1 Introduction

From the confluence of two social phenomena—an aging population and an increased dependence on digital technology— a critical need has emerged for older citizens to develop digital literacy. Over the next 35 years, the United States is projected to experience rapid growth in its older population, with nearly 20 percent of the population being 65 or older by 2030 [1]. Although older adults are also the fastest-growing group of Internet users in the United States [2], 41 percent of older adults still do not use the Internet at all [3].

In a review of the literature on computer use and older adults, Wagner, Hassanein, and Head [4] discovered a variety of reasons for the low level of digital adoption. Some have to do with a lack of access to computer hardware or the Internet, and others are related to physical and cognitive changes that people experience as they age. Melenhorst, Rogers, and Bouwhuis [5] argue that while it is widely believed that cost is one of the main factors that deter older adults from using technology, a larger reason is the lack of perceived benefit: older adults either do not find that the technology meets their needs or do not understand the technology well enough to appreciate what it can do for them. Mackie and Wylie [6] agree that technology acceptance depends on several factors, including the degree to which the technology meets the user’s needs and the user’s awareness of the technology and its purpose. A study by Czaja, Guerrier, Nair, and Landauer [7] confirms that older adults are more likely to use computers if they perceive the technologies and the tasks for which they can use the technologies to be useful.

2 Barriers to Digital Literacy

Listening to older adults and addressing their learning needs as we develop the technologies of tomorrow is a moral imperative. Since 2011, the Breaking Digital Barriers group at Michigan Tech has run a program called BASIC (Building Adult Skills in Computing) that pairs our students with community members, most of them 60 years of age or older, who are seeking help with computing technology. We have reported on some of the recurring themes in our interactions [8]:

  • Anxiety stifles exploration. The experience of using a computing device is known to cause anxiety in older people, and our experiences bear this out. A common concern for participants in our group is that something they do will “break” their investment. Even routine activities cause anxiety as users fear accidentally going “off script”.

  • Danger online. Many learners are fearful of going online because of stories of fraud and identity theft they have heard in the media. Without a basis of understanding for how malware and other threats work, they have no model for how to minimize their threat level.

  • Context sensitivity and non-obvious affordances. A shift toward mobile devices with small displays and a shift toward “clean” design have led to a decrease in affordances and other cues in user interfaces. To use these interfaces effectively, the user must be willing to explore the space and uncover the functionality. A change triggered by an inadvertent action makes older users feel anxious and out of control.

  • Details obscure abstraction. Not so long ago, users typically accomplished activities like email and word processing through dedicated applications specific to a particular personal computer. The movement toward mobile computing devices and cloud-based storage and applications has abstracted those activities into general “services”. For many of our learners, conceiving of computing in this abstract way runs contrary to their script-based style of learning about computers.

  • Functionality across devices. Most older adults do not use the same services across different devices; rather, they use different devices for different tasks. For example, seniors might not use a tablet to check the weather if they associate that ability with a PC. Also, many older users do not realize that content on the Internet, especially “cloud” services, is accessible and consistent across devices.

Issues like these cannot be explained satisfactorily through traditional factors like age-related cognitive, perceptual, or motor changes. Usability tests focused purely on external behavior, like eye tracking or measuring response time, are insufficient. In general, a focus purely on end goals (“completing the task”) without taking the method of learning into account will hide important cognitive and social barriers to digital adoption.

3 Our Social-Cognitive Approach

Bandura’s Social Cognitive Theory (SCT) informs our approach to digital literacy [9, 10]. According to this theory, an individual’s functioning is the product of an interaction between cognitive, behavioral, and contextual factors. It emphasizes the social context of learning and the importance of observation. In opposition to a behaviorist approach, learning and the demonstration of what has been learned are separate, so learning involves not just the acquisition of new behaviors, but also acquisition of knowledge, cognitive skills, concepts, rules, values, and other cognitive constructs.

SCT provides a foundation for interventions designed to improve people’s learning. Below we review several key SCT principles [9, 10] and describe how our instructional practices are connected to each.

  • Observational Learning/Modeling. SCT’s most basic instructional implication is that learners require access to models of the knowledge, skills, and behaviors they are expected to learn. Multiple types of models (e.g., instructors, peers) and various forms of modeling (e.g., cognitive, verbal, mastery, coping) should be used. Instruction must support learners’ engagement in observational learning.

    Our BASIC program pairs each learner with a tutor who models behaviors and strategies that we hope to reinforce in the learner. The simplest form of modeling is when our tutors demonstrate how to conduct an action. To make their actions and intentions more obvious, they typically vocalize what they are doing while they are doing it.

