The Collaborative Spectrum
At the end of 2018, Sam VanHorn and I conducted research to understand information sharing practices amongst teenagers. Specifically, we focused on understanding the attitudes, expectations, and self-reported activities of information sharing of teens with their peers and teens with their parents/guardians. The goal of the research was to understand how both current and historical location information fit within broader contexts of information sharing in order to inform product development. This research sought to contextualize location information within the requirements for sharing it.
As researchers, chief amongst our concerns was structuring data collection to elicit responses that were detailed and meaningful, and, at the same time, acknowledging the discomfort of discussing what one does or does not share with friends and family. That is to say, the subject matter (information sharing) and the participants (driving-aged teens) presented a methodological challenge.
In the end, we adapted a group facilitation method referred to as the spectrum line into a collaborative activity we called the collaborative spectrum. The collaborative spectrum balanced our concerns for voluntary disclosure and power imbalances against those for detailed data collection and quick insights.
Data Collection with Teens
All user-centered approaches to product development advocate for talking to users. But, not all users are the same when it comes to talking to them, and not all topics can easily be talked about with anyone. In the case of information sharing with and amongst teens, both of these caveats are true. Information sharing is a sensitive topic in general and specifically for teens. Although myriad concerns exist, we had two main methodological concerns when developing a research plan and protocol:
- Power Imbalance: How to structure the study that accounts for the power imbalance of age/role between teens/participants and adults/researchers?
- Exposure: How to collect data to avoid individuals feeling exposed by the data collected? In other words, how to support voluntary disclosure relative to sensitive information?
Power Imbalance and Group Interviews
Beyond issues of informed consent, age is a concern when doing research with teens—specifically, with respect to the power imbalance that age differences present. When considering how to structure data collection, we felt that although one-on-one semi-structured interviews are appropriate to gather data about opinions, attitudes, and perceptions, these types of interviews were inappropriate with teens. As a person who conducts a lot of interviews, I am fully aware of how uncomfortable interviews can be. And, as a person who once was a teenager, asking teens questions about what they do and don’t share with my families and friends—meaning what they keep to themselves—can feel invasive and punitive.
As a result and consulting past studies, I decided group interviews were an appropriate avenue. In the most basic way, group interviews means that the researchers (Sam and I) were outnumbered by participants. In this setting, we sought to establish an atmosphere of camaraderie amongst participants, where sharing felt more like commiserating and empathizing rather than being interrogated. Even more, being outnumbered has its advantages a researcher — discussions can be group-directed. When done well, group-directed discussions provide opportunities for participants to teach researchers, and, as a result, give a voice to the expertise of participants (a core tenet of participatory design).
A group interview served two additional purposes from the perspective of data collection. First, groups allow individuals to reflect on their peers’ behaviors, attitudes, and opinions rather than a researcher interrogating a participant’s life. In other words, groups provided spectrums of opinions, and the opinions of others require individuals to negotiate their own. Second, group interviews allow groups to build consensus about key themes, ideas, and issues. Unlike a one-on-one interview, in a group setting, ideas circulate amongst participants, get turned over, and rethought. Obviously consensus reduces individual differences that might surface from one-on-one interviews. As a first stage of research, we felt establishing consensus provided a means to develop hypotheses that can be validated, unpacked, and complexified over time.
Voluntary Disclosure and The Collaborative Spectrum
The nature of the study presented another challenge — how can we ask teens to discuss what they do and don’t share with family and peers? Asking the question outright could result in participants shutting down (“why do you want to know?”) and asking the question to obliquely could introduce interpretive biases. Additionally, asking questions on this topic without sufficient guardrails around the conversation could risk of the conversation veering off into territories that are uncomfortable for participants (and unhelpful to researchers). In short, our concern was with the feeling of exposure that the topic presented — that is, an individual feeling unwilling to share no matter how much we had assured them responses were confidential.
To address this issue, we structured a large part of the group interview around an activity: the collaborative spectrum.
