Chatting while you work

Abstract

Purpose To understand chat reference user needs at the University of Arizona Libraries, this study aims to analyze the differences in READ (Reference Effort Assessment Data) scores and content from two different chat reference origins: the library website and, course sites within the campus course management system, Desire2Learn. Design/methodology/approach The authors have used a mixed methods approach, whereby ordinal regression has been performed on READ scores and grounded theory on chat transcripts in the analysis of themes. This content comparison offers insight into the variations of how and why users ask for help at different chat reference service points. Findings The results of the content comparison inform the understanding of the reasons behind library users’ needs and questions at various stages in their academic workflow. The analysis also has implications for future chat reference training, as well as underscores the need for continued integration of library services and resources into online educational technology systems. Practical implications The findings support the importance of collocating research/reference support and course content, and underscore the value of library integration with course management systems. The analysis also serves as user-needs assessment data, with the potential to inform library services beyond reference, such as information literacy instruction or collection development. Originality/value This study aims to fill a gap in the literature by investigating differences in chat reference transcripts based on chat reference origin.

Does the READ Scale Work for Chat? A Review of the Literature

“…Larson et al (2014) highlighted consistent application of the scale as an issue when a chat service employed both local and consortial chat agents. However, the rest of the articles did not have any crossinstitution norming, though five (in addition to Gerlich and Berard), performed some form of inter-rater norming for their own data (Belanger et al, 2016;Kemp et al, 2015;Keyes & Dworak, 2017;Stieve & Wallace, 2018). In the remainder, no mention was made of any norming of READ Scale ratings, other than Warner et al (2020) who explicitly stated that interrater reliability was not tested as part of the study.…”

“…While the READ Scale is a 6-point scale, the 1-4 range of the scale was heavily used, while the 5-6 range was not. Belanger et al (2016), Gerlich andBerard, (2010), Kohler (2017), Stieve and Wallace (2018), and Ward and Jacoby (2018) found that zero chats were rated a 5 or 6. Kayongo andVan Jacob (2011), Cabaniss (2015), and Mavodza (2019) found that the majority of chat interactions fell within the 1-3 range.…”

“…Kohler (2017) reported Rockhurst University used the READ Scale to understand "the effort, skills, knowledge, teaching, techniques, and tools" used by librarians (p. 138). Keyes and Dworak (2017), Mavodza (2019), Stieve and Wallace (2018), Valentine and Moss (2017), and Warner et al (2019) did not address whether they were assessing the complexity of either questions or answers.…”

Constructing citations: reviewing chat transcripts to improve citation assistance as a service

“…Articles reporting the results of research on digital reference including chat have been published since the 1990s (Matteson et al , 2011). Topics covered in the last few years of literature have ranged from teaching or instruction in chat (Dempsey, 2016, 2017; Hervieux and Tummon, 2018; Jacoby et al , 2016; Schiller, 2016), the types of questions asked in a chat service (Bourgeois and Bealer, 2020; Brown, 2017; Chen and Wang, 2019; Mavodza, 2019; McKewan and Richmond, 2017; Ozeran and Martin, 2019; Stieve and Wallace, 2018), staffing hours or needs, patron satisfaction with the service (Brown, 2017; Mungin, 2017), the role of cobrowsing (Wan et al , 2009) and communication issues (Westbrook, 2007). Probably the largest and most prolific area has involved service quality.…”