For this post the subject of glanceability will be described, discussed and reflected upon based on the three seminar texts. The papers are called: Designing and evaluating glanceable peripheral displays, Exploring the design space of glanceable feedback for physical activity trackers and Evaluating peripheral displays, they all present conducted research with in the field based on glanceability and peripheral displays. 

Designing and evaluating glanceable peripheral displays

Summarizing the text in one sentence it is about increasing knowledge among designers when it comes to designing and evaluating glanceable peripheral displays, something that is done through a set of guidelines for designing as well as a framework.The text is making a qualitative analysis since what is described is 10 in depth interviews rather than 100 on the surface interviews. Of course establishing if a text is qualitative or quantitative can be hard sometimes, this because some research is based om mixed methods, meaning that both numbers and a deeper meaning behind the results is presented.  

Authors write that peripheral displays can positively impact task flow, resumption and reacquisition. This because they are about maintaining the primary task flow,  this by monitoring secondary tasks to determine whether or not to resume or switch between tasks. One of the core design goals when it comes to peripheral displays is that the display should be glanceable and non-interruptive, in this, glanceability is important because it improves our ability to use a display operationally. In addition to this it improves multitasking.  The main characteristics of glanceability described in the paper are the following:  Symbolism, dimensionality and entity. Perceptibility, interpretability, memorability and aesthetics. Abstraction is used, its distinct, matches user expectations, and maintains consistency. 

As mentioned, peripheral devices improve our ability to multitask, this because we can just quickly glance at them while doing something else. According to the author, increasing the knowledge into this type of devices can provide a better support for multitasking. They are providing guidelines for designing for these types of displays, using the wealth of abstraction techniques (change detection, feature extraction), design variables (colour, shape) and design characteristics (dimensions, symbolism). This type of displays helps for example information workers with balancing tasks. 

One of the problems brought up in the text is the lack of consistent terminology, this can be a problem in a few different ways. When people speak from different contexts or grounds when they refer to things or words, meaning the same or different things. It can slow down the process of developing concepts like these, since the meaning behind what is said is unclear. Of course developing a consistent terminology can be hard, who decide on which word to use or stick with? It is something that is worth focusing though, both because it provides a common ground between designer as researchers and because it provides a better ground for the research conducted. 

The goal of the paper is to inform the design of displays (like task management), authors believe that improving the glanceability of these displays will enable more efficient for multitasking. The goal is to study something that can later be designed. What is important to note is the difference between glancing and peripheral vision, when you glance at something you have to look down or to the side, peripheral vision is about more than that, it’s about knowing and feeling what’s around you, acknowledging the space you live in.  According to the authors, glancing is refers to the perception and interpretation after the user is paying attention to the interface.

The methods used in this analysis are mixed, a literature survey that they interviewed designers from was conducted. They also planned two empirical studies to measure the glanceability of abstract representations with different design characteristics. In addition to this they explored abstraction techniques in the lab. In the study the conducted interviews were performed with 10 peripheral display creators.  

Exploring the design space of glanceable feedback for physical activity trackers 

This text is about how to best design glanceable behavioural feedback in activity trackers. According to the authors glanceable feedback for behaviours change should be abstract, integrating with existing activities, supporting comparisons to targets and norms, being actionable, having the capacity to lead to checking habits and to act as a proxy to further engagement. Glanceability is defined in terms of how quickly and easily feedback is able to convey information after one pays attention. Behavioural feedback should be displayed for a limited amount of time, reinforcing re-engagement habits. Feedback should be short and quickly apprehensible. Summarizing what the authors mean by glanceable behavioural feedback, it is about providing feedback of the users behaviour in a device such as a smartwatch that can quickly be glanced, this feedback can come in the form of how many steps that has been taken that day or how active the user has been. The goal being to affect the users behaviour, making him or her being more active or taking more steps. The design space of glanceable feedback is about how we as designer chose to design the interface, something that the researchers behind this text has done in a few different ways. 

The design qualities presented in the paper are the following: 

Abstract: abstracting data allows the user to process and perceive information with minimal consciousness, enabling quick awareness and reflection on ones behaviours. (Step count through abstract forms such as circles).

Integrates with existing activities: embedding feedback into frequently occurring activities makes the feedback more likely to be glanced. Users check their smartwatch 60-80 times a day, they therefor decided on checking time in the study, feedback being placed in the background. 

Support comparisons to targets and norms:  Feedback that presents progress in comparison to a target can be easier for the user to process, helping the user to evaluate their behaviour compared to a goal rather than presenting raw data. Comparing results with others, with the norm. 

