Visitor Observation: Privacy Issues

During my PhD I spent some time tracking and timing visitors to learn more about visitor behaviour in the exhibitions I was studying (more on the history and applications of visitor tracking here). Recently, I was asked about the privacy implications of doing such research. What steps do we need to take to ensure we’re a) staying on the right side of the law and b) respecting visitors’ rights to informed consent and ability to opt out of participating in research?

On the first part (i.e. The Law), I’ll tread carefully since I’m not a lawyer and it will vary in specifics from place to place anyway. However, in a general sense, museums will generally count as a “public place”, and people can reasonably expect to be seen in public places. Therefore if you’re just documenting visitors’ readily observable public behaviour, and nothing about them that may allow them to be identified as individuals, you’re probably in safe territory. However, it would be wise to check whether your museum is classed (in a legal sense) as a “public place” – for instance an entry charge may implicitly impose an expectation of some level of privacy on the part of paying guests.

So how about different approaches to informed consent?

The first consideration is cuing – do you tell visitors they’re going to be watched and/or listened to at the start of their visit? If so, then you are studying cued visitors – and gaining informed consent is relatively straightforward. When you approach potential participants, you explain the benefits and risks of participating, and they can decide whether they want to be part of it or not. The downside of cuing, of course, is that you’re probably no longer going to be documenting natural visitor behaviour – people tend to do different things when they know they are being watched.

Depending on what you’re studying, this may not be an issue – and, like contestants on Big Brother, visitors tend to forget they’re being watched or listened to after a while, even if they’re rigged up with audio recording equipment (Leinhardt & Knutson, 2004). Also, if you’re going to be tracking the same group of visitors over the course of a whole visit, which could mean following them for 2-3 hours, then you really do need to cue them first – otherwise, frankly, it just ends up getting creepy and weird for all concerned.

If you’re tracking visitors across a whole site, sooner or later they’re bound to notice you. Awkward. You’d be better off telling them first.

In contrast, tracking and timing uncued visitors through a single exhibition gallery can be done discreetly without visitors becoming aware they are being tracked (assuming you are not trying to hear what they are saying as well, meaning you can observe from a reasonable distance). It still takes a bit of practice, and is easier in some exhibitions than others. Even so, if someone approaches you and asks what you’re up to, the right thing to do is fess up, explain what you were doing, stop tracking that person and try again with a different visitor.

If you’re taking this uncued approach to visitor observation, you’re in a far greyer area with respect to informed consent. The usual approach is to post a sign at the entrance to the museum or the gallery informing visitors that observations are taking place, and giving them steps to take if they wish to opt out of being observed. In practice, this might be notices telling visitors which areas to avoid if they don’t want to be watched, or having a mechanism for visitors to opt-out by wearing a lapel sticker or wrist band (although chances are this won’t be necessary – it never came up in my research and my experience tallies with other researchers I’ve spoken to).

What about when you’re recording?

Things can get a little more complicated when you go beyond simple observation and field notes to audio or video recording visitor behaviour. It’s one thing to watch publicly observable behaviour, another to have that behaviour recorded, replayed, and deconstructed ad infinitum. This doesn’t mean it’s not done – audio recording at individual exhibits dates back to at least the 1980s and Paulette McManus’s landmark study of visitors evidently reading labels more than it might first appear (McManus, 1989). In that study, specific exhibits were hooked up to a radio microphone linked to a tape (tape!) recorder, and an observer unobtrusively watched the exhibit from a safe distance, making field notes to aid subsequent interpretation (Leinhardt and Knutson also emphasise how important observational data is to back up audio recordings, where there are frequently snippets that make little sense if you don’t have additional details about what was happening at the time). As far as I can tell, visitors were uncued in this study.

Audio recording of uncued visitors poses fewer difficulties than video recording, as people can’t (easily) be identified based on voice recordings alone. Things get tricker when you get to video, of course. My first exposure to video-based visitor research was seeing Christian Heath speak about his and Dirk vom Lehn’s work in V&A’s British Galleries in the early 00s (Heath and vom Lehn, 2004). In this case, although they specify that visitors explicitly consented to being part of the research, it’s not obvious whether this was done in advance, or after the fact by approaching visitors once they’d left the exhibit of interest (and then discarding the data of those who have refused to participate prior to analysis). This ex post facto approach is a way you can ensure both uncued visitor behaviour and informed consent, but as I have no direct experience of this, I don’t know how high the refusal rate is and how complicated it is to ensure data is discarded appropriately as required.

Irrespective of the type of informed consent, there is the issue of data storage. Gone are the days of tapes that could be kept under lock and key. You’ll need to have a data retention policy in place to ensure that anything that could potentially identify participants is kept secure, safe from those who have no need to access it . . . and from accidental syncing to your public Facebook feed.

Disclaimer: This is just general advice based on my own experience and what I can glean from some of the literature. Different parts of the world and different ethics committees may have different views, and the specifics of any given piece of research may make a difference as well.


Heath, C., & vom Lehn, D. (2004). Configuring Reception: (Dis-)Regarding the “Spectator” in Museums and Galleries. Theory, Culture & Society, 21(6), 43–65. Leinhardt, G., & Knutson, K. (2004). Listening in on museum conversations. Walnut Creek, CA: AltaMira Press.
McManus, P. (1989). Oh, yes they do: How museum visitors read labels and interact with exhibit texts. Curator: The Museum Journal, 32(3), 174–189.