A Sandwich-model experiment with personal response systems on epigenetics: insights into learning gain, student engagement and satisfaction

Abstract

Current trends in Higher Education Pedagogies include an ongoing discussion about active learning strategies. Technology-based interventions such as personal response systems (PRS) have gained momentum, especially since the advent of cloud-/web-based solutions. One model that supports the transition from traditional lecturing towards active learning by maintaining a balance between instruction and self-learning is the 'Sandwich Model'. In the present study, we investigated the impact of the Sandwich Model combined with PRS in student learning, engagement and satisfaction by a randomised trial in a large undergraduate biomedical/medical sciences class. A teaching session on epigenetics was delivered either as a traditional lecture (C-group) or as a PRS-including Sandwich-based session (S-group). The major finding of our experiment was the significantly enhanced performance of the S-group over the control, suggesting that the Sandwich Model improves learning gain. We also provide strong evidence that the Sandwich Model enhances student engagement and satisfaction. However, the effect of the Sandwich Model in learning gain and student attitudes was not dependent on PRS incorporation per se and students seemed to favour non-PRS activities over PRS, as evidenced by their feedback. Although further experimental research is needed in order to conclusively compare and contrast PRS and non-PRS activities regarding learning gain, we propose the usage of the Sandwich Model with a variety of in-class learning activities, both PRS and non-PRS-based. Altogether, our work shows that the Sandwich Model is a powerful pedagogical approach that exerts a positive impact on student perceptions for learning and satisfaction and that can support the teaching of challenging biomedical concepts, such as epigenetics.

Keywords: active learning; clickers; higher education didactics; personal response systems; sandwich principle; teaching epigenetics.

Fig. 1

Study design and the architecture of the Sandwich model. (A) Summary of study design. In total, 99 students participated in the study. Of them, 45 were taught by the traditional way of lecturing (C‐group) and 54 by the sandwich model including PRS activities (S‐group). In‐class learning was evaluated at the end of each teaching session with an identical for both group questionnaire (QUE1). (Appendix S2). A week later, students were asked to provide feedback on their experience and learning by a second questionnaire (QUE2, Appendix S3). (B) Sandwich model with PRS. Schematic representation of the sandwich model. The beginning and finishing phases are depicted as the burger bun. PRS activities are presented by a slice of tomato and the non‐PRS activities by sliced of onion. The burger stake represents the main teaching units delivered by the lecturer (collective learning). The break in the middle of the lecture is depicted as a piece of lettuce. (C) Diagram of the sandwich architecture of our study experiment. PRS activities are represented in red (tomatoes) and non‐PRS in light magenta (onions). Details on learning objectives (brown, meat) can be found at the presentation source file (Appendix S1). The activities are placed in chronological order.

Fig. 2

Comparison of performance at MCQ test (QUE1) between the control group (C‐group, traditional lecture, n = 45) and the hypothesis testing group (S‐group, sandwich model, n = 54) (unpaired t‐test: P < 0.001, two‐tailed, t = 5.861, df = 97).

Fig. 3

The Sandwich Model and attention span: (A) Schematic representation of delivery design with learning objectives (LO) and active learning activities (ACT) for the S‐group and C‐group, respectively. Details on learning objectives and activities can be seen at Appendix S1 and Table S3. (B) Time slot of content delivery and learning outcome for the C‐group and S‐group. Corresponding questions (Q1…Q12) from the QUE1 (Table 2, Table S3) are associated with the time point of delivery within didactic hours. Y‐axis shows the percentage of correct answers per each question. X‐axis shows the time slot (minutes from lecture start) rounding the minute (i.e. slide related to Q1 was presented at 7’35’’ taught). It is important to underscore that S‐group was offered 12 interactive activities of a total duration ca. 15 min. In reality, the C‐group had a rather longer break and a slightly earlier lecture closure (ca. 8 min).

Fig. 4

Comparative results for PRS and MCQ for S‐group. Six questions were presented at the S‐group (n = 54) as multiple‐choice questions with Turning Point in‐class at various time points across the 2 h of lecture, and they were part of the MCQ test (QUE1) at the end of the teaching session (see also Table S4). (A) The correct responses as percentage per question are depicted for PRS and non‐PRS (MCQ) setting. A clear increase on 5/6 questions was observed. (B) The score of correct responses was significantly improved at the MCQ in comparison with PRS setting (mean with SD is depicted) (**: paired t‐test; P‐value = 0.0023, two‐tailed, t = 5.702, df = 5, SD = 18.15, SEM = 7.408, 95% CI = 23.20–61.28, R ² = 0.867).

Fig. 5

(A). Correct responses per group divided by category of questions (PRS‐employed or non‐PRS activities). For C‐control group, no activities were involved but the scores on the same MCQ questions (highlighted in light orange for non‐PRS and green for PRS) were employed. (B) Adjusted predictions of PRS question categories by group with 95% CIs.

Fig. 6

Students’ Feedback and Perceptions for the Sandwich Model and PRS: Results from QUE2 depicted by Likert R package. (A). A comparative view between C‐group and S‐group is provided per question for the F1–F9 questions which both groups answered. (P‐values, Mann–Whitney U‐test) (B). Results for the Sandwich‐related questions (F10–F15) which were asked only to S‐group cohort (see also Appendix S3, QUE2 and Tables S5,S6).

Fig. 7

Analysis of students’ free‐text feedback regarding student engagement and satisfaction: Likert charts arising from the three free‐text feedback questions (QUE2, F16–F18) related to student satisfaction (left up) and engagement (right up) for the S‐group and the C‐group. Evaluation of comments was performed by three independent observers who scored comments in a five‐point scale (from very negative to very positive). The median score of the three evaluators was used to classify each comment to one of the five points (Table S7). P‐values were calculated by Mann–Whitney U‐test. Sample comments for each of the coding category are provided at the lower panels for satisfaction and engagement, respectively. A full list of all comments received per group of study and their assigned coding category can be found at Table S7.