Actionable: Effective glanceable feedback interfaces should not only inform but also instigate short, goal related actions. Suggests actions such as taking the stairs or moving. 

Leads to checking habits: The frequency of glances as well as the overall engagement with feedback decreases over time. Might make the individual relapse. Glanceable feedback should be able to sustain the frequency of glancing over the long run. Two approaches for the creation of checking habits: Novelty(the feedback constantly provides new information) and scarcity (suggests that checking habits may be created if feedback is turned into a scarce resource.. Limited seats and such. 

Acts as a proxy to further engagement: Individuals quickly lose interest in deep data exploration. Glanceable feedback can be designed with the goal to create aha moments, acting as cues for further engagement with the feedback. Presenting  feedback that raises questions rather than provides answers. 

So what are the advantages of glanceable feedback? And how can it be used to affect user engagement? According to the authors looking at the watches, seeing how much (or how little) you have moved during the day can make you as a user engage more, maybe walk home instead of taking the bus, all because of what your watch told you. One advantage of glanceable feedback that is described in the text is that it requires minimal attention to be perceived and processed. They also argue that It is also proven that glanceable feedback can increase long term commitment to physical exercise and that monitoring of behaviour can lead to early correction of slips and relapse. This because It reminds and makes the user aware and affects his or her behaviour.

In the study they used 12 participants (all male with the median age of 25) and the tests were conducted over a 28 days period. They were exposed to each of the interfaces for seven days. This period of time might be enough for investigating the experience beyond the initial reaction, but if they’d want to investigate the behaviour more time would have to be spent. It gives short term results rather than long term. The goal was to compare concepts in terms of their adaption, how participants engaged with them, and what impact they had on their physical activity. Personally I think that the study should have had a broader range of participants, including women and different ages. They did include different countries but I don’t think that’s enough. One option could be rephrasing the research question as well “What impact does activity tracks have on the activity of males within the age of 25?” better phrased of course, but something like that. 

They chose their concepts based on different criteria’s. Then they developed the interfaces as watch faces for android. Normly and ticktok, and goal competition and gardy. To evaluate the results of the study, skype interviews introducing each new interface as well as talking about the usage and experience of using the last one were conducted, these interviews lasted for 15 minutes. The results were that this type of activity tracker works best for people who have the will but not the means to change their behaviour. The participants checked interfaces 107 times which is higher than previous studies.  80% of all usage sessions were glances which lasted for 7 seconds, which was shorter than previous studies. Used to check the time or notifications. Participants felt motivates to walk when sensing the could catch up with others or stay ahead. This motivation disappeared when the differences grew bigger. Having to check in for the walking goal every hour made for more frequent interactions as well as more walking. Users wanting credit for their walks. People either hated ticktock or liked it, depending on how active they are in themselves. Trying to figure out what is going on but then it’s not interesting, participants get bored. People want to see the numbers. If you have walked a lot you won’t walk that much more, if you are further away from your goal you will walk more. Tictock, normly was more unmotivating in a lack of numbers. Displaying behavioural feedback for a limited amount of time (ticktock) let participants to re-engage and walk more frequently. Feedback about competition of traditional step goals best supported reaching ones target step count. 

During the seminar it was suggested that this experiment/research could be conducted again in this day and time, but made collaborative, engaging more than one user in the tracker. I found this idea really interesting, I own a smart watch myself and I am a big fan of the numbers but sometimes I do get unmotivated, and getting notifications about that it is time to be active can sometimes be more annoying than good. The thought of being able to engage with my friends however, is intriguing and I feel like it would encourage me at least to be more active. 

Evaluating peripheral displays 

The last text is about different criteria’s and techniques for evaluating peripheral displays, it measures the techniques pros and cons as well as how well the achieve their goals. Some of the questions I prepared for the seminar are presented below: 

• Why is awareness and distraction hard to measure? Could there be other things other than those brought up in the text that are “better” for measuring?
• How do the presented models differ from each other?
• What are the criteria presented in relation to the activity theory framework and what do they mean?
• What do they mean by stating that the methods used are holistic? 
• When how and why can CUEPD be used?
• How do the use of the presented methods differ from each other?
• How can awareness and distraction be measured and why? How are the terms defined?
• Why does the email domain benefit from peripheral displays? Is there any other domain that could benefit from this too?
• What is the purpose of the study?
• Were the used methods and models appropriate for the purpose of the study?
• Which are some of the mentioned limitations about the techniques mentioned in this study? 

Authors describe how evaluation criteria and techniques are still being created for peripheral displays, in doing this they describe models and methods that have been tailored to specifically to evaluate peripheral displays (measuring how well they achieve their goals). Authors then continue with presenting criteria used in past evaluation of displays, regarding issues such as learnability and distraction. After explaining how these criteria’s have been addressed in the past, a case study is presented (two email peripheral displays) that demonstrates the pros and cons of various evaluation techniques.  

