Text from the interactive explanation

From researcher to communicator: Creating a specialized training framework for CERN Science Gateway guides

Oddershede, Kristian Dahlmann

A master's thesis from Aarhus University & CERN, explored

More than 750,000 people have visited CERN's Science Gateway. Its exhibitions are staffed by real CERN scientists and engineers in red jackets. Many visitors didn't notice them, or thought they were security.

Now showing: a shift before the new training.

Where the tracked guide's time goes (this simulation, live)

Silent behind an exhibit 0%

Talking with a colleague 0%

On the open floor 0%

Answering visitors 0%

Started a conversation 0

Family (adults, children)

Tracked guide

Colleague (untracked)

Tracking dots (SEEZ-style)

Guide starts a conversation

Visitor starts it

Approach declined

“Guides? I did not notice there were guides there?”

“Ohh, the ones in the jackets? Yeah, I did notice them, but I thought they were there for safety or something.”

The puzzle

Why would a physicist stand silently for most of a shift?

The author, a physics master's student and experienced museum educator, spent a four-month internship at CERN (August to November 2025) doing action research: studying the guides and changing how they were trained. Before reading on, pick the explanation you find most likely.

Choose a hypothesis to see what the survey of 56 guides and the field observations say.

1 Role ambiguity

Who speaks first?

CERN's exhibitions use guides , not educators. An educator runs something structured and scripted: a tour, a show. A guide has unstructured, walk-up conversations with whoever happens to be there. Nobody had told the guides clearly whether that meant waiting to be asked, or starting the conversation themselves.

Your prediction

Make a guess, then reveal.

Measured Survey of guides trained the old way, N = 56 (thesis Fig. 14).

“What do you spend the most time on during a shift?”

Measured Each guide chose two activities; bars show counts out of 56 (thesis Fig. 13, redrawn). Only 77% named “sitting on a bench” and only 52% “checking your phone” among the two things they did least.

The twist from the tracking

Standing still was not even an option in the survey. Yet when three randomly chosen guides were tracked for an hour each, it was the dominant activity, time-wise, for all three: silent, behind one exhibit. Talking with colleagues came second or third, then answering visitors once they approached.

Measured SEEZ software tracking, N = 3 guides, Discover CERN (thesis Fig. 11 and Section 6).

“I have just been walking around for an hour.”

“if you don't approach the [visitors], or if you don't look at their eyes and smile and say hello, it is almost never happened to me that someone came and asked me a question.”

So what is a guide for? The thesis' answer

“Science museum guides should facilitate visitors' interactions with the exhibits while assessing visitors' personal contexts and adapt accordingly. Guides should support shared experiences between visitors in groups.”

CERN adds two things: relate the content to CERN, and to the guide's own work.

2 Engagement skills

The opening line

In a classroom everyone already knows who teaches and who listens: an unspoken didactic contract . On a museum floor nobody has agreed to anything. So a guide's first sentence proposes a contract, and the visitor accepts or declines it, often out loud. At CERN, real declines sounded like No, we don't need help or We can try ourselves.

Illustrative simulation A diffusion cloud chamber like the Discover CERN exhibit, whose screen lets visitors release α or β particles. α tracks: short, thick, straight. β tracks: long, thin, wiggly. Rare long straight tracks come from space. Drag across the top of the chamber (or press “Sleeve trick”) to clear the haze.

2a · Replay a real conversation

Observed by the author in August 2025 at the cloud chamber in Discover CERN. The guide: an engineer at one of CERN's big detectors, with more than a year of guiding experience. The visitor: an adult whose family was elsewhere.

2b · Your turn

Same chamber. A parent and a child are pressing the buttons. You are the guide in the red jacket. How do you open?

Pick an opening line to see the contract it proposes.

Outcomes are qualitative readings from the thesis, not probabilities. Lines 3 and 4 adapt the thesis' Examples 1 and 2 (Example 2 was about turning a magnet upside-down).

3 Expert blindness

Three kinds of knowledge

Knowledge has to be translated twice. Exhibition designers turn what scientists know into exhibits. Then, on the floor, a guide turns the exhibit into a conversation. The thesis names this Science Museum Didactic Transposition (SMDT), adapting a classroom theory to museums. Tap the parts of the diagram.

Exhibition design (designers)

The didactic situation

Only when guides are scientists

Redrawn Redrawn and animated from Fig. 3 of the thesis.

