Mar
13

Research in the Age of Context

posted by: Boaz Gurdin

Scobleizer recently proclaimed that we are entering an “Age of Context.” That’s good news to us in the user research world, because capturing contexts of use and making them actionable for design is what we do. In this post, I want to reflect on methods from our user research toolbox that can be particularly useful for people who want to build context-sensitive apps. In particular, I think it’s time to revisit Grounded Theory as a core method in user research.

Scobleizer uses the term “context” to describe the trend of services like Google Now that anticipate user needs and interact with users differently depending on who they are and where they are. He calls out five proliferating technologies enabling “contextual services”: (1) sensors, (2) wearable devices, (3) big data, (4) social data, and (5) location data.

In the UX world, Jared Spool reaffirmed that we are entering a new frontier of context-aware design. Using examples such as the Apple Store App that has different modes for in-store and out-of-store shopping, he asks how to foreground context in design thinking and design deliverables.

Contextual Inquiry

Contextual inquiry seems like the obvious user research method to use for context-sensitive design – it has context right there in the name!

Contextual inquiry simplifies ethnographic methods into five formalized diagrams:

  1. flow model describing organizational roles and how they interact
  2. sequence model showing individual and group workflows
  3. artifact model showing wireframes of existing designs
  4. physical model showing the environment in which the actors interact, and
  5. cultural model showing goals, values, and attitudes that influence behavior and the cultural groups influencing those beliefs.

Contextual inquiry is a solid foundation for user-centered design, but it rarely leads to context-sensitive design, at least according to the Scobleizer definition. Contextual inquiry is good for generating a complete feature list, but it doesn’t make it easy to see how features should behave differently in different contexts.

Grounded Theory

We need a method that will help us design systems that reveal information and functionality to users in relevant situations, like the Apple Store App that only surfaces retail store features when the user is physically in a retail store. To design these systems, we need to figure out what situations trigger what behaviors. For example, being physically located in an Apple Store is a situation that may trigger scanning a barcode to pay for a purchase. Grounded Theory methods can help define situations and predict outcomes, allowing us to surface features that support and influence those outcomes.

Grounded Theory, developed in the 1960s for analyzing hospital work, is a method for developing theories about how the social world works. The analyst’s job is to figure out what causes what in their domain of interest. Analysts do this by observing real-world situations and constantly comparing and contrasting each situation with previous situations to see what different conditions lead to what different consequences.

The analyst hypothesizes causes and effects, creating a theory that is grounded in his or her observations. The causes and effects can be concrete or abstract, as long as the analyst clearly describes the indicators of the abstract concept, so people can spot the concept when they see it. After making an initial hypothesis of what causes what, the analyst seeks out more and more varied situations, trying to find situations that the hypothesized model fails to predict and revising the model with new variables to improve the prediction.

Learning Grounded Theory

A few key methods from Grounded Theory are:

  • Coding – not computer programming! In Grounded Theory, codes are the categories of objects and actions/interactions you notice in the world you are describing. For example, if you were describing cafes, you could have codes like “drinks,” “decor,” “business talk,” “social talk,” and even more abstract concepts like “connecting.”
  • Conditional/Consequential Analysis – Once you have codes, write and diagram your thoughts about the relationships between them. How do they vary? How do they relate to each other? (X is a type of Y, X causes Y, etc.) How do variations in one or more categories predict variations in other categories? Start thinking about cause and effect. Write questions and hypothesize answers about how causes and effects vary in different situations.
  • Maximum Variation Sampling – To answer your questions about cause and effect relationships, you want to find as many different examples as you can. Try to think up all the different variations of the cause and all the different variations of the effect and look for examples.

Some good resources for learning more about Grounded Theory methods are:

Applying Grounded Theory

With these Grounded Theory methods, you can develop a model of the important actors and behaviors within your domain and their cause-effect relationships. What do you do with this? Well, if you’re an academic social scientist, this is when you write your book. But for the rest of us trying to infuse context into the design process, we now have a specification for the relevant contextual variables in the domain that we care about, and from there we can design systems that are triggered by the conditions specified in our grounded theory, hopefully leading to consequences that better serve the goals of the users and the other stakeholders that we are designing for.

The Future of Context-Sensitive Design

The first wave of context-sensitive apps will likely take advantage of the elements of context that are most readily available using existing technology: apps like the Apple Store App that change modes based on location; apps like the Google Calendar/Maps/Android suite that alarm you earlier about events based on traffic conditions; apps like the Yahoo homepage that personalize content and advertising based on who is viewing.

Using Grounded Theory, we can go beyond basic technology-inspired contexts like location and start designing for a wide variety of domain-specific, socially relevant contexts. By paying attention to how user behavior varies by conditions, we can develop tools that more effectively help users achieve their desired consequences.

In the Age of Context, it’s time to incorporate Grounded Theory methods into our design process.



One Response to “Research in the Age of Context”

March 14th, 2013 at 3:43 pm

[...] serve fundamentally different media needs for me. I use them in completely different contexts (hey, there’s that word again!). Twitter is immediate: What are the people I care about personally or professionally talking [...]

Leave a Reply