Predicting comments on Facebook photos: Who posts might matter more than what type of photo is posted

Addict Behav Rep. 2022 Feb 25:15:100417. doi: 10.1016/j.abrep.2022.100417. eCollection 2022 Jun.

Abstract

The number of likes and comments received to social media posts and images are influential for users' self-presentation and problematic Facebook use. The aim of this study was to highlight the most relevant factors predicting the popularity (i.e., the probability to receive at a least a comment) of Facebook photos based on: (i) Facebook user-related features; (ii) Facebook photo-related features; and (iii) and psychological variables. A mixed approach was used, including objective data extracted from Facebook (regarding users' characteristics and photo features) as well as answers to a questionnaire. Participants were 227 Facebook users (M = 25.01(1.05) years). They were asked to answer a questionnaire and provide a copy of their Facebook profile data. A total of 180,547 photos receiving a total of 122,689 comments were extracted. Results showed that user-related features (Facebook network and activities) were the most relevant in predicting image popularity accurately. It seems that who posts a Facebook photo matters more than the type of photo posted and the psychological profile of the user. Results are discussed within a psychological perspective. Future research should look at the sentiment (positive vs. negative) of the comments received by different types of photos. This is the first study exploring what makes a Facebook photo popular using objective data rather than self-reported frequency of Facebook activity only. Results might advance current methods and knowledge about potential problematic behaviors on social media.

Keywords: Comment; Facebook; Machine learning; Objective data; Photo.