Measure what the experiment achieved

NASA’s 1982 executive summary reports 35 approved United States experiments, of which 20 were completed, and discusses additional demonstrations and events separately. These are categories in a particular retrospective report, not a worldwide count of everything done with Hermes. The report also describes funding difficulties that prevented some approved work from proceeding.

Approval and completion answer different questions. Approval says that a proposal passed a selection process; completion says that the activity was carried out under the report’s definition. Neither alone establishes that every research hypothesis was confirmed or that every participant benefited.

Using the report’s own two experiment counts, 20 / 35 is approximately 57 percent. That is the completion share of the stated approved group. It is not a 57-percent spacecraft reliability figure. To explain the uncompleted work, consult the reasons and scope of the records rather than assigning all of it to hardware failure.

CSA records Canadian university telemedicine experiments using Hermes during March–June 1977. NASA’s reports also describe education and public-service uses. These demonstrations raise several distinct questions: did the connection work, could the participants use it, and did it improve the intended activity? A successful answer to one does not settle the others.

Imagine a fictional remote class with an 80-percent quiz average, compared with a different class averaging 60 percent. The difference is 20 percentage points, but it is not by itself proof that the satellite connection caused better learning. Prior knowledge, teaching time, questions and participant selection could differ.

A stronger comparison starts by defining the outcome and making the groups or repeated conditions reasonably comparable. Record who participated, what was taught, what was measured and what else changed. A small pilot can reveal a practical problem or motivate a larger study without proving that the same result will occur everywhere.

Keep the denominator attached to the result

The project summary explains that NASA supplied spacecraft time without charge to approved United States users, while experimenters still needed funding for their activities and ground equipment. Free access to one component did not make the complete experiment free. This distinction matters when asking whether a demonstration can become a continuing service.

In an invented classroom budget, a session has 450 cost units of fixed setup expense and five units per participant. With 30 participants, the total is 450 + 5 × 30 = 600 units, or 20 units per participant. With 50 participants, it is 700 units, or 14 each. These are teaching numbers, not historical Hermes prices or a current service quote.

The lower per-person figure does not automatically make the larger class better. Equipment, staffing and teaching quality may impose limits, and the simple model assumes the fixed cost stays fixed. A useful final assessment reports the technical result, the user outcome, the costs included and the remaining uncertainties together.