Abstract: Recent advances in natural language processing methods provide exciting new ways to approach the analysis of psychological open-ended data. However, many of these methods have not been translated to the needs of psychologists working with text. I’ll introduce methods to create and validate extensive dictionaries of psychological constructs using platforms such as Wordnet (a network of semantic relations) and techniques such as word embeddings (Nicolas, Bai, & Fiske, in prep.). Our proposal consists of a series of steps that utilize an R package we developed to access resources such as Wordnet in order to expand an initial list of “seed words” by obtaining their synonyms, antonyms, and other semantically related terms. Then, we introduce word embedding techniques as a method to evaluate the dictionaries reliability. Finally, we propose avenues to examine the convergent validity of dictionaries created with this method. We illustrate this process by creating novel dictionaries of stereotype content, an important construct in social psychology that currently lacks specialized and validated dictionaries for its measurement in text.