ACM Interactions

Authors:
Interaction Research Studio

---

 

ProbeTools are fully self-contained digital devices robust enough to be used in the field. Each one offers a unique and engaging way for people to tell you about themselves and their everyday lives. At the outset of a study, you configure the devices and lend them to participants to use independently. Once they're done, they return the ProbeTools to you. Simply download the pictures and recordings, and enjoy interpreting new glimpses into other people's worlds.

TaskCam lets people take pictures in response to prompts (tasks) displayed on a small screen on the back of the camera. Participants can scroll through tasks, select one, and take a picture in response. Each new image is stored along with the current task that prompted it. Both the tasks and the pictures are stored on an enclosed Micro SD card, allowing you to customize the tasks before your study and download the pictures when participants have returned the device.

  Insights

TaskCam is an updated version of a classic cultural probes camera, and the workhorse of the ProbeTools family [1].

VisionCam captures time-lapse images when it is activated, using computer vision to retain only contour animations—similar to line drawings—to protect privacy and enhance aesthetic interest. It is designed to be a non-invasive way to record events over time at home or in public spaces. Participants can choose where to place the camera and adjust the density of contours to control what is shown and in how much resolution. As with all the ProbeTools, when participants return the devices, the results can be downloaded for your review.

TaskCams can be made in a variety of configurations and housed in paper or 3D-printed cases.

VisionCams come in vertical or horizontal orientations. Swiveling the lens cap activates the camera. Ribs on the casings facilitate attaching accessories, including commercially available USB battery packs for power.

Interviewer can be made with 3D-printed or textile cases to suggest various cultural connotations. Pressing the button plays a question, after which participants have several seconds to record their answer.

Interviewer asks people questions that you record, then pauses for a short time to record their answers. Digital audio processing is used to change both the questions and the answers to anonymize voices. Participants can borrow the Interviewer, loaded with the questions you want to ask, and use it at their own pace wherever they please—they can even answer questions more than once. When combined with playful questions, its disarming form can evoke spontaneous and intimate replies.

ProbeTools create a space of possibilities for learning about people—from the answers they give to your questions, to pictures they take themselves, to animations recorded passively. Use them together or separately to gain new access to what your participants get up to when you're not around. Better yet, ProbeTools are completely reusable. Simply clear the memory and load up new tasks or questions, and you're ready to run a new study.

TaskCam is made with a few basic parts, including a custom Arduino shield [center) sold at cost on www.probetools.net.

The shield snaps apart, allowing different configurations for different housings.

The VisionCam uses a standard Raspberry Pi 3.

  Making Probetools

There's only one catch to ProbeTools: You can't buy them online or in stores. You have to make them yourself!

ProbeTools are DIY devices—fully finished designs that people can make themselves with little or no technical expertise. Most of the electronics simply slot or snap together; many versions do not even require soldering. For the most part, they rely entirely on off-the-shelf components that can easily be ordered online, combined with software that we make available for download. The only exception is the TaskCam, which relies on a custom Arduino shield that we sell on a not-for-profit basis from our website: www.probetools.net.

The finished devices can be housed in a variety of ways. For instance, we offer templates for cases made of card or paper, patterns for sewing textile cases for the Interviewer, and complete specifications for a range of 3D-printed housings. Our aim is to make the ProbeTools easy to make and use for even the least technically minded designer, practitioner, or researcher. Think of them as the flat-pack furniture of the technology world (except with better instructions).

ProbeTools can also be seen as open source products, in which all the specifications needed to build a hardware product are made available for anyone to build, modify, and improve. We have released full editable specifications for the ProbeTools hardware and software to encourage customization. We welcome contributions from people who want to add new features, try new components, offer new housings, or spin off completely new designs.

VisionCam can be placed to record everyday events unobtrusively.

Participants can answer Interviewer's questions whenever and wherever they like.

People can take pictures in response to TaskCam's prompts, or find the prompt that best corresponds to a photo opportunity.

