Too Positive to Be Tweeted? An Experimental Investigation of Emotional Expression on Twitter and Instagram

Abstract

The literature on emotional expression on Social Network Sites (SNSs) is still in its infancy. It is assumed that SNSs are subject to a positivity bias: individuals share positive aspects of their lives on SNSs rather than negative ones. However, sentiment analysis studies have shown that this bias might not be appropriate for all SNSs, particularly Twitter. This research aimed to understand how the emotions of a message impact the choice of the SNS used to publish it. Four pre-registered experimental studies were conducted (N = 449). Participants were presented with a message–text and image–and asked if it were most likely to be published on Twitter or Instagram. The emotional valence of this message was manipulated in experiments 1a and 1b, as well as emotional arousal in experiments 2a and 2b. The results indicated that Instagram was preferred for positive messages. But the platform was also chosen for messages displaying low negative emotions, such as boredom or lassitude. Twitter was associated with messages displaying highly negative emotions, such as anger or distress. This research emphasizes the social norms that govern both platforms and demonstrates the interdependence between SNSs architecture, user motivations, and social context.