Tap a part of the diagram

Green: designers turn professional knowledge into content. Blue: guide, content and visitor meet in the didactic situation, where “the mediated” happens. Yellow: a shortcut that only exists because CERN's guides are themselves scientists and engineers.

The formal definition

“SMDT is the process by which professional knowledge is converted to content by a science museum, and is further converted to the mediated through the didactic situation.”

Which kind is it?

Sort each sentence. Is it professional knowledge , the museum's content , or an invitation to explore ?

Sorted 0 of 6.

Build the next two sentences

You are at the cloud chamber with a family. Choose what you say next; watch which arrows of the diagram light up.

Visitor's role: pick two sentences.

Conceptual, not scored.

Expert blindness, observed

One guide explained ELISA , a small working particle accelerator in the exhibition, using “RF cavity”, “Ionization” and “MeV” without explanation, to families and to expert physicists alike. In the interview, the same guide said:

Other guides start explaining the LHC with up- and down-quarks, while many visitors aren't comfortable with what a proton is. Many guides had tried only a few exhibits themselves; the new training tells them to walk the exhibitions as a visitor, without their CERN badge.

Said versus seen

All 4 observed guides said they adapt their explanations to visitors; 1 of 4 was observed doing so fruitfully.

Observed Group A: four guides, each observed for a shift and interviewed afterwards.

The intervention

Five modes: a job description you can actually do

The old two-hour training was a lecture-style tour by the exhibition designers: the stories of the exhibitions, a few exhibits in detail, ELISA, how to do demonstrations. Nothing on what a guide is expected to do, or how to approach a visitor. The budget could not grow: trainings must fit on Mondays, the only day Science Gateway is closed, and guides volunteer alongside full-time jobs.

How the trainings changed

The ELISA training is only needed to guide at ELISA; the voluntary engagement training is for experienced guides.

Inside the new two hours (minutes)

From the thesis Fig. 6 and Appendix I. The training is built on social constructivism (knowledge is built “like LEGO bricks clicking together”, through doing and talking) and inquiry-based learning: it treats itself as a learning environment, the way guides should treat visitors.

1Introduce the exhibition

Stand in front of your exhibition and briefly introduce its central elements. Visitors learn you're a guide they can talk to, “so you get more questions!”

“This exhibition is called Discover CERN and is all about what we do, right here at CERN…”

2Demonstrations

A scripted 5–10 minute “cooking recipe” at one exhibit per room, with visitors interacting. Start with a few people; more will join. Ideally at least 3 per two-hour shift.

3Approach visitors

Go up to visitors who are playing or looking, for example at the cloud chamber. Don't stop their play; support it by asking questions as they try things. “Don't ask questions that sound like a test.”

4Join an activity

Be one of the players a group game needs, like Operate the Crane or Proton Football. Playing makes you approachable, and it is easy to connect the game to CERN or your own job afterwards.

5Approachable walk-through

Walk slowly through the middle of the room, smile, don't avoid eye contact, let visitors come to you. A fallback for resting that “should not be the main part of your shift”.

“Always do one of these five modes, and you are doing what is expected of you.”

In every training, trainees were asked which mode allows the deepest conversations, which to start a shift with, and which they were most excited to try. Each time, every mode was chosen by at least one trainee: “no mode is strictly better than the others — the diversity of the modes is a part of the framework's strength.”

Plan a shift

The shift simulator

A shift is two hours: eight 15-minute slots. Give each slot one of the five modes, or the old habit, and watch the same floor play out. Gold rings are conversations the guide started; white rings are ones visitors started.

Toy model. The rules paraphrase observations in the thesis; the numbers it produces are not measurements.

This shift so far

The thesis suggests future research could “count the exact number of proactive interactions initiated per guide per shift”. The gold counter is that idea in miniature.

Families arrive at a steady rate, visit about 4–6 exhibits and leave. About 40% are “shy”: they never walk up to a guide by themselves. This is grounded in the guide community manager's remark that a guide who only answers questions “will surely only reach visitors who come to you… who are already brave enough to approach you”, while demonstrations reach visitors “who would never approach you”.

Visitors only walk up with a question if a guide is near, on open floor and not busy. The chance is low behind an exhibit or while chatting, higher during a walk-through, and higher still after the family was greeted (Mode 1).

Mode 1: the guide stands near the entrance and greets arriving families (gold), marking them “greeted”.

Mode 2: the guide runs a demonstration at one exhibit; nearby families gather (one gold contact per family, shy ones included).

Mode 3: the guide walks to a family playing at an exhibit and asks a question; some decline (grey).

