A focus on the preparation for future learning (PFL) requires the development of new types of assessment. Most assessments employ “sequestered problem solving” in which students work without access to resources for learning. Like members of a jury, students are shielded from outside sources that might contaminate their performance. Though sequestered problem solving may be a good measure of full expertise, it can be a blunt instrument for assessing whether someone is ready to learn. The observation echoes Vygotsky’s (1934/1987) arguments for evaluating a child’s zone of proximal development.

Like a gardener who in appraising species for yield would proceed incorrectly if he considered only the ripe fruit in the orchard and did not know how to evaluate the condition of the trees that had not yet produced mature fruit, the psychologist who is limited to ascertaining what has matured, leaving what is maturing aside, will never be able to obtain any kind of true and complete representation of the internal state of the whole development…” 

Dynamic assessments measure what a student can learn given resources or scaffolds. Without dynamic assessments, it is difficult to evaluate whether an instructional method has successfully prepared students for future learning. For example, we asked students to analyze simplified data sets from classic memory experiments. When students completed these activities, they did poorly on tests compared to students who wrote a summary of a chapter on the same memory experiments. However, when students in both conditions received a learning resource in the form of a follow-up lecture, the results reversed themselves. On a subsequent assessment, the data analysis students showed that they learned much more deeply than the summarizing students and made twice as many correct answers. These results are useful because they show that instructional techniques that look poor on initial assessments of sequestered problem solving may still turn out to be very useful when evaluated with dynamic assessments. Discovery activities, for example, often do not lead students to the target knowledge, and therefore the activities can appear inefficient. However, if done well, they may be very useful when coupled with subsequent opportunities to learn the target knowledge, even in the form of a dry lecture. It is important to develop PFL assessments, not only because preparation for future learning should be a significant goal of instruction, but also because educators may otherwise overlook the value of some forms of instruction that are especially valuable in the long-run.

The study of transfer is an excellent domain to help clarify the position we are advocating towards instruction and assessment. The position we are advocating is not to develop general skills or habits of mind. Rather, we presume the best preparation for future learning and transfer comes through the development of content-rich knowledge. The top panel of the figure summarizes the experimental design used in many instructionally relevant transfer studies. Students learn Topic X by one of two instructional methods. Afterwards, they receive a transfer problem that is relevant to topic X but that has a different surface form. Researchers then compare which of the conditions leads to better performance on the transfer problem to draw conclusions about effective instruction. This is not a bad model, because transfer problems are particularly good for comparing methods of instruction; they are often more sensitive to differences in understanding than direct tests of memory.

The PFL perspective suggests a different approach, shown in the middle panel of Figure. As before, students study Topic X in one of two ways. The difference is that students from both conditions then receive equal opportunities to learn from a new resource. For example, they might read the same chapter on Topic X after completing their instructional treatments. After the opportunity to learn, students then receive a transfer problem that depends on material included in the learning resource. Researchers can compare performance on the transfer test to determine which method of instruction better prepared students to benefit from the learning opportunity. We label this approach a “double transfer” design, because students need to transfer what they learned from the instructional method to learn from the resource, and they need to transfer what they learned from the resource to solve the target problem. This seems like a more complete model of transfer, because it considers both the transfer that helps one to learn and the transfer where one applies that learning. For example, students bring knowledge to school that helps them learn, and what they learn helps them solve new problems. The bottom panel in the Figure shows a "double transfer" experimental design that we have employed. This design helps to show that two methods of instruction may look the same until we assess how well they prepare students to learn. In this design, students learn Topic X by one of two methods. Afterwards, half of the students in each condition have a learning opportunity and then receive the transfer problem per the double transfer design in the middle panel. The remaining students from each condition go directly to the transfer problem per the standard transfer design shown in the top panel. The outcome that we find, given appropriate instruction, is that the students who receive the different forms of instruction perform about the same on the transfer task when they do not have access to the learning opportunity. Thus, by the standard transfer approach, the two instructional methods look the same. However, when the students have access to the learning opportunity, they look quite different on the transfer problem. The students from one method of instruction do much better, whereas students from the other method show no benefit from the learning opportunity. The double-transfer design reveals that one method of instruction better prepared students to learn. In combination, the experimental design will help to show that some methods of instruction better prepare students for future learning, while the design also shows the need for assessments that can detect this preparedness.