SSHOC D4.12 Guidelines for the integration of Audio Capture data in Survey Interviews
Bibliographic record
Abstract
This deliverable is the first associated with Task 4.4. Voice recorded interviews and audio analysis in the Social Sciences and Humanities Open Cloud project (SSHOC). This task will collect audio data in the form of voice recorded interviews from the Generations and Gender Survey. This audio data will then be processed and analysed by colleagues at CLARIN to contribute with automatic speech recognition, speaker attribution, part-of-speech-tagging, named entity labelling and other NLP tools. Survey methodologists and Social Scientists from the EVS and GGP will work together with data scientists and oral historians from CLARIN to develop a survey module specifically adapted to integrate audio recordings and their processing into the traditional data collection process. Thematically they will focus on the qualitative assessment of value statements. Once fielded, CLARIN will adapt existing auto- transcription tools to the specific needs of the audio survey data and make the transcribed files available for analysis. Data Archiving and Networked Services (NL) (DANS-KNAW) will oversee archiving and dissemination of the data, drawing on their significant experience with oral histories. In these guidelines, we set out the questionnaire and fieldwork principles for the implementation of the Audio Survey Modules. These will then be implemented in fieldwork in early 2021 and the data processed and analysed before the end of 2021 (Month 36 of the project). A specific questionnaire and guidelines are required as the aim of the project is unprecedented in several respects: Audio data is sometimes captured by surveys as part of data quality control, but the technical implementation and questionnaire content are rarely if ever optimized to ensure that the digital language data generated from the interview is optimized for linguistic analysis. What is unique about this task is that researchers from CLARIN were involved from the early stages of design in order to ensure that the substantive focus and technical implementation of the project would be able to produce audio data and transcripts that produce meaningful results when analysed with the tools at CLARIN’s disposal. These guidelines proceed as follows. First, the various aims of the project are outlined from the perspective of survey infrastructures (GGP & EVS) and linguistic infrastructures (CLARIN). These, help shape the technical requirements that are then laid out in section 4. Finally, the questions which we intend to field are then laid out.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.162 | 0.227 |
| Meta-epidemiology (narrow) | 0.003 | 0.008 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.010 | 0.009 |
| Science and technology studies | 0.005 | 0.005 |
| Scholarly communication | 0.012 | 0.006 |
| Open science | 0.008 | 0.012 |
| Research integrity | 0.007 | 0.006 |
| Insufficient payload (model declined to judge) | 0.259 | 0.219 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".