Keywords

1 Introduction

Older adults are the fastest growing group of Internet users in the US [8], but 41% of older adults still do not use the Internet at all [12]. Several studies have indicated that many older adults are ready and willing to learn about technology, even though it might involve a great deal of time and effort on their part [2]. The benefits clearly outweigh the time and effort required: technology can greatly enhance the lives of older adults by providing social interaction, tools for banking and shopping from home, entertainment, and ways of getting information about healthcare or even contacting healthcare providers. The issue is that older people must be able to acquire the knowledge and skills they need to use technology effectively [11]. Training is critical, especially as success at initial training is one of the primary predictors of whether or not adults continue to use computers [10].

Training can help older adults understand their devices and the software that runs on them, but many psychologists believe that the training needs to be targeted to take advantage of the strengths and compensate for the weaknesses of the adult learner [7]. Several older adult learning training models have been developed. For example, the Industrial Gerontology Model specifies five factors that should be considered when developing training programs for older adults: motivation, structure, familiarity, organization, and time [15]. A meta-analysis found that the factor contributing to the largest amount of variance in adult learner performance is self-pacing [5]. Self-paced, goal-oriented training is also recommended because this type of training minimizes working memory demands, which can be an issue for older adults [9].

Our BASIC (Building Adult Skills in Computing) program [1, 13] is one of many nationwide initiatives aimed at improving the digital literacy of older adults. In contrast to more standard prescriptive, task-focused programming, our self-paced tutoring program focuses on developing digital flexibility, problem solving abilities, and self-efficacy—higher-order skills that are essential for keeping up with the rapid pace of technological change and transferring knowledge from task to task or system to system. Our approach, grounded in Social Cognitive Theory [3, 4], emphasizes the roles of observational learning and behavior modeling, learner-led goal setting, and the development of self-efficacy.

Nearly all of the training programs cited in the literature are formal or workplace-based classes rather than the one-on-one tutoring sessions that our program offers. However, our program was designed to incorporate many of the characteristics noted above, such as self-pacing and goal-orientation. As we work to formalize our program, we wanted to discover which other techniques and strategies have emerged from tutoring sessions led by our most experienced tutors. For example, many of our tutors have developed analogies that they frequently use to help learners relate new technology to older systems. Others have honed their skills identifying and addressing errors in learners’ mental models of computing. To elicit this expert knowledge, requires a systematic approach. The Critical Decision Method (CDM) [6] is a three-step, semi-structured cognitive task analysis technique that has been successfully used to drive the development of expert systems and training programs.

In the current project, we have adapted the CDM protocol to elicit knowledge from experienced digital literacy tutors in our program. This paper describes our protocol, outlines the steps involved in our tutoring process and the goals of each step, and presents the data we gathered from a set of 12 interviews we conducted with tutors of varying backgrounds. The interviews yielded a wealth of information, and our goal was to classify and organize the data to reveal common themes, tutoring strategies, and obstacles to successful tutoring.

This study was guided by the following research questions:

  • What commonalities exist in tutors’ stories about their tutoring sessions?

  • Which types of strategies are most often used to guide adult learners?

  • What types of barriers arise that could derail the patron’s learning?

  • What are the key attributes of successful tutoring sessions?

The purpose of this narrative study is to analyze the activities, strategies, and tools used to help solve technological issues for older adults in BASIC tutoring sessions.

2 Methods

Participants. Participants included six tutors from our BASIC program, each with two to six years of tutoring experience with older adults. All were current or former Michigan Tech undergraduate (1) and graduate students (5) and faculty (1). Each tutor was interviewed twice, resulting in a total of 12 interviews.

Procedure. The Critical Decision Method (CDM) is an incident-based cognitive task analysis technique. The purpose of the technique is to elicit expert knowledge using a semi-structured interview to help experts tell stories from their field. The interviews were conducted in three sweeps.

  • Sweep 1: Selecting an Incident. During Sweep 1, we asked the tutors to identify several memorable tutoring sessions. From those sessions, we selected one successful session and one unsuccessful session for further discussion. Success was defined as follows:

    • A successful tutoring session was one in which the tutor (1) understood the patron’s request, (2) identified the source of the problem (technical, educational, etc.), (3) determined a solution to the problem or an answer to the question, (4) explained the information clearly to the patron, and (5) felt confident that the patron left the session better equipped to deal with a similar situation in the future.