Mode 4: the guide joins Operate the Crane or Proton Football; nearby families join the game; afterwards questions flow both ways.

Mode 5: the guide strolls through the middle of the rooms; only bold families may walk up (white).

Old habit: the guide stands behind the cloud chamber, silent or chatting with a colleague; only bold families who come close may ask.

Energy: active modes drain it; walk-through and the old habit recover it. At zero the guide switches to a walk-through until recovered, echoing a newly trained guide: “When I'm tired of talking and just walking, I don't feel bad because this is part of [the 5 modes]”.

The exhibits used for demonstrations here are placeholders; the floor plan is a schematic inspired by the thesis' Fig. 11, not to scale.

The evaluation

Did it work?

No single method could answer that, so the thesis triangulates : five independent ways of looking, aimed at the same guides, like detector layers around one collision. Tap a beam.

“The different methods largely converged.” Field notes, observations, interviews, tracking and the survey pointed at the same three challenges. A follow-up interview with CERN's guide Community Manager, Niklas Herff (April 2026), adds a colleague's view.

Context: 103 active guides did at least one shift between August and November 2025; 37 were regulars (at least one shift a week on average).

Old training vs new training, observed side by side Observed · group A, 3 + 1 guides

“After the training, I felt quite ready. And during the training, I found that these five [modes] are quite important. And even though you think that they are naturally inside you, knowing those five, like they are actually five things. Even though you naturally do [them], during the exhibition [guiding], sometimes in my brain, I [think] ‘and now I'm doing this [mode] actually’.”

“So basically the training was just like if I am a public person and I just go inside and see what's there.”

“How successful was the training in preparing you?” (old training, 0–10)

Net Promoter-style score −9

Measured N = 56, thesis Fig. 18. Around 35% did not feel prepared for their first shift and did not find the training satisfactory. The author notes the question was not worded as a true Net Promoter question, so the score deserves less weight than usual.

“How often do you do demonstrations?”

Measured N = 56, thesis Fig. 16. Two regular guides who said they did many demonstrations were actually giving ordinary lecture-style explanations.

“Would you recommend guiding to a colleague?” (0–10)

Net Promoter Score +45

Measured N = 56, thesis Fig. 20. Guides like the role.

“What are your two main motivations for guiding?”

Measured N = 56, thesis Fig. 21. Yet guides rarely talk about their own CERN work: only 7 of 56 list “explaining what you do at CERN” among their most-time activities, and in the field notes guides discussed their own role at CERN only a handful of times, always after a visitor asked.

What colleagues said

After each new training the informal feedback was entirely positive. Three trainees independently said versions of “it was the best training they had ever attended for anything”.

Verbatim Niklas Herff, CERN guide Community Manager, April 2026

What the evidence can and can't say yet

The survey was meant to be repeated with newly trained guides in March/April 2026 for a before-and-after comparison. After a restructuring at CERN, no new guides were trained that winter and spring, so it never ran. The evaluation therefore rests on observations, interviews and feedback, and group A is four guides. The author calls the conclusions “indicative rather than conclusive” and frames the work as an in-depth case study of CERN.

Back to the floor

From researcher to communicator

After the thesis was submitted (August 2026)

The new trainings have restarted. The new training is mandatory for new exhibition guides, and existing guides must retake it, with the new framework, before 28 September to continue guiding.

What comes next, according to the thesis

Count the exact number of proactive interactions each guide starts per shift.

Study what happens when an experienced trainer drops by on a new guide's first shift.

Have a dedicated person keep evaluating guides and refining the trainings, someone who also guides.

Keep treating training as a learning environment, so that “…CERN can ensure that the transformation from researcher to communicator becomes a successful endeavour.”

Beyond CERN, SMDT offers a vocabulary for unstructured museum conversations, an area where research is scarce, and the 5 Modes is a practical tool that other institutions using scientists as guides could adapt.

Source

Kristian Dahlmann Oddershede (2026). From researcher to communicator: Creating a specialized training framework for CERN Science Gateway guides. Master's thesis (60 ECTS, Physics), Department of Physics and Astronomy, Aarhus University. CERN Document Server, version v1. Supervisors: Hans Otto Fynbo and Kristian Hvidtfelt Nielsen. DOI: 10.17181/p2cm4-3z487

Content © 2026 Kristian Dahlmann Oddershede, CC BY 4.0. Charts redrawn from the thesis' survey figures; floor plan is a schematic inspired by Fig. 11 (not to scale); simulations are illustrative.