Hi all,
I am currently investigating the differences in the use of language between online groomers and online daters.
In academic analyses, scholars have characterised online grooming as deceptive.
Deception itself, is characterised by a greater frequency of negatively emotionally loaded words and a greater infrequency of other (you, your etc.) and self (I, I'm etc.) references.
Consequently, I have predicted that online groomers conversations will likely show such markers of deception more frequently than online daters.
I have counted the amount of negative words, self references and other-references in both the online daters and online groomers conversations.
Does anyone have any advice on what statistical tests I could run to answer my hypothesis? I'd really appreciate any help!
Emma, if you want to compare the features of two different kinds of behaviour within social networks, you first need an independent way to determine which contribution belongs to which kind. I figure this could be done through an expert rating.
Once you have that "ground truth" (i.e. you know which message or which thread belongs to the categories of grooming and dating respectively), it appears to me that an analysis of variance (ANOVA) is a proper way of testing your hypotheses. You would analyze for a "main effect" of certain word frequencies.
- Badges
- Science topic
- Similar topics
- Social Science
- Sociology