    • An unsuccessful tutoring session, in contrast, was one in which the tutor either (1) did not understand the patron’s request, (2) was unable to clearly explain information to the patron, or (3) did not feel confident that the patron left the session better equipped to deal with a similar situation in the future. Note that we were not necessarily interested in whether the tutor successfully solved the patron’s problem. In some cases, doing so may have required advanced technical support or even a new computer. These situations would not be classified as unsuccessful unless the tutor was unable to communicate with the patron as to why he could not solve the problem at that time.

  • Sweep 2: Constructing a Timeline. After the events were selected, we worked with the tutor to collaboratively draw a timeline depicting the major events that occurred during one of the tutoring sessions. For example, we asked the tutor to consider what happened when the patron arrived, the order in which questions were asked and addressed, any interruptions that occurred during the session, etc. As we moved into Sweep 3, we used the timeline to guide our interview and to elicit additional details about each of the primary events.

  • Sweep 3: Deepening. In Sweep 3, we went beyond the surface details to focus on what the tutor was thinking during the session. For example, we were particularly interested in the information that our expert tutors use to guide their decision making throughout the help session. How does the tutor identify the patron’s knowledge gaps? How does the tutor determine the best way to explain a difficult concept? If there are multiple solutions to a problem, how does the tutor determine which solution best meets the patron’s needs? Sweep 3 also aimed to identify common knowledge gaps and socio-technical barriers that patrons frequently encounter.

    CDM recommends several categories of deepening probe questions that have proven effective in eliciting expert knowledge [6]. We adapted those questions to our specific domain. A table of these probe questions is included in the Appendix A.

After Sweeps 2 and 3 were completed for one tutoring session, participants were given the opportunity to take a break, and then the process was repeated for the second session. The pair of interviews took approximately two hours to complete, with each interview ranging from 21 to 70 min.

Coding and Analysis. Interviews were digitally recorded, transcribed, and then independently coded by a pair of researchers (the first and second authors). The initial coding focused on a combination of preset and emergent codes. Preset codes included several broad categories, including cues, information use, analogies, standard operating procedures, goals and priorities, and mental models. After the researchers discussed and combined codes, the coded text was re-classified along three dimensions: step in the tutoring process, agent (patron, tutor, task, device, communication, and session), and attribute (agent characteristic, barrier, strategy).

3 Results and Discussion

Analysis of the refined classification codes revealed a sequence of five essential steps to the tutoring process. Not all of our tutors follow these steps exactly, but the majority incorporated the described practices at some point in their tutoring process. These steps are illustrated in Fig. 1. Below we describe the process in detail, and discuss common barriers and strategies at each step.

Fig. 1.
figure 1

Model of tutoring process

Full size image

Step 1: Introduction

Tutoring sessions should begin with an introduction between tutor and patron.

Goal. The goal is to establish a working relationship between the tutor and the patron and to determine the patron’s motivation for using technology.

Description. The introduction should begin with the tutor introducing herself to the patron and asking some non-intrusive questions (where are you from, is this the first time you’ve used our program, what brings you in today, etc.). Much of this step might take the form of small talk, with the patron also asking questions of the tutor. The tutor should use this exchange to determine the patron’s motivation for using technology. For example, some patrons want to use email or social media to stay in touch with family or friends, while others want to apply for jobs via the Internet. Some patrons may have purchased the device on their own and want to put it to use. Others may have received a device as a gift or feel forced to adopt the technology because there is no alternative means to their desired outcome. Armed with this information, the tutor can begin to assess the urgency or importance of the patron’s issue and identify some of the non-technical barriers they may encounter during the session.

Barriers. Patron-tutor rapport and trust is the foundation to a successful tutoring session. Accordingly, tutors tend to struggle when patrons are passive, difficult to talk to, or seem to dislike small talk. Interestingly, the patron’s motivation for seeking help and the nature of the desired task can be barriers to building rapport. Patrons who seek assistance to achieve a singular high-consequence outcome, like applying for a job or making money on Craigslist or eBay, may not really be interested in learning and may become easily frustrated by small talk that they perceive as unrelated to their personal agenda. This can be compounded when patrons arrive at the library already overwhelmed and stressed out by their circumstances.

Strategies. To build rapport, tutors must be patient. Tutors should engage in small talk, but spend time carefully listening to identify the patron’s specific interests. These will be critical later in the session as the tutor tries to connect aspects of the digital tasks to topics and activities that the patron will find personally relevant. Reading the patron’s non-verbal cues can often be key to establishing a strong relationship and choosing appropriate tutoring strategies. Some tutors may be more adept at this than others, but tutor training programs should emphasize how to identify subtle body language, like cautious or tentative movements and tensing, that may indicate the patron a novice or particularly stressed.

