Scaffolding refers to any sort of learning assistance provided after initial instruction, including context-sensitive help. The term is borrowed from the construction industry, where a temporary structure is built and used to construct or modify another structure (van de Pol, et al, 2010). As such, scaffolding is temporary and removed when no longer necessary. Within education, the social learning theory of Vygotsky (1976) is generally credited with providing the theoretical basis for the practice, where he describes the zone of proximal (next) development:
Zone of proximal development: “The distance between the actual developmental level as determined by independent problem solving and the level of potential development as determined through problem solving under instructor guidance or in collaboration with more capable peers” (p. 86).
|1. Vigotsky's zone of proximal (next) development|
Characteristics of scaffolding
The literature is largely consistent in describing features of scaffolding (Azevedo & Jacobson, 2008; Lee & Kolodner, 2011;, Smit, et. al., 2012; and van de Pol, et. al., 2010).
- Contingency. Support is provided only when and to the extent it is needed, implying the necessity for accurate diagnosis on the part of the support provider, human or machine.
- Temporary. Support is reduced and removed as the learner gains competence and is able to answer questions and solve problems on his or her own.
- Transfer of responsibility. Responsibility for successful performance is gradually transferred from the support provider to the learner, both cognitively and metacognitively.
“If a student, for example, works on a series of tasks and the instructor adapts the support responsively to the understanding of the student, the instructor is teaching contingently. If the student gains understanding, the instructor can fade the support over time. While fading the support, the instructor can also transfer the responsibility to the student so that the learner will take more and more control over his/her learning.” (van de Pol, et al, 2010).
You may be thinking that there is only one purpose for scaffolding – helping learners master the content or performance. Of course, with humans it’s not that simple. We can consider scaffolding as “situational leadership” for learners. In his One Minute Manager, Ken Blanchard (1982) describes the situational leader as “providing for the employee what the employee is unable to provide himself.” In the case of learning, situational support can focus on cognitive, metacognitive, and affective issues. Kim & Hannafin (2011) and van de Pol, et. al. (2010) describe six purposes of scaffolding:
- Cognitive structuring. Providing explanatory and values structures (mental models, schema) that organize and justify.
- Reduction in the degrees of freedom. Taking over those parts of a task the student is not yet able to perform, or breaking it into smaller steps and thereby simplifying the task for the learner and reducing cognitive load.
- Direction maintenance. Keeping the learner and the learning focused on specific goals.
- Metacognitive coaching. Modeling cognitive strategies and helping learners identify and modify their learning and problem solving strategies.
- Recruitment. Getting the learner interested in a task and helping him adhere to the requirements of the task.
- Contingency management/frustration control. Facilitating learning using rewards as well as keeping learners motivated by minimizing frustration (identifying the source of frustration and removing or mitigating it).
A number of specific scaffolding tactics are provided in the literature. Here we divide the list into active interventions and on-demand resources available for student retrieval.
- Feedback and Feed Forward. Providing information regarding the student’s performance and making suggestions for improvement. This can be seen in multi-phase assignments where milestones are established for each component of a larger project, with the instructor and/or peers providing feedback to be incorporated into the final submission. Please refer to the Feedback article.
- Problem Fading. Presenting whole problems with the solution carried out until the final step, requiring the student to complete it (backward fading). Alternatively, the more difficult steps are provided and the learner completes the easier ones (whole task fading). Either way, the process is repeated with more and more of the problem unsolved, requiring the student to complete more and more (Clark & Mayer, 2011). See Fading.
- Hinting. Providing clues or suggestions to help the student move forward. The entire solution or detailed instructions are deliberately withheld.
- Instructing. Telling students what to do or describing how something must be done and why. This is another use for milestones.
- Explaining. Providing more detailed information and clarification.
- Modeling. “The process of offering behavior for imitation.” This can include demonstrating particular skills such as search strategies and software simulations.
- Questioning. Asking questions that serve to guide learner thinking.
On-demand (fixed) resources
We should note that on-demand resources do not properly fit our definition of scaffolding unless accompanied by a diagnostic agent that fades support and transfers responsibility as the learner is able to perform more and more independently. Many learners will naturally depend less on support as they gain competence, however, this cannot be assumed. The creation of online scaffolding is addressed in the Course building section.
