Explainer Research, Knowledge, Thinking & Analytics · Monitoring, Evaluation and Impact Measurement · Issue 2026-W46

After a science-trust initiative launches, what should be measured?

The World Health Organization has launched a new initiative to strengthen trust in science. An announcement, however, does not show what has changed for an audience. This analysis proposes a way to distinguish reach, understanding and evidence-sensitive trust, then test a small communication decision. It is an ARPS framework, not an evaluation of the initiative.

By ARPS Editorial Team Last reviewed 11 November 2026 About 5 min read (estimate)

Source date 2026-11-10

A lens between research sheets and a balance represents examining evidence before judgment.

The World Health Organization has launched a new initiative to strengthen trust in science. An announcement, however, does not show what has changed for an audience. This analysis proposes a way to distinguish reach, understanding and evidence-sensitive trust, then test a small communication decision. It is an ARPS framework, not an evaluation of the initiative.

A launch creates an evaluation question

On 10 November 2026, WHO launched Together for Science and Health. Its priorities include bringing science and society closer, strengthening the use of evidence in health policy and improving communication about science. These are stated intentions. The announcement does not establish that trust has increased or that policy has improved. [1]

For an evaluation manager in a research institution or a science communication team, the useful question is what change would justify further investment. Our proposed distinction separates recognition of an institution, understanding its argument and being able to challenge that argument with reasons. A message could succeed at one and fail at another. An evaluation brief should therefore avoid compressing all three into a single indicator labelled “trust.”

This also changes what counts as a disappointing result. Greater recognition accompanied by weak understanding is a different problem from informed disagreement. The former might require a clearer explanation. The latter might require a more honest conversation about choices and costs. A publicity count cannot distinguish those situations.

Specify what the audience is being asked to trust

“Trust in science” might mean confidence in a method for testing claims, confidence in a particular organization or support for a specific decision. We propose separating these before writing a survey question. Someone who disputes the cost of a policy is not necessarily rejecting the scientific finding behind it. Someone who trusts an organization’s name might still be unable to explain its evidence.

Start with one claim, a defined audience and a particular use. For example, can readers of a research summary identify what was actually measured and distinguish the result from its interpretation? That is a comprehension task. Can they name information that would change their assessment of the result? That moves closer to evidence-sensitive trust. Neither task should become a test of agreement with the institution.

The distinction makes an unsuccessful message easier to diagnose. If readers remember the initiative’s name but miss a study’s limitations, increasing exposure may address the wrong problem. If they understand the evidence but disagree about how costs should be distributed, repeating scientific facts may leave the actual disagreement untouched. These are decision categories we propose, not established findings about WHO’s audience.

Build a measurement ledger with a rival explanation

Before a campaign, an evaluation manager can write three short entries. The first defines the immediate output: who encountered the message and through which channel. The second identifies a change in understanding that can be examined using a new example, rather than repetition of the original wording. The third names a plausible rival explanation: a different audience, an easier question or a changed response setting.

This ledger is not a validated psychological scale. Its operational purpose is to expose what a number cannot establish. Visits describe reach. A correct answer on an unfamiliar example describes performance on that task. Similar performance at a later point adds information about persistence. None of these alone demonstrates a society-wide change in its relationship with science.

A short open question can complement a closed one: what missing information would you need before reaching a judgment? Before reviewing answers, the team should define how it will distinguish relevant requests for evidence, repeated slogans and value-based objections. Where resources allow, another reviewer can examine ambiguous responses. The team should acknowledge that the coding still involves human judgment.

Test what happens when a limitation is harder to see

Consider a hypothetical team preparing two summaries of the same finding. One explains the sample’s limitations in the main text; the other places the qualification in a footnote. The result and numerical claims remain identical. Readers can then be asked whether the finding applies to a case outside the study population, with the task kept as comparable as possible and message allocation recorded.

The purpose is not to prove that one layout always builds trust. A clearer limitation might make readers less certain while making their judgment more appropriate. We therefore propose assessing calibration to the evidence before assessing reassurance. Confidence that outruns the evidence should not automatically count as a communication success.

If allocation is not randomized or participation is too limited, describe the exercise as exploratory and avoid an audience-wide effect claim. Record the recruitment method and the question alongside the result. Otherwise, an apparently impressive improvement could simply reflect which readers chose to participate.

Keep the exercise from becoming a referendum on the institution

The strongest objection is that a single task captures too little of trust. That objection is fair: relationships with institutions also involve experience, interests and values. Use the task to evaluate a particular message, not to rank communities or institutions. Optional short interviews may help explain hesitation, with clear consent and appropriate protection of participants’ information.

An international questionnaire should not be assumed suitable for Arabic or for a Gulf audience. Check the meaning of the question with the intended readers first. Collect only the personal information the exercise actually needs.

The next useful decision is to choose one message and state, before distribution, what the team will measure and what it will not claim. Which result would make your team revise its message even if reach increased? Readers wishing to follow peers and make professional connections can choose a community through ARPS communities; ordinary membership is free.

References

  1. WHO launches Together for Science and Health movement to strengthen trust worldwide — World Health Organization · 2026-11-10

References open in a new tab.

Discuss this with practitioners

Create a free account to save articles and join the Trainers community discussion.

Create a free account

Discussion 0 published comments

No published comments yet. Be the first to share your experience.