EPFL_Method_Appendix_Instrument_RQ_Mapping.csv

Question No.,Survey Item (Summary),Data Type,Coding / Analysis Method,Primary RQ Link,Purpose in Framework
Q1,Organisation size,Quantitative,Used as control variable in regressions,RQ1–RQ3 (control),Adjusts for scale bias in adoption behaviour
Q2,Industry,Quantitative,Used to cluster by sector and infer adoption archetypes,RQ1,Context for stage classification
Q3,Familiarity with Hugging Face,Quantitative,Converted to AI overall maturity stage,RQ1 / RQ3,Proxy for AI maturity level
Q4,Type of AI solution used (Open / Proprietary / Hybrid),Quantitative,Converted to Open-source adoption stage,RQ1,Classifies adoption stage
Q5,Key decision factors when choosing AI solutions,Qualitative (multi-choice),Coded into Gate or Lever categories using decision factor map,RQ1,"Identifies dominant adoption drivers (Compliance vs. Performance, etc.)"
Q6,Integration challenge rating,Quantitative,Likert scale → descriptive and correlation,RQ2 / RQ3,Indicator of integration readiness
Q7,Plan to increase open-source use,Quantitative,Treated as dependent variable in regressions,RQ3,Measures intention / outcome of adoption
Q8,Motivation for choosing open-source or proprietary AI,Qualitative (open text),Deductive coding with Gate–Lever framework,RQ2,Explains reasoning behind adoption path
Q9,Example of a use case showing value from open-source AI,Qualitative (open text),"Coded for Lever outcomes (performance, cost, time to value)",RQ3,Identifies realised benefits post-adoption
Q10,Challenges when adopting open-source AI,Qualitative (open text),"Coded for Gate barriers (compliance, security, documentation, etc.)",RQ2,Identifies governance and technical blockers
Q11,Desired support / features for Hugging Face,Qualitative (open text),"Coded for Lever needs (support, SLAs, customisation)",RQ3,Reveals missing enablers for scaling
Q12,How organisation evaluates open-source vs proprietary trade-offs,Qualitative (open text),Coded for Gate–Lever interaction,RQ3,Explains decision-making logic