Abstract:

Recent studies have shown that answering interspersed questions regarding a previously

examined word list enabled greater recall of new information than individuals that did no

study or restudied initial information (Chan, Manley, Davis & Szpunar, 2018). Within the

study, the effectiveness of different strategies to potentiate new learning were analysed and

compared. The results revealed that retrieval-based strategies most effectively potentiate new

learning relative to restudy. Retrieval proved to more effectively potentiate new learning

relative to those participants that performed list discrimination. The results make profound

contributions to the understanding of TPNL and enable the development of new educative

practices that facilitate optimal new learning through retrieval-testing.

Test-potentiated new learning or TPNL is defined as a phenomenon through which

testing on previously studied information facilitates new learning (Chan et al., 2018). The

existence of this positive influence of retrieval practice on subsequent encoding demonstrates

the phenomenon of TPNL. Encoding refers to the initial learning of information. This paper

will investigate the effectiveness of differing encoding conditions to potentiate more effective

encoding techniques as part of a strategy change mechanism for TPNL.

Retrieval has been shown to play an important role in improving subsequent recall by

reinforcing the episodic context of a previous learning experience (Whiffen & Karpicke,

2017). Retrieval-based testing has also demonstrated a TPNL effect evident in higher recall

than non-tested groups on a new identically studied second list (Chan et al., 2018

Experiment1-4). Whilst not based on TPNL, Whiffen & Karpicke (2017) compared the

recall performance of subjects that restudied or performed list discriminations, and those that
made pleasantness ratings (semantic) and categorical judgements. By analysing final free

recall, temporal and

semantic clustering, Whiffen deduced that temporal list discrimination judgments enhanced

temporal and semantic organisation and recall relative to restudy. Meanwhile, Chan et al.

(2018) implemented a 25-minute lag between original learning and new learning, to test

whether a switch in context from encoding to retrieval was responsible for TPNL. Despite a

lag, testing still out-performed no-testing and restudy, supporting a strategy change as the

mechanism for TPNL (Chan et al., 2018). However, frequent switching between encoding

and retrieval has been shown to impair new learning, as increased frequency of task

switching negatively recall (Davis, Chan & Wilford, 2017).

The current study design involved sets of interspersed testing on original learning prior

to an identical new learning test to examine the effect of new testing conditions such as list

discrimination, retrieval, category judgments and restudy on TPNL. Whiffen focused on

testing the theory that retrieval-based testing improves retention relative to restudy and

semantic judgments by making individuals rethink the relevant episodic context of learning.

Extending the found benefits of List Discrimination for temporal and semantic encoding for

retention (Whiffen 2015), the current paper seeks to compare their effectiveness within the

context of TPNL. Further understanding of the mechanisms of TPNL has the potential to

renew educational scenarios to incorporate specific learning strategies which aid new

learning.

Chan et.al (2015, Experiment 1-4) has shown that interspersed retrieval-testing is

significantly more effective in potentiating new learning relative to restudy and no study.

Thereby, in this paper, the retrieval testing condition will have significantly greater mean

recall rate than the restudy condition. Since, List discrimination judgements have been proven

as an effective form of retrieval-testing that enhances temporal organisation, it is reasonable

to hypothesise that they will allow for significantly greater recall compared to restudy.

Furthermore, the retrieval condition will have a significantly greater mean recall rate than list

discrimination, as list discrimination involves a lesser degree of retrieval and greater degree

of restudy than the retrieval condition (Whiffen & Karpicke, 2017), which is fundamental for

TPNL (Chan et al., 2018). Category judgments do not provide the necesary retrieval-based

testing scenario that potentiates new learning (Chan et.al.,2018 Experiment 1-4) and may

convolute retrieval with encoding of new categories that has been shown to impair new

learning (Davis et al., 2017). Consequently, the categorical judgment condition will not

exhibit a significantly higher mean recall rate relative to restudy.

Results:

M denotes the Mean Recall Rate/Proportion of Words recalled from List 3. Participants

in the retrieval condition demonstrated a significantly higher M than those in the restudy

condition, p=0.00001 (figure 1.1). Participants in the List Discrimination condition showed a

significantly higher M, than those participants in the restudy condition p=0.022 (figure 1.1).

Participants in the Category Judgements condition did not have a significantly higher M than

the restudy condition, p=0.44 (figure 1.2). The retrieval condition demonstrated a

significantly higher M than the List Discrimination condition, p=0.031 (figure 1.1)
Discussion:

The results of the experiment have provided evidence consistent with the hypothesis

that the retrieval condition would most effectively potentiate new learning. Additionally, list

discrimination effectively potentiated new learning relative to restudy (figure 1.1). The

results confirmed that no significant link was identified between the category judgment

condition and the effective potentiation of new learning relative to restudy (figure 1.2).

The effects of a lag interval on TPNL have highlighted that interspersed retrieval-based

testing effectively potentiates new learning based around a strategy change phenomenon

(Chan et.al 2018, Experiment 3-4). In this current paper, the effectiveness of corresponding

retrieval-based conditions such as List Discrimination Judgments and Retrieval to potentiate

new learning can be explained by their ability to provoke a strategy change. Consequently,

these testing conditions prompt new and superior encoding techniques that are evident in

their higher mean recall rate (M) from List 3 relative to restudy (figure 1.1). It can be

concluded that the category judgment does not effectively provoke a strategy change to

facilitate TPNL relative to restudy, as it doesn’t meet the conditions of retrieval outlined by

Chan et al. (2018) and Whiffen & Karpicke (2017) and mixes encoding with retrieval which

impairs new learning (Davis et.al., 2017).

The experiment has shown that the integration of interspersed retrieval-based testing

such as List Discrimination Judgements and Free Recall has a relevant place within the field

of education, as a result of its significant contribution to TPNL. Chan’s experiment was

designed to understand the mechanism that underlies TPNL and limited its testing to three

conditions. The current study has expanded on Chan’s methodology to incorporate further

conditions in a previously untested context of TPNL.

The conclusion that the retrieval condition was the most effective method supports the

findings of Chan et.al (2018) and warrants the exploration into implementing similar retrieval

conditions into different levels of education to benefit new learning. However, the current

experiment was limited to a measurement of average words recalled from List 3 and does not

consider other measurements such as clustering and inference used in the Whiffen &

Karpicke (2017) study. These additional measurements would facilitate the understanding of

the exact semantic/episodic learning strategies that aid TPNL. In order for TPNL to be

employed as an effective learning strategy future experimentation should analyse these

additional measurements. Feedback should also be gathered from participants as a positive

testing experience may be linked to greater inclinations to engage in a particular learning

strategy. By addressing these shortcomings within further experimentation and practical

integration of retrieval-testing conditions in education, the benefits and understanding of

TPNL will only grow.

 References:

Chan, J. C., Manley, K. D., Davis, S. D., & Szpunar, K. K. (2018). Testing potentiates new learning

across a retention interval and a lag: A strategy change perspective. Journal of Memory and

Language, 102, 83-96. 

Davis, S. D., Chan, J. C. K., & Wilford, M. M. (2017). The dark side of interpolated testing:

Frequent switching between retrieval and encoding impairs new learning. Journal of Applied

Research in Memory and Cognition, 6, 434–441.

Whiffen, J. W., & Karpicke, J. D. (2017). The role of episodic context in retrieval practice

effects. Journal of Experimental Psychology: Learning, Memory, and Cognition, 43(7), 1036.