    Tutors model not only behavior but also problem solving and exploration. We believe it is important for our learners to see that learning the skills to find the answer is often more important than knowing the answer. Tutors also model their emotional reactions to not knowing the solution. It is critical for tutors to demonstrate to learners that, while it is reasonable to feel annoyed when something is not easy to figure out, there is no need to feel anxious; there are ways to find a solution and recover from mistakes.

  • Outcome Expectations. Instruction should help people see that situated learning and the demonstration of that learning lead to personally valued or important outcomes. Lessons should emphasize real-world applications and the relevance of material to the learners’ lives.

    Rather than delivering predefined training sessions, we invite learners to come to us with their specific needs. We then tailor our one-on-one tutoring sessions accordingly. Addressing the learner’s specific problem often affords opportunities to address specific digital literacy competencies along the way. In this way, the skills we teach are tied to problems of personal value to our learners. We also invite our learners to bring their own devices to our tutoring sessions. This ensures that learners are developing skills on the devices they will be using in their day-to-day lives. To support an understanding of functionality across devices, we often encourage learners to use multiple devices in the same session. For example, when a learner is interested in learning about the cloud, we may work with him or her to synchronize files across devices and access information from the cloud from a variety of platforms and devices.

  • Goal Setting. Instruction needs to help students set effective goals—goals that are attainable, clear, specific, and moderately challenging. To facilitate progress and self-efficacy, learning goals should be attainable with moderate levels of effort. Goals that learners set or endorse themselves have a bigger effect on their behaviors than do goals that are assigned.

    As mentioned above, our tutoring sessions are driven by the requests and goals of our learners. Even in the cases in which learners specify a general goal of “learning about computers”, our tutors spend time talking to them about their lives to identify potential computing needs and to choose learning goals that may be most relevant to each learner’s life.

  • Perceived Self-Efficacy. People will be more active, effortful, and effective learners when they are confident in their ability to complete tasks successfully. Instruction should be designed to help learners develop and sustain self-efficacy: the belief in one’s capabilities to organize and execute the courses of action required to manage prospective situations.

    Our learning sessions are hands on, and whenever possible, we ask the learner to “drive”. Although we may model behavior by demonstrating a sequence of steps, our tutors ask the learners to repeat the steps themselves to help ensure they will be able to address the problem on their own at home. As noted above, our tutors do not always know the answers to a learner’s problems and may need to seek assistance online or from another tutor. In doing this, we hope to reinforce that even “experts” need help finding the answer and that having questions about how to do things on a computer or handheld device is normal and not something to feel ashamed about.

  • Self-Regulation. All students should be supported in their efforts to be self-regulated learners. Three processes involved here are self-observation (monitoring one’s own behaviors and outcomes), self-judgment (evaluating whether one’s actions are effective), and self-reaction (responding to the self-evaluations by changing, rewarding, or discontinuing the behavior). Instructors can promote self-observation by helping people learn how to monitor different aspects of their learning behavior through aids such as checklists.

    Our program currently offers only limited support for self-regulation. Tutors provide models of self-observation, self-judgment, and self-reaction, and they can encourage similar behavior in learners, but the limited contact time makes it difficult to practice these behaviors. Future work may include self-guided learning tasks done outside of the group sessions. Also our planned tool support for wayfinding (discussed in the next section) offers an opportunity to record learning progress and present it to the learner.

4 Future Directions

Talking to seniors is important, but so is talking to the tutors with whom they work. Community educators around the world are using novel approaches to address the numerous socio-technical barriers facing older adults, and they have a wealth of knowledge that can be used to improve educational programs. In many cases, however, their knowledge is tacit and must be elicited. We have adapted an incident-based cognitive task analysis technique, the Critical Decision Method (CDM) [11, 12], to elicit this tacit knowledge from our own tutors. The CDM uses a three-sweep, semi-structured interview to help experts tell stories from their field. We plan to expand this interview project to tutors from other programs in North America. We will use the experiences of these community educators to formalize our training program, identify opportunities to improve our current program, and motivate the design of technology to support the acquisition and maintenance of digital literacy.

One direction we are pursuing is the development of digital tools to support online wayfinding behavior. Wayfinding is the process of “organizing and finding a way through dynamic explorations and analyses [13]”. Our observations suggest several barriers that older adults, in particular, may face when wayfinding online. These include the high level of visual complexity in online interfaces, obscure or hidden affordances, and the difficulty of remembering past wayfinding successes.

We propose a socio-technical approach to developing wayfinding skills among older computer users, through a scaffolded adaptive web navigation tool used in conjunction with small-group, active-learning sessions. The technical tool at the heart of this learning activity is a browser plugin that will provide just-in-time guidance, highlighting wayfinding options with a high likelihood of success based on analysis of web page structure and previous navigation activity. The tool will also allow users to annotate their wayfinding actions and share them with other learners.