Facebook graph search for refined screen-based data collection in CMC: A pilot study for Fiji English

• Facebook Graph Search is introduced for linguistic research.
• A pilot study in CMC and Fiji English demonstrates the use of Graph Search.
• Age and context (Facebook wall post vs. comments) influence language choice.
• Age influences wall content, post format and comment structures in Fiji.
• Non-standard features typical of CMC show local influences.

In sociolinguistics, the advantages of web-based research are often lost because it can rarely satisfy the need for social variables without additional offline data. Facebook Graph Search is a new query tool, which allows highly specific retrieval of Facebook users and can thus mitigate this dilemma. In this paper, Facebook Graph Search will be introduced and a pilot study in computer-mediated communication and Fiji English will demonstrate the suitability of Graph Search for exploratory research in a speech community. The pilot study shows that there are differences in language choice based on age and context of use: English is preferred over local indigenous languages by younger users and in wall posts. The data further revealed not only the use of typical CMC features, but also locally influenced features and spellings.