A lot of specialist knowledge is needed to engage with most open source products, from programming to 3D CAD modeling. This means that, in practice, they are mainly accessible to expert users. We see DIY devices like ProbeTools as spanning from the world of open source products toward the easy-to-build electronics kits sold by companies like Technology Will Save Us (www.techwillsaveus.com). By designing them to be extremely easy to make and use, on the one hand, and absolutely open to modification, on the other, we hope to have made ProbeTools accessible and appealing to audiences ranging from technological newbies to hardcore tech fanatics.

  Using Probetools

Once you have made the ProbeTools, you can hand them over to your study participants, who will use them to document their worlds.

Each of the ProbeTools is designed to be simple and intuitive to use.

Participants can take their TaskCam with them wherever they go. Scrolling through prompts on the camera, looking for one that inspires a picture, can be a pleasant diversion.

Alternatively, the TaskCam can be used spontaneously when your participant sees something that reminds them of one of the tasks. Either way, TaskCam shows how many times they've responded to a task, allowing them to keep track of how they're doing.

ProbeTools are aesthetically finished and technically robust so participants can bring them along in their day-to-day activities.

VisionCam can be used in two ways. You can instruct participants to use it in specific locations or situations—for example, to record an evening meal, activities in their front hallway, or even a walk to work. Alternatively, you can use more open-ended requests—"a place to slow down," for example, or "the taste of morning"—to prompt participants to find their own ways to use the device.

Interviewer can be used whenever participants have the urge. They might decide to answer all the questions in a single session, perhaps somewhere private, at home or in a secluded spot outdoors. Or they might carry it with them as a kind of companion, answering a question or two in their off moments. There's no limit to the number of times they can answer the same questions, so different situations might give rise to different responses.

Images and recordings can easily be retrieved from the ProbeTools and used to inspire design sessions.

  Probetool Returns

Interpreting probe returns can be daunting. It's not always clear why a particular picture has been taken in response to a certain task. Is it meant to be the cat in the foreground that is "beautiful," or the view out the window? Why is a picture of a bus an appropriate response to "something broken"?

There are a few tactics for using probe returns that can help. One is to look for landmarks— individual returns that stick out as particularly salient or interesting. Another is to focus on textures across returns, ignoring details to get a feeling for the combination of person and place that they reveal. Sometimes TaskCam questions don't matter so much, for instance, as the unusual views into people's lives that they elicit. Finally, it is useful to imagine the person behind the returns. What does an interview response suggest about motivations, hopes, fears? What does a VisionCam animation suggest about somebody's home life? Can you extrapolate to how your participants might orient to other situations? You won't be sure of your answers, but your guesses might still lead to productive ideas.

---

It's not always clear why a particular picture has been taken in response to a certain task. Is it meant to be the cat in the foreground that is "beautiful," or the view out the window?

---

Someone once compared interpreting probe returns to flying a kite. Anchored by their origins in participants' realities, your interpretations are borne aloft by imagination but will travel only so far as you choose to let them go.

There are no right answers when it comes to probes. They are designed to be empirically grounded, yet open to interpretation. They rely on imagination and provisional speculations. Their purpose is not to produce validated, detailed accounts of participants' lives, but rather hints and clues that can be a starting point for design. The ProbeTools embody this ethos, offering new ways to collect evocative data for design.

  DIY Designs

There are several reasons for designing ProbeTools as DIY devices.

Primarily, we want to support their use by as many researchers and practitioners as possible, because we believe probes to be a useful method for design.

ProbeTools are also part of our broader investigation into how design research products can circulate more widely. Typically, practice-based researchers produce only one or a few multiples of the things they design. That means that, as James Pierce [2] has pointed out, most of us never get to live with or gain firsthand experience of the artifacts made by other practice researchers. Instead, we encounter them through articles or lectures, or occasionally in exhibitions. This is a shame insofar as these artifacts are thought of as reflecting new knowledge, not just in their conceptual design or appearance—which are relatively straightforward to convey using words and images—but also in the lived experiences their use engenders and their changes over time—which tend to be more difficult or impossible to describe satisfactorily. From this point of view, making ProbeTools available as DIY products is a test of a new strategy for circulating research products.

