2 Situational Episodic Context Theory 2.1 Extending the Episodic Context Account Some scientists studying the testing effect misinterpret Karpicke et al.’s (2014) ECA, assuming that the theory, as it was initially proposed, involves a reinstating of the entire episodic learning context, including visuospatial context (CITATIONS). In actuality, when Karpicke, Lehman, and Aue proposed their theory, they explicitly said that it was the temporal context that was reinstated. A personal communication with Karpicke confirms this (CITATION). Their experiments show that the temporal context is reinstated, at least for word pairs, under the condition of cued and free recall, and that this drives the RPE (CITATION). Some, such as Hong et al. (2019 ), have tested individuals for word lists while changing surrounding visual contextual details (e.g., font color), and found no evidence for a differential testing effect based on those superficial contextual details. While it is interesting to know, this is not what Karpicke and colleagues had in mind when they proposed their theory. I will now attempt to explain the reasoning behind their theory.
When people learn, they have a distinct learning experience that exists as an episode in time. Then, when individuals need to recall information they studied previously, they mentally reconstruct the temporal context of when the item was first studied. This updated temporal context is likely what Karpicke et al. were referring to. According to them, the temporal contextual details (i.e., time-line of events) direct a search process in the association cortex, acting as filters to narrow the search set to relevant information obtained during a short time period (CITATION). Though I will not attempt to test it in this paper, I hypothesize that if the details of the temporal context are highly relevant, they will be more likely to be reinstated and thus strengthen the related memory traces, leading to faster, more successful recall. I further propose that the brain compares the two (or more) temporal contexts, identifying the similarities and differences. By doing this, the brain filters through the unrelated contextual details to delineate the common features of an idea or stimulus.
The central idea is how past experiences relate to the present experience. Whenever a person makes a decision, they recall, consciously or not, prior information, experiences, and emotions related to the decision. These memories guide their subsequent decisions to those that are best suited to meet their goals given the specific circumstances. In a related fashion, episodic memories of learning semantic information guides one’s mental search to memories most related to the question being asked on a test. It is therefore advantageous to the individual to bring to mind those contextual details that will help them recall the desired information.
Karpicke et al.’s theory is consistent with the idea that free recall tends to show a much stronger RPE than recognition or even cued recall and also with Carpenter and DeLosh’s (2006 ) finding that impoverished cues provoke stronger RPEs, because with free recall, there are minimal cues present in the immediate environment; so the person must mentally recall and reconstruct prior episodic contexts for experiences where the same or similar information was encountered.
Furthermore, Karpicke et al. hit on something important regarding the concept of cue overload. When studying the testing effect or memory in general, forgetting is often seen as undesirable. However, I contend that this idea is misguided at best and detrimental at worst. Forgetting is actually a vital component of the memory consolidation process. The important thing is what is forgotten. The brain uses relevant, distinct, and memorable (not just any) contextual details from past experiences to recall the necessary information. If too many contextual or semantic details are remembered, the search set is broadened instead of narrowed, making it more difficult to recall the precise information. If a few important details are remembered, narrowing the search set, when the mental representations of the concept are strengthened and made more accessible, the brain (likely) “marks” unnecessary or irrelevant details to be forgotten. That, in turn, would lead to a sharper and quicker recall of the target information because all of the mental "clutter," so to speak, is cleared away.
Returning back to Hong et al.’s (YEAR) objections to the ECA, they found that font color did not predict a RPE. However, font color is not very important for reading and is typically unrelated to the content, and most people can read words in any color that they can see, provided there is adequate light, contrast, and spacing. Changing the color of the font doesn’t change the meaning or even the implications of the text very much, if at all. Thus, it makes sense that font color would not be a very poignant contextual detail to aid in recall. It would only be important if different font colors implied different things for the contexts that were well-known to the participants.
Thus, while I believe that temporal contextual details are very important, other episodic contextual details can also aid in recall if they are sufficiently associated with and specific (distinct) enough to identify the target information. This idea is the crux of this paper. Through several experiments, I will attempt to show that the SEC theory is accurate not only because of temporal context reinstatement but also because of visuospatial context reinstatement as well. In short, I will try to show that the testing effect is strengthened by the overall episodic context.
It is important to note however, that not all episodic contextual details will strengthen the testing effect. I propose that only those contextual details which are situational will be useful for increasing recall. I agree that Karpicke et al. (2014) had the right idea, but I believe that their theory only partly explains the testing effect. For example, part of situational episodic context (SEC) includes temporal context; people might remember in what order they learned different information. But it is also possible that they remember a visuospatial component as well; people might remember where they were and what they were doing when they encountered a test question or read something from a textbook.
In this experiment, participants will either study or test, and they will either do so while riding the bus or in a classroom setting. If my theory of SEC is correct, then all of the testing groups should show a testing effect but more so for the testing group that rode the bus.