Task force approves minutes, plans sensitivity tests and debates ChatGPT use for verbatim analysis

The Tax Relief Evaluation Task Force opened by approving the minutes of its Feb. 5 meeting by voice vote. A member moved to approve the minutes; another seconded and the chair recorded approval with no roll‑call tally recorded in the transcript.

Members then turned to process issues and next steps. Lynn presented a revised outline and said the group needs an integrated draft to finalize executive summary and recommendations. The task force agreed that Ellen and John will refine the analytics and methodology sections and run sensitivity analyses on application uptake (examples discussed: an additional 500 or 1,000 qualifying applicants) to show the effect on the RTE breakeven and tax shift.

On methods, Lynn described using an automated assistant to process qualitative responses: “I was brave enough to use ChatGPT to upload an Excel sheet and ask ChatGPT to analyze the qualitative answers we got, you know, over 900 lines.” Several participants urged consistent prompts and cautioned against reporting percentages from open‑ended verbatims, noting response bias; Shannon (researcher) said qualitative open‑ended responses typically should be analyzed by theme rather than prevalence percentages.

The task force agreed three tentative dates for follow-up meetings (Feb. 26 morning, March 12, March 26) and assigned tasks: correct and document analytic calculations, prepare sensitivity scenarios, and circulate an integrated draft for Meredith (assessor) to review. No final policy decisions were taken.

The chair closed by noting the team will work on outreach to landlords and that the integrated draft should include clear methodology and caveats about representativeness.