Step 2: Triage

Once the patron-tutor relationship has been established, the next step is to assess the situation.

Goal. The tutor’s primary goal at this step is to gather information to develop a mental model of the patron’s desired task and level of digital literacy.

Description. To begin, the tutor asks the patron to explain verbally, or demonstrate on the device, what he would like to do. The patron’s explanation and interactions with the device provides the tutor with many useful insights about the patron, task, and device—each of which may drive the tutor’s approach in later steps. As the patron discusses or demonstrates the task, the tutor should consider the following:

  • Patron.

    • What is the patron’s general level of digital literacy?

    • How much is already known about the specific task, relative to how much needs to be taught?

    • What is the patron’s style of communication?

    • Is the patron forthcoming or does the tutor need to actively elicit information?

  • Task.

    • How complex is the desired task?

    • What underlying knowledge or skills are required to complete the task?

    • Can the desired task be accomplished within the time allowed by the tutoring session or should the task be divided into smaller subtasks? If so, which subtasks should be prioritized?

    • Does the task require a one-time change or fix that the patron will never need to do independently?

  • Device.

    • If the patron is using her own device rather than one provided at the tutoring site, the tutor should inspect it or ask questions to determine its age and condition.

    • Does the tutor have experience with that device or have a similar device on hand?

Barriers. The special terminology of digital technology can be a significant barrier. Terms differ between platforms and software services, and they are in a constant state of flux. Misinterpretation of terms can lead to divergences in understanding that are difficult to identify and roll back.

The fast pace of change in the technology is another barrier. Patrons may experience a constant barrage of small updates or a sequence of intermittent and disruptive major upgrades; they often ignore notifications of such changes, and express frustration and lack of motivation after upgrades cause noticeable changes to established workflows [14]. Hardware and software may be old and outdated—and given many patrons’ low level of disposable income, they often are. The incompatibilities faced when trying to use new services on an outdated platform can be confusing and demoralizing.

Patrons’ digital literacy levels vary significantly; when they register lack of understanding, it can be unclear what tutors should ask to move the process forward. Cognitive and motor issues can also exacerbate other barriers, as can emotions of fear, embarrassment, and anxiety.

Strategies. Tutors must start out slowly with patrons. A best practice is to listen actively, with neither tutor nor patron using a computing device. Tutors should ask probing “why” questions and encourage patrons to articulate the problem, using their own vocabulary. Through these articulations, tutors try to gauge patrons’ level of digital literacy. Cues that may indicate low levels of digital literacy include: use of non-technical or imprecise language; difficulty explaining a problem, even if capable of demonstrating it; little or no experience with email; only knowing one way to access a website (using a shortcut on the desktop to access email); lack of knowledge of common email or word processing functions, like the distinction between “reply” and “reply all” or text formatting options; and slow or tentative mouse, touchpad, or finger (on a touch screen device) movements.

During this time, it is important to manage expectations, set time limits, and prioritize goals.

Step 3: Planning and Preparation

Once the tutor believes he has a clear understanding of the problem space, it is time to begin addressing the problem.

Goal. The goal of this step is to develop a plan for helping the patron and determine the information, skills, and resources needed to implement the plan.

Description. Due to the breadth of issues and questions that patrons bring to the tutoring sessions, the tutor will frequently encounter a task or device with which he is unfamiliar. In these cases, the tutor may need to look up information on the Internet, explore possible options on the device, or consult with another tutor. In situations in which the patron is eager to learn and the process required to find the answer is not overwhelming, the tutor may choose to model how to find the solution, using a talk-aloud strategy to explain his thoughts and actions. If the patron seems overwhelmed or if the process of looking for a solution is unnecessarily complicated, the tutor may simply choose to look for resources independently. The tutor then uses the gathered information to decide how to proceed in helping the patron. If multiple options are available, the tutor may select one based on several factors, including the option’s complexity, the time required to teach the option, and the patron’s preferences. A final requirement of this step is to create a safe learning environment. This may involve preparing test documents or dummy accounts for patrons to use as they learn the task and practice their skills.

Barriers. Numerous barriers may present at this step, especially if the tutor has had difficulty understanding the patron’s description of the problem. Sometimes the task or goal that the patron describes is quite different from the one actually wanted. This may not be readily apparent until after the tutor has developed a plan of action. In other cases, the patron may have provided inaccurate or incomplete information, or the problem may be so complex that the tutor has difficulty prioritizing the steps or determining where to start. Finally, when exploring a patron’s device in search of solutions, the tutor may encounter unexpected issues. One common problem is a lack of organization of digital content. Files may also be poorly named. Across a wide range of tasks, tutors reported that they needed to introduce file management skills and work with patrons to set up new folder structures and file naming conventions.

