Ben Collier, PhD Book a call →free, 15 minutes

Version A + Version B

Evaluations

Two analyses, one teaching record

These two interactive reports were made from overlapping sets of Benjamin Collier's de-identified teaching evaluations. They use different time windows, which is why their headline response counts differ.

02

The denominator, made explicit

727 recent responses + 430 earlier responses = 1,157 total responses

Version A combines all 44 comparable five-point sections from 2011 to 2026. Version B uses the 24 sections since Fall 2023 for its headline and places the earlier CMU-Q record in a separate section. The pages are therefore different views of overlapping evidence, not conflicting counts.

Version A

Full-history teaching story

Model
GPT-5.6 Sol High
Coverage
2011 to 2026
Comparable sections
44
Section-level responses
1,157

A visual narrative with the long timeline, comment themes, course-level quotes, improvement ideas, and methods.

Open Version A

Version B

Recent-first evaluation notebook

Model
Claude Opus 5.5
Headline coverage
Fall 2023 to Spring 2026
Headline sections
24
Headline responses
727

A recent-first analysis with section and course charts. Its separate historical CMU-Q section contains 428 responses from 19 sections; the two-response 2011 record is not included there.

Open Version B

Why B's page does not say 1,155: its 727-response headline intentionally excludes the historical section, and its separate earlier series starts in 2012. Version A also includes a two-response 2011 section, bringing the complete comparable total to 1,157.