Romantic Self-Presentation in Date-Me Docs, 2024

This study employed natural language processing (NLP) techniques to explore shared themes and variations in romantic self-presentation strategies in “Date-Me Docs”— recently popularized web-based longform personal advertisements for online dating. 251 usable profiles were sourced from a publicly available directory of Date-Me-Docs (https://dateme.directory/) after omitting 108 documents which were inaccessible or blank (e.g., due to a change in site permissions), one document which contained less than 50 characters, and four documents which were not written in English. The average word count for the dataset is 972 (SD = 1006), with an average of 27 (SD = 18) words per sentence. The authors are 59.6% male (36.4% female, 4% non-binary; 0.40% undisclosed), 73.9% monogamous (18.3% open to both monogamy and polyamory, 7.8% polyamorous; 8.4% undisclosed), 84% straight (i.e., “M” interested in “F” or “F” interested in “M”) and have a mean age of 32.5 (SD = 7.3) years. In terms of childcare intention, 46.3% of authors indicated they “want kids” (32.9% “might want kids”, 17.6% “do not want kids”, 3.2% “has kids”; 13.9% undisclosed). Authors also provided their location, location flexibility (i.e., how willing they are to move), and community affiliation (Effective Altruism (EA), Tech, and/or Rationalism). After pre-processing the unstructured text data, the corpus was converted into a document-feature matrix where the rows represent documents, columns represent terms (768 in this dataset after cleaning), and values represent the frequency of the term’s appearance in a specific document. This matrix contained the features (i.e., document terms) that were used as variables in topic modeling with Latent Dirichlet Allocation (LDA) to identify shared themes and topic distributions over documents. Topic modeling revealed ten distinct themes spanning personal preferences, self expansion and growth, open communication, and social responsibility. The topics (and their definitions; % of total) were: Altruism (Desire to contribute to societal well-being; 14.3%), Expectations (Specific expectations for a long-term relationship; 11.2%), Collectivism (Strong connection with others & the world; 10.6%), Dialogue (Intellectual curiosity & exchange; 10.6%), Lifestyle (Conventional life choices, especially children, monogamy, and substance use; 9.5%), Learning (Learning & experiencing aspects of the world; 9.1%), Socializing (Casual socializing, including entertainment; 9.1%), Hobbies (Creative & fulfilling leisure activities; 9%), Goals (Building towards a better future; 8.4%), and Activity (Physical activity, exercise, and traveling; 8.3%). Other descriptive information about the documents were also gathered (e.g., punctuation, numbers, symbols, URLs, tags, emojis), and each document was rated on various positive and negative emotions (e.g., anger, anticipation, disgust, fear, joy, sadness, surprise, trust). This study was exempt from ethics board approval, as the data were publicly accessible.