The third reason we have made ProbeTools as DIY devices is more political. We are interested in circulating ProbeTools widely not only to promote an approach to research, or to disseminate the outcomes of our practice, but also to explore the possibility of designing and distributing computational products independently from the large, profit-making companies that currently dominate our digital lives. To be sure, we enjoy the latest technological offerings from Silicon Valley as much as anybody. But along with many people these days, we are also concerned with the costs—the incursions into our privacy, the amplification of attention-seeking voices, the commodification of the lowest common denominators of everyday life. We're curious to see if a form of DIY products that span from open source to flat-pack designs might provide an alternative, less-encumbered approach to designing and delivering computational products.

  Cultural Probes

Cultural probes are collections of evocative tasks given to volunteers to elicit inspiring responses. Probes were invented in 1999 by Gaver, Dunne, and Pacenti for a project set in three different European communities [3]. The impetus was to avoid questionnaires or focus groups, which were felt to prefigure or homogenize responses, and instead find a more exploratory, expressive form of self-documentation.

Most cultural probes are collections of materials posing tasks for volunteers. These typically include a variety of printed items such as maps, postcards, or diagrams with requests for people to add information in the form of annotations, drawings, or stickers. For example:

• A set of self-addressed postcards that ask questions ("What would you write your local politician?" "Tell us a joke," "What do you like about where you live?")
• Maps with stickers allowing people to mark where they live, where they go to see other people or to be alone, or where they would like to go but can't
• Readymade diagrams (e.g., an image of the solar system) with a request to indicate friends and family.

Examples of images guiding the construction of TaskCam Paper Long. Full, illustrated "recipes" for making ProbeTools can be found on www.probetools.net.

Probe tasks often take the form of printed items for accessibility.

For instance, people annotated a diagram of Dante's Heaven and Hell with the names of friends and family.

Other tasks have used electronic and physical artifacts. Top to bottom: disposible camera with requests; audio recorder for dreams; pinhole camera; listening glass.

Tasks and materials can change from study to study—there are no predefined ingredients. Often they also include simple devices such as cameras or recorders with requests for certain kinds of content. ProbeTools are designed to augment or replace these devices, which are increasingly becoming outdated, or too fully featured to offer the constraints that are useful for probes.

  Designing Probe Tasks

Important as the probes' material forms are, it is the design of the tasks that determines whether or not probes are intriguing and revealing.

The best probe tasks balance empirical encounters with playfulness and surprise. For participants, they undermine ideas about research to encourage informal intimacy and creativity. For researchers, they produce observable evidence with enough uncertainty to leave room for the imagination.

In designing probe tasks, the aim is to hit a sweet spot between focused inquiry and entertaining self-expression. Good probe tasks are almost flirtatious, allowing innocent replies while opening the door to more intimate responses. Thus, playful, open, or even absurd requests are more rewarding for participants and surprising to researchers than straightforward ones. But clever opportunities for play that don't reveal anything meaningful aren't useful either. Balance is essential in designing probe tasks.

Probe returns are often rich and detailed, but also mysterious or downright puzzling. This is considered a strength of the approach.

It should be clear that the epistemological commitments of probes are different from many other user study methods. Rather than emphasizing the collection of representative, comprehensive and verifiable data, probes reveal fragmentary insights into peoples' lives, balancing empirical encounters with a sense of mystery that leaves room for, and ideally stimulates, designers' imaginations.

ProbeTools are designed to offer affordances that are useful for creating cultural probe studies. For instance, the automated animations produced by the VisionCams sacrifice some of the information conveyed by video not only to protect privacy but also because the results are more ambiguous, open to interpretation, and aesthetically compelling. Moreover, we make ProbeTools available as DIY designs to allow people to build upon these characteristics for their own studies. Of course, it is possible to use ProbeTools to serve more traditional forms of user study—just as it is possible to disguise a conventional questionnaire as a cultural probe. We hope, however, that by making ProbeTools available, we will encourage the use of the methodological approach behind them, as well as the devices themselves.

