Tell-Tale Words: Linguistic Cues Used to Infer the Expertise of Online Medical Advice

This article analyzes the linguistic cues used by naive perceivers to assess the expertise of online medical advice. We develop a theoretical framework of linguistic correlates to perceived expertise and test it on a corpus of 120 online medical advice messages, written by either medical doctors or laypersons. Linguistic Inquiry and Word Count (LIWC) analyses show that messages were perceived as more expert if they contained more words (an indicator of uncertainty reduction), fewer I-pronouns and anxiety-related words (indicators of psychological distancing), and more long words and negations (indicators of cognitive complexity). These linguistic cues explained over a third of the variance in expertise ratings. Although unaware of the author of each message, perceivers were able to discern between messages written by doctors versus laypersons. However, only long words were helpful in making this distinction. Results advance the literature on linguistic correlates of psychological processes.