Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Public Engagement topic
No spam. Unsubscribe anytime.
Acton resident tells finance committee town surveys can be biased; shares local results
Summary
Resident and Transportation Advisory Committee associate Eric Heals told the finance committee that surveys tied to specific events or venues can produce biased, nonrandom samples. He presented his Aug. 22–Sept. 21 survey of 4,035 emails and said six priorities rise above his calculated margin of error.
Eric Heals, an associate member of the Transportation Advisory Committee, told the Town of Acton finance committee that many local surveys are statistically invalid and prone to selection bias, and he presented results from a survey he ran this summer.
Heals said his questionnaire ran Aug. 22–Sept. 21 and was sent to 4,035 email addresses; 1,798 opened the message (about a 44.6% open rate), 236 clicked a link (about 5.8%) and 86 completed the survey (about a 2.1% completion rate). Using a 0–4 rating scale he said he computed a margin of error of 0.42 and identified 2.42 as the break‑even score above which a priority shows statistically clear support in his sample.
"There is, unfortunately an art to creating an inaccurate survey," Heals said, using a movie‑marathon example and a subway‑rider hypothetical to illustrate how surveying a self‑selected audience skews results. He said that town‑meeting turnout can act like an "echo chamber" for some issues and may not reflect the broader town population.
Heals summarized his ranked priorities and said six items stood clearly above the margin of error: advanced town meeting participation (mean 2.98), Main Street intersection and complete‑streets improvements (2.79), final stormwater infrastructure approval and implementation (2.76), a volunteer recruitment plan (2.62), proactive strategy for gas‑line replacement (2.59) and a study/implementation plan for fire/EMS deployment (2.43).
Committee members thanked Heals for the work and raised concerns about self‑selection bias. One committee member noted that producing a truly random sample would be costly and time‑consuming; Heals agreed and said his survey was not an official town survey and that a formal, randomized town survey would likely produce higher response rates if done through town channels.
The finance committee did not take formal action on the presentation; members asked to continue discussion about survey quality and public engagement at a future meeting.
