Traditionally, when social scientists wanted to analyze large amounts of documents, they have resorted to using manual coding techniques. This process can be made easier by using machine learning approaches. One such approach, called topic modelling, can find which words commonly occur together and in doing so provide the researcher with semantically coherent topics. This thesis utilizes topic modelling to investigate Nordic Psychology, a psychology journal published in the Nordic languages. Articles published between 1949 and 2005 are examined to map out how discourse has changed during the second half of the 20:th century. Psychology textbooks and researchers active in the late sixties frequently refer to something called the cognitive revolution taking place. Accounts of this revolution paint a picture of something resembling a paradigm shift. This thesis therefore sets out to look for signs of the cognitive revolution being a paradigm shift. The topic model used in this thesis does however not find the traces of a paradigm shift within the dataset, suggesting that if a paradigm shift did take place, it was not reflected in the Nordic Psychology journal.