Success in the abstract: exploring linguistic and stylistic predictors of conference abstract ratings

The purpose of this study is to examine the relationship between features of conference abstracts and reviewer scores. Previous researchers have described the discourse structures of abstracts (e.g., Blanco, 2005) and have reported on the characteristics of successful abstracts (Cutting, 2012; and Kaplan et al., 1994). We build on this research by considering both accepted and rejected abstracts, investigating the relationships between a wide range of linguistic and stylistic features and reviewer scores. The corpus used in this study consists of 287 abstracts submitted to the 2009 Second Language Research Forum. Each abstract was scored according to four criteria by three independent abstract reviewers. Additional coding was carried out by the researchers to measure ten stylistic features (e.g., word counts and number of citations). The Biber Tagger was used to annotate grammatical and lexico-grammatical features. A series of t-tests and Pearson's correlations revealed a number of significant relationships between linguistic and stylistic features and abstract ratings. The results of a stepwise multiple regression showed that 31 percent of the variance in abstract scores can be predicted by a combination of the following: more words, citations, the presence of a Results section, more nouns, no errors, and fewer first-person pronouns. In addition to presenting a comprehensive description of the stylistic and linguistic characteristics of conference abstracts, the findings from this study offer practical recommendations for successful abstract writing in applied linguistics.