Peripheral displays require a new set of criteria related to attention issues, since they aren’t usually the focus of user attention and always used while multitasking, in the text awareness and distraction is presented as criteria in relation to these issues. Awareness refers to the amount of information shown by a display that the user is able to register and use. Empirically this can be measured through behavioural change. In this study it is measured by knowledge questions and self-reports. Post usage questionnaire. Distraction refers to the amount of attention a display attracts away from a user’s primary task. Measured in lab studies in terms of directly observable properties of the users behaviour such as changes in performance on a primary task. In this study it is measured by system logs focusing on performance changes in the primary task resulting from display use. Speed and accuracy. Self-reports and interviews. Gathering data about awareness and distraction is challenging, awareness is difficult to evaluate because interactions with peripheral display are brief and changes in behaviour may be small and unnoticeable. Distraction is hard to measure without further distracting users. Here it is argued that peripheral displays have the following characteristics: provide realistic usage experience and probe the use of the display according to trade-offs among interruption, reaction and comprehension.  

Two evaluation frameworks and two methods have been created specifically for peripheral displays. These criteria’s should be taken into consideration when designing and evaluating peripheral displays. They attempt to standardize questions that evaluators ask about their displays to gather data about important criteria. These are the IRL model as well as the activity theory framework. The IRL model classifies different types of notification systems. User goals are based on the interruption, reaction and comprehension caused by a system. Can be used to suggest useful empirical and analytical evaluation metrics for tailoring usability methods. Its targets being interruption, reaction and comprehension. The Activity theory framework has the following criteria: appeal, awareness, distraction, ability, effects of breakdowns. Design dimensions relative to the framework are scope (number of activities supported), classes of activities supported (primary, secondary, pending), criticality (from low to high importance). 

The methods described are the following: CUEPD which stands for context of use evaluation of peripheral displays. This is an evaluation method that relies on active user participation and emphasizes the experience of using peripheral displays, this by user scenario building, enactment and reflation. The method can be used when there is a working prototype to improve future designs. It suggests survey question categories such as noticeability, comprehension, relevance, division of attention and engagement. The other method is a modified version of Nielsen’s heuristic evaluation method and reflects ambient display design goals. It is meant for use in the early stages of design, suggesting usability goals to the designer as they iterate.   

In this they present two broad categories important to peripheral display evolution:  traditional usability and attention to user issues. Some of the most common graphical interface user criteria (traditional): Effectiveness, efficiency, learnability, memorability, flexibility, usefulness, user satisfaction. From these, learnability, error visibility, usefulness and user satisfaction will be used in relation to peripheral displays. 

Learnability is the amount of time and effort required for users to operationalize their use of a peripheral display. Error visibility and recovery refers to how apparent errors or breakdowns are to the user and how easily users can recover from them. (since the displays aren’t that visible users might not notice errors). Users should be able to distinguish between an inactive display and a broken display. Usefulness is about the degree to which the display provides value to the user. Measured through interviews. User satisfaction refers to a user’s qualitative happiness or unhappiness with a display. Authors argue that user appeal is an important criterion, which is measured through surveys and interviews. Design mechanisms presented to accomplish many of the criteria are abstraction, glanceability, user customization, sufficient information design, easy transition into more in depth information, consistent visuals.

In the text a case study is conducted and it’s about two e-mail display evaluations that shows the pros and cons of techniques. In the study they showed sender groups in emails so that the user could determine if it’s important enough to interrupt the task. Which was thought to give them the chance to multitask better. In this the authors interviewed 10 administrative assistants about their needs for managing emails. Then an evaluation of both Tinker and the Orb was done with 6 graduate students. They were given descriptions and images of the displays in two scenarios. 1. Notification version of each display that would change only when a new email arrived. 2. Ambient version that cycles through information about the emails. 5/6 preferred the notification version since it gave away a clear notification when a new email arrived, and it was believed to be less distracting. Based on the research they iterated on the designs. Orb changed colours and brightness. Ticker, rolling, notifications 25s, summary text. Orb users had lower awareness than the ticker users. Speculate it has to do with learnability, the orb took longer to learn. The field study had four participants, two who tested tinker and two who tested the orb. In total the study had 26 participants between 18-23 (half testing orb and half testing ticker).

This study highlights the importance of evaluating peripheral displays using multiple methods to gather data about the six evaluation criteria: awareness, distraction, learnability, error visibility, usefulness, and user satisfaction. 

Final reflections on the texts and subject: