Iterative Chat Transcript Analysis: Making Meaning from Existing Data
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Objective – In order to better contextualize library data about patron satisfaction with reference services, we analyzed an existing corpus of chat transcripts. Having conducted a similar analysis in 2010, we also compared librarian behaviors over time. Methods – Drawing from the library literature, we identified a set of librarian behaviors closely associated with patron satisfaction. These behaviors include listening to and understanding patrons’ needs, inviting patrons to use the service again, and providing instruction or completing a search for patrons. Analysis of the chat transcripts included establishing a coding schema, applying these codes to individual chat transcripts, and analyzing these codes across the corpus of transcripts for frequency and correlation with other codes. The currently presented analysis used chat transcripts from the fall of 2013 and seeks changes in librarian behavior over time in order to gauge the success of establishing best practices and improving training standardization over the last three years. Results – The analysis shows that librarian behaviors have changed over time, pointing to what campus librarians are doing well, and that implementation of best practices at a campus level after the 2010 analysis may have increased these positive behaviors. The analysis also shows opportunities for further standardization and reinforcement of best practices. Conclusion – Qualitative analysis of already-collected data serves as a model for other units and suggests areas for process improvement, including enhanced coder training and code schema design. Further analysis of chat patrons’ questions is also warranted, including investigation of the relationship between subject- and location-specific questions and referrals.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.007 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.501 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it