Strategies. Although some patrons may be surprised to learn that the tutor does not automatically know the answer to their questions, many seem delighted to have “stumped an expert.” Tutors should be encouraged to tell patrons when they do not know the answer, precisely so they have the opportunity to model their process of finding the solution. As the tutor explores the device or software, he or she should talk aloud and take the opportunity to introduce new terminology. If a patron has been using non-technical language (for example, a patron may refer to documents as “letters” or refer to Microsoft’s Internet Explorer or Edge browser as “e” due to the form of the common icon for that application), tutors can mirror this language, but also take the opportunity to connect these terms to their technical component. Another benefit of modeling this search for solutions is that the tutor can reinforce that anyone can find answers through the use of “regular” language for searching.

This step also presents an opportunity for tutors to commiserate with the patrons. When tutors encounter poor design or clunky solutions, they can express their own frustration. In doing so, tutors can help reduce the patron’s stress and anxiety and reassure them that this sort of frustration is common and normal.

Finally, tutors should be encouraged to embrace a sandbox philosophy of play and practice. Creating test documents and dummy accounts allows patrons to practice skills without fear of destroying their real files and digital content. It sets up an opportunity for safe, consequence-free play and exploration before transitioning the patron back to the real files to complete the intended task.

Step 4: Implementation

Goal. The penultimate step in the process is to actually accomplish the patron’s goal by teaching the patron how to solve the problem and/or fixing the problem for the patron.

Description. This step involves the tutor working with the patron to execute the plan produced in step 3. At this point, the tutor will decide whether the patron will “drive” with the tutor providing ongoing guidance, as needed, or if the tutor will demonstrate first and then ask the patron to repeat the process. As the tutor and patron work through the problem, the tutor may rely on analogies that relate aspects of the digital tasks to other technology or entities for which the patron may already have a mental model.

In select cases, the tutor may determine that the best solution is to complete the task for the patron. Typically this only occurs when the tutor is confident that it is a one-time fix that the patron will not need to repeat on her own, for example, when changing an arcane system setting.

During this implementation process, it is possible that the tutor will encounter a challenge executing the plan or identify an unexpected gap in the patron’s knowledge. In these situations the tutor may return to Step 2 (Triage).

Barriers. Implementation may be a period of information overload, where patrons can get lost or sidetracked. Patrons may even react with anger. As the solution plays out, motor skills issues may manifest themselves, and patrons who shun the tutor’s exhortations to explore and play may fail to make progress. There is a risk of patrons losing the logical thread of the workflow but failing to admit it; in such cases, the actions do accomplish the goal, but in a way that remains “magical” from their perspective.

Conflicts may emerge at this step if there is a misperception, either on the part of the tutor or of the patron, regarding the patron’s level of understanding. A patron may become stubbornly insistent that he understands, not willing to admit that he lacks information. Notes from previous sessions can be a hindrance in this regard: written notes from earlier sessions may be out of date but still convey a false sense of authority. Extensive, repetitious, or disorganized notes can also be a distraction from attention to the tutor and the device and problem at hand.

Strategies. Tutors should encourage patrons to “drive” (manipulate the digital device) as much as possible, with emphasis placed on play, practice, and exploration. In rare occasions, tutors may intervene and perform manipulations—for instance, in the case of a one-time setting change that requires delving into potentially confusing details. In such instances, the tutor must explicitly state why she is intervening. In all cases, either driving or not, the tutor “thinks aloud,” modeling her thought process to encourage independence.

Many patrons have knowledge of older technologies or non-technological systems that tutors can leverage to help patrons better understand digital technology. Many tutors have developed sets of analogies that they frequently use:

  • An arts and crafts analogy is helpful for digital cut-and-paste operations.

  • Experience with Internet Explorer can inform use of File Explorer.

  • To explain security-related topics, pop-ups can be likened to someone calling on the phone; if the communication is initiated on the other side, the recipient should demand verification. Likewise, spam can be compared to junk mail.

  • To discuss the cues that should be used when assessing the trustworthiness of websites, compare well-established websites to frequently visited Main Street businesses.

  • Compare URLs to physical stores or mailing addresses.

  • Motivate regular computer updates by invoking recommended regular servicing for cars.

  • Compare folders to kitchen drawers and the files to the items you place in them.