- Progressive disclosure. A series of prepared explanations, moving incrementally from general to detailed or “shallower to deeper,” containing progressively more information, direction, and simplification (Bell & Zemke, 2007; Fernandez, 2003).
- Progressive questioning. Closely related to progressive disclosure, this approach uses questions and learner responses to guide learning. An initial question is posed and, based on the response, additional questions are presented that narrow or widen the scope, retrace previously asked questions, and provide hints or other information.
- Procedural Guides. How-to instructions, which can also be separated into progressively more detail (Kim & Hannafin, 2011).
- Rubrics. Assessment tools for qualitative rating of authentic or complex student work. It includes criteria for rating important dimensions of performance, as well as standards of attainment for those criteria. The rubric tells both instructor and student what is considered important and what to look for when assessing (Jonsson & Svingby, 2007). See Rubrics.
- Modeling can be made available through videos and animations.
- Faded problem sets, machine scored; can be provided via computer or online.
- Case-based scenarios (Lee & Kolodner, 2011) describe previous examples of design or repair problems that can serve as fodder for application to the current one. Previous cases can be used to provide suggestions to proceed in a new situation and how to get around common obstacles. They can also be used to guide evaluation, develop evidence to inform decision-making, develop predictions about the way solutions-in-progress will work, and multiple perspectives on the same issue.
- Hypertext and Hypermedia are made available through the use of hyperlinks within text-based resources. When clicked or tapped, the link opens an additional text or multimedia window providing the additional support. Cognitive load theory (Clark et al., 2006; Gerjets & Scheiter, 2010) suggests that embedded hyperlinks increase extraneous cognitive load for novice learners engaged in linear instruction, with each link presenting the reader with a choice to make – follow or ignore. In these cases, the links should be placed in their own subsections near the text. In any case, hyperlinks should open into new tabs or windows to avoid learners becoming “lost in hyperspace.”
- Templates guide the construction of documents, presentations, experiments, etc. by asking questions to be answered or directing the content to be entered within provided spaces, resulting in a completed or near-completed product. In many instances, templates operationalize scoring rubrics for learners.
- Taking over lower-order tasks such as performing calculations in order to free cognitive resources to concentrate on the learning task (reducing extraneous cognitive load).
Automated Scaffolding with Technology
There are many research efforts aimed at building automated scaffolding systems. The 2009 book Artificial Intelligence in Education (Dimitrova, et al., 2009) provides a wide-ranging look at these efforts, focusing on “building learning systems that care: from knowledge representation to affective modeling.” Far beyond the scope of this article, it serves as a reminder of the progress being made in artificial intelligence in general, and education in particular. Among the topics in the book:
- Intelligent tutoring systems
- Task-based vs. affect-based feedback
- Using natural language processing to interact with students
- Automated assessment of verbal reading
- Engaging collaborative learners with helper agents
Fernandez (2003, 2004, 2005) describes a system of cognitive scaffolding in a web-based learning environment using a combination of a learning management system assessment tool for its presentation, data storage, and feedback capabilities, and a mathematics symbolic manipulation package (e.g., Mathematica, Matlab), for its analysis capabilities. The system is designed to “advance students up a cognitive ladder” by determining the next appropriate question based on student answers and history (using item response theory), stopping when the student has either reached an impasse or achieved the learning objective. An example of automated scaffolding can be seen in the Web-based Inquiry Science Environment (WISE). Signup is required.
A 2004 study (Mercer, et. al.) found students in the metacognitive scaffolding condition made greater gains both on group and individual measures of non-verbal reasoning than students in the control condition.
Azevedo et al. (2005) demonstrated that students receiving human adaptive scaffolding received significantly higher test scores and made greater shifts in their mental models than those receiving no scaffolding.
van de Pol, et. al. (2010) reviewed an additional seven effectiveness studies and concluded that scaffolding is effective for cognitive and metacognitive activities, but there is insufficient evidence to confirm its use for affective (motivational) issues. Testing automated emotional support, Mostwo, et. al. (2002) found no significant difference in persistence between experimental and control groups.
Scaffolding is a very useful concept for online learning. We separate it from initial instruction, plan for it, and make it adaptive to individual learner needs. Artificial intelligence will play a significant role in automating scaffolding systems.
Up to Home