

Materials: Dataset and codebook available on OSF