  • Explain context sensitivity through candles: “If the candle is in the kitchen the power’s probably out. If the candle’s in the bedroom, it’s probably romance.”

The patron must be protected from embarrassment and fear. Belittling the patron, and making him feel ignorant or stupid, even inadvertently, is strictly to be avoided. To build up patrons’ self confidence, tutors can get them to articulate what they already know, and emphasize that the fault lies with the technology, not the patron. Modeling assertiveness is important: tutors should encourage patrons to take action boldly once they have thought it through sufficiently. Tutors must be aware of anxiety cues and confusion cues, such as hesitation, withdrawal, slower responses, and lack of questions. For patrons experiencing stress, tutors must talk quietly and slowly, and work deliberately.

The sandbox approach, practicing on low-stakes or no-stakes data, is an effective strategy for reducing anxiety. When in the sandbox or manipulating real data, tutors may use back-off statements (e.g., “We don’t have to do it this way...”) and point out escape routes to earlier safe states. Best practices for staying safe online, such as choosing safe websites and avoiding and detecting viruses, can be woven into the discourse.

Step 5: Conclusion

Goal. The final step requires the tutor to recap what was accomplished and tie up loose ends.

The tutor concludes the session by reviewing with the patron what was done and answering any outstanding questions. If there is time left in the session, the tutor may allot time for the patron to continue practicing the task, providing feedback and assistance when necessary. If there is work remaining to be done, the tutor will explain that and indicate how the patron might proceed (for example, by coming back next week, watching videos online, or having hardware repaired). If the patron took notes during the session, the tutor can review them to ensure there aren’t errors and even provide a list of suggested websites or resources that the patron may find helpful. Even if the issue could not be addressed, the tutor tries to end the session on a positive note by reminding the patron what was accomplished and what he learned.

Barriers. The ultimate goal of our tutoring program is to prepare patrons to become independent technology users. Realistically, this is not something that can typically be accomplished in a single tutoring session. Limited time is a barrier. One purpose of this final step, therefore, is to provide sufficient information and resources that the patron can continue to practice and learn at home. Unfortunately, we find that solutions don’t always stick. Some patrons forget information between tutoring sessions (or return to seek reassurance that they are doing the task correctly). Others don’t practice outside of tutoring sessions.

Many of our patrons rely heavily on taking notes. During the interviews, tutors commented that while note-taking is a useful memory aid, over-reliance on the notes is a barrier to developing true digital literacy skills. Patrons learn to follow steps, but not to explore. Eventually, those who fall into this habit will return to the tutoring program when their step-by-step instructions no longer work for them, perhaps after an update to a website or an application.

Some patrons who are quite capable of completing tasks under the guidance of a tutor continue coming to the program. To some extent, this may be driven by a lack of self-efficacy and a desire to seek assurance from an expert. Others come back to the program when they have something else they want to learn or have additional issues, such as a new device to set up. For other patrons, their return may simply be driven by a desire for the social component of the program.

Strategies. To help overcome these barriers, the session should end with a plan of action. Should the patron plan to attend the program in the future? What can the patron do to practice and maintain the new skills at home? For those patrons who took notes, the tutor may consider reviewing them to check for errors. It is important to avoid embarrassing the patron by calling out errors, but this gives the tutor an opportunity to reinforce the correct steps. The tutor should encourage the patron to practice without the use of the notes, just referring back to them when the patron gets stuck. Finally, the patron should be armed with appropriate resources. Some tutors have a list of trustworthy websites that they share with patrons during the session. Others focus on modeling how to search for answers on Google, YouTube, or other websites.

4 Future Work

Results of the cognitive task analysis have revealed numerous best practices that we believe should be incorporated into tutor training curricula. Our tutoring process is emergent, and effective techniques continue to arise as our tutors gain experience and new tutors appear with new approaches. The results of the cognitive task analysis described here, however, provide us with an initial framework of replicable practices within which we can continue to refine our process.

It is important to note, however, that learning does not begin and end at our tutoring sessions; patrons must reinforce what they have learned through independent practice. Despite our tutors’ best attempts to encourage exploration, practice, and play within our tutoring sessions, encouraging patrons to engage in these behaviors independently is a remaining challenge. Patrons may not feel confident about setting up their own sandboxes at home or exploring without the watchful eye of their trusted tutor. Clearly, new techniques are required to reinforce these behaviors as patrons engage in digital tasks. To that end, we plan to focus on web navigation behavior and develop learning support through web components and browser add-ons that mimic tutor best practices—not as a substitute for face-to-face tutoring, but as a complement.