  References

1. Boucher, A., Brown, D., Ovalle, L., Sheen, A., Vanis, M., Odom, W., Oogjes, D., and Gaver, W. TaskCam: Designing and testing an open tool for cultural probes studies. Proc. of the 2018 CHI Conference on Human Factors in Computing Systems. ACM, New York, 2018, Paper 71; https://doi.org/10.1145/3173574.3173645

2. Pierce, P. On the presentation and production of design research artifacts in HCI. Proc. of the 2014 Conference on Designing Interactive Systems. ACM, New York, 2014, 735–744; https://doi.org/10.1145/2598510.2598525

3. Gaver, B., Dunne, T., and Pacenti, E. Cultural probes. Interactions 6, 1 (Jan. 1999), 21–29; https://doi.org/10.1145/291224.291235

4. Gaver, W., Boucher, A., Jarvis, N., Cameron, D., Hauenstein, M., Pennington, P., Bowers, J., Pike, J., Beitra, R., and Ovalle, L. The Datacatcher: Batch deployment and documentation of 130 location-aware, mobile devices that put sociopolitically relevant big data in people's hands: Polyphonic interpretation at scale. Proc. of the 2016 CHI Conference on Human Factors in Computing Systems. ACM, New York, 2016, 1597–1607; https://doi.org/10.1145/2858036.2858472

  Authors

Andy Boucher is a designer and maker who has developed a practice-based approach centered on the production of interactive devices and design-led methodologies for user-studies. He co-leads the studio with Bill Gaver. [email protected]

Dean Brown is a product designer who makes objects, installations, and visitor experiences, with a dedication to understanding context, materials, and process. [email protected]

Bill Gaver does design-led research developing technological products that explore unconventional values and the methods and concepts to support them. He co-leads the Studio with Andy Boucher. [email protected]

Naho Matsuda is a design researcher experienced with a diversity of media including performance, objects, installation, print, and writing. [email protected]

Liliana Ovalle has a background and personal practice in product design and also supports user engagements and graphic design in the studio. [email protected]

Andy Sheen is a creative technologist in the studio, with a background in audio and hardware design. [email protected]

Michail Vanis is a creative technologist with a background in interaction design, electronic engineering, and computer science. [email protected]

  Footnotes

The Interaction Research Studio at Goldsmiths, University of London, brings together designers, technologists, and social scientists to pursue practice-based research on technology and everyday life. [email protected]

  Sidebar: MAKE, USE, INTERPRET

ProbeTools are designed to be used in a three-stage process:

• Researchers make the devices and configure them for their study.
• Participants use the ProbeTools to document their lives.
• Researchers interpret the images and sounds when participants return the tools.

The result is a kind of dialogue between researchers and participants, with plenty of room for surprising twists and interesting digressions.

  Sidebar: RETRIEVING THE RESULTS

You can ask participants to return the ProbeTools to you by a deadline, or simply when they feel they've spent enough time with them. Returns can be made face-to-face, occasioning conversation about the tasks and how participants chose to address them. ProbeTools are also robust enough to return by mail: The feeling of overcoming distance to communicate can also be evocative for design.

It's exciting to see the images and recordings piling up as you download them from the devices' SD cards. We usually organize them by participant and task, but many strategies are possible.

  Sidebar: BATCH PRODUCTION AND DISSEMINATION

We have explored batch production as a tactic for increasing the number of people who can gain first-hand experience with the research devices we design. Over a series of projects, we produced different computational products in increasingly larger numbers—from about 12, to 20, to 35—culminating in 130 technically sophisticated mobile devices for use in a large-scale field study in London [4].

The results were amazingly rich and rewarding. But the cost, both in effort and money, was prohibitive. It took many person-years, and tens or even hundreds of thousands of dollars, to produce even 130 devices. For research purposes, then, our conclusion is that it doesn't make sense to batch produce more than about 20 copies of a research product depending on its complexity. For dissemination, on the other hand, this is far too few. Batch production does not seem a viable option for circulating research products widely. Far more promising, it seems, is to design DIY devices that people can make themselves.

 

Copyright held by authors. Publication rights licensed to ACM.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2019 ACM, Inc.