Building a Collaborative Spectrum: “Always, Sometimes, Never”
What Sam and I called “Always, Sometimes, Never” was an activity to build a spectrum of information disclosure. Our activity was based on a tried-and-true group facilitation activity called the spectrum line. The spectrum line is often used during participatory workshops around contested issues, such as public perception of driverless vehicles. The spectrum line fosters discussion about the range of individual opinions relative to an issue by asking individuals to position their opinion relative to two poles. The Seeds for Change facilitation manual explains the spectrum line as follows:
Start by creating an imaginary or real line through the room (chalk or masking tape on the floor are good for indoor spaces). One end stands for “I agree completely”, the other end for “I disagree completely”. Outline the issue under debate and formulate it into a statement to agree or disagree with. Ask people to position themselves along the line according to their views. They may try out several spots before making a final choice.
Using the spectrum line as a starting point, we created our collaborative spectrum activity as follows. On a wall, we placed a strip of blue painters tape. Labels of “Never” (far left) and “Always” (far right) were placed along the line. Above the line we placed a label that read “Friends/Peers” and below the line we placed a label that read “Family/Parents.” Using the wall space, we asked participants, as a group, to decide on the location of slips of paper. The slips of paper corresponded to particular pieces of information. For example, slips included “Phone password”, “Where you are right now”, “Who you are interested in/dating”, “What you did on Friday night”, and so on. These pieces of information corresponded to the behavioral goals of information requests (Smetana et al. 2015).
We had two reasons for structuring the collaborative spectrum as such. First, we wanted to create a durable artifact that we could analyze and show as data. Although pictures of a group positioned along a line could be one such artifact, doing so risked creating a feeling of exposure by singling out individuals for their unique position. When the spectrum line is typically used, individuals are asked to explain themselves in front of the group, which we felt undermined the type of environment we sought to foster. Instead of individuals being data points, the slips of paper served as markers for group consensus. As a facilitator introduced the slip of paper, the group would direct them to move the slip right or left, even using other slips a reference points (“a little more right than phone password”). Each individual was able to contribute but not obligated to do so, and each contribution was meaningful in the moment as the slip moved around.
Second, given the nature of the data, we were curious about the relational aspects of information relative to the audience and other types of information. For example, how did a group place “how you got to an event” relative to other bits of information and for different audiences. Structuring the activity around data collection allowed us to quickly surface patterns about what was expected, optional, and private (Image 2). As time is a constant constraint on industry research, the collaborative spectrum made analysis more straightforward in exchange for subtlety.
Limitations and Adaptations
As mentioned, the collaborative spectrum has its limitations. Like group interviews themselves, the collaborative spectrum as described here leads to consensus. Where differences in opinions are available in our transcripts or notes, they are not shown in the spectrum as a representation of each group. Finding a means to capture differences requires researchers to consider other means of data collection or subsequent studies. One way to adapt the collaborative spectrum for the purposes of showing differences is inviting each participants to place a slip of paper on the spectrum for a given topic.
Another limitation of the collaborative spectrum is that the terms of engagement are bounded by the researcher. In the activity we used, the slips of paper bounded the types of information we discussed, and the meaning of the spectrum itself (from always to never). This bounding was important for our purposes, but also foreclosed the possibility of other types of information. For example, on several occasions groups placed slips off the spectrum entirely — an indication that not all information falls on a spectrum of always to never! Incumbent upon a researcher is considering the trade offs between limiting responses and opening them up. A possible adaptation of the collaborative spectrum to unbound responses is to have participants propose what topics to include (for us, what information is on the slips).
***
The collaborative spectrum is a highly adaptable group activity. Where we used it to account for methodological issues related to power imbalance and exposure, it can be useful in a variety of settings to address other concerns and contexts. Even more, as a means of understanding group opinions, the collaborative spectrum provides a way to materialize consensus and analyze patterns quickly. We hope you’ll find an excuse to use it!
Written by Thomas Lodato
Thomas is a Senior UX Researcher at Verizon Connect. He’s an alum of the Georgia Institute of Technology, where he focused on the future of work. At Verizon Connect, he brings his background in participatory design into his research. He feels most at home on inline skates. For more, follow him: @thomas_lodato
