Efficient coding of eyewitness narratives: A comparison of syntactic unit and word count procedures

Abstract

Researchers rely on two types of coding systems to evaluate eyewitness narratives. Features-of-events checklists offer coding simplicity but are impractical when target events are complex. Units-of-information (UOI) systems score all information reported, regardless of event complexity, but are difficult to implement. To test whether simpler systems would alter conclusions about memory performance, transcripts from children (3-8 years of age) who participated in an eyewitness study, originally coded using a UOI system called syntactic units (SU), were recoded using two word count procedures. Correlations between SU, modified word count, and raw word count values were high, and the proportion of information that was inaccurate was comparable across systems. Considering their high interrater reliability, procedural simplicity, and convergence with SU coding, word count procedures are efficient alternatives to UOI coding.