Master's Dissertation First Full Draft

Chapter 6

Discussion

6.1. Research question 1

Research Question 1: Are there differences in learning (measured by the ability to correctly answer questions from each of Butler’s (2010) categories of question, with a delay of approximately one week between reading task and test) when text is read (only) on paper compared to when this is done using a tablet PC?

For the purposes of comparing various device interfaces, the reading-only condition is arguably the most evenly-balanced or equal of the conditions used in these experiments, as it substantially reduces potential confounding variables such as device familiarity and technological proficiency. That statistically significant differences were shown to exist for both the between-subjects and the statistically more powerful within-subjects samples adds further support to the notion that, rather than being approximately equivalent activities, important differences in fact exist between reading from a tablet (i.e. computer screen) and reading from paper. The small sample size obtained for participants in the tablet condition of the between-subjects group is a potentially confounding variable. It must be noted that other potential confounding variables such as zoom level, head and neck position, and others are not necessarily eliminated by reading only and must be controlled for in other ways. However, it must also be noted that the vast majority of participants indicated strong (and well-established) general tablet device proficiency on the relevant items contained in the demographic questionnaire (Appendix C) Additionally, the majority of participants had been using tablet PCs in the classroom for approximately 3 years, decreasing the likelihood that details such as these acted as significant confounding variables.

The finding of both sample groups (with moderate-to-large effect sizes) that participants in the paper condition performed better on factual questions lends weight to the assertions of Wolf (2008), Wolf and Barzillai (2009) and Wolf et al. (2012) and others that electronic screen-based reading promotes more superficial reading of textual content, resulting in participants either recalling factual details less clearly when tested or possibly engaging less strongly with these factual details while actually reading the text.

However, this assertion struggles to explain the non-significant differences observed for questions teaching conceptual recall. The result for the statistically more powerful within-subjects condition was non-significant and its effect size was substantially smaller than the effect size observed for factual questions. One possible explanation for this is that the type of skimming or shallow(er) reading which is encouraged by the digital word (or at least some iterations of it) involves the reader aiming to ascertain the gist of the text and its content, without paying as much attention to many details, in the same way as some individuals often tend to remember the gist of a news story without remembering any of the names of the people, companies or things involved, or in the same way as individuals may tend to remember what a product does/how it works (in terms of key selling points) without remembering details such as its official name. This would allow tablet readers to gain a general idea of how concepts in described in the text work, even if their engagement with factual details is far more limited, enabling them to answer conceptual questions more accurately than was the case for factual questions.

The slightly higher performance for conceptual understanding for participants using paper in the within-subjects condition may be explained by existing evidence of the importance of visuo-spatial processing and spatial elements found in conceptual processing (cf. Chapter 2.3.5) – the advantage given by the more stable page layout and ontological connection between word and substrate allows for stronger visuo-spatial mental simulations, which are important for conceptual understanding. That participants were given 5 minutes to revise before completing their tests suggests that paper’s more stable visuospatial layout may make a difference for reading- and revision-based learning.

It is possible that this tendency of the digital word to encourage a greater degree of skimming or shallow(er) reading is in part due to its virtual nature – it is by definition insubstantial (in that it is unstable, impermanent, fluid) and therefore can subconsciously be treated less seriously (as a study tool). An everyday example of this, as Mangen (2013a) notes, is that legal documents are still largely printed out and often only seen as being official or ‘real’ when converted to paper-based format. The multifunctional nature of tablets may also play a role – a tablet can be used to play games, recreationally watch videos, browse the web or social media platforms, in addition to reading through and studying from electronic textbooks. A textbook, on the other hand, is generally only used to study from and therefore may encourage a more studious attitude in its user. However, more careful and controlled research experiments are required to investigate this matter more fully, examining issues such as reading speed, scrolling behaviour.

6.2. Research question 2

Research Question 2: Are there differences in learning (measured by the ability to correctly answer questions from each of Butler’s (2010) categories of question, with a delay of approximately one week between reading task and test) when text is read and notes are taken on paper compared to when text is read and notes are taken using a tablet PC?

In a reverse of the reading-only conditions, participants in the reading and note-taking condition showed no differences in performance for factual questions, suggesting that adding the action of note-taking may mitigate against the digital word’s tendency to encourage skimming and encourage learners to focus more carefully on facts and details presented in the text. This may be because making notes requires participants to extract information to write out as notes and factual details are generally easier to note than the action of summarising conceptual explanations, encouraging the learner to focus more carefully on factual details. The between-subjects condition displayed no difference in performance for conceptual questions, with an effectively zero effect size. The more statistically powerful within-subjects condition, however, displayed a statistically significant difference (higher score for paper), with a moderate effect size. Given the reasonably large sample size for the within-subjects group (N = 42), which is largely comparable to the N = 59 (paper) and N = 26 (tablet) sample sizes obtained for the between-subjects conditions, the within-subjects’ result does appear to suggest real but relatively weak effects.

The act of studying a text by taking notes has some important differences in comparison to reading-only – for example, the former requires the learner to both mentally and physically (i.e. motorically) reconstruct portions of the text as part of the learning process. The stronger performance of participants in the paper (i.e. handwritten notes) condition on conceptual questions is consistent with Longcamp et al. (2003; 2005; 2006) and James and colleagues’s (2006; 2009; 2010; 2012) findings on the neural importance of handwriting’s motoric elements and its relation to reading and Mueller and Oppenheimer’s (2014) general findings that handwritten text engages greater levels of processing than typing and leads to better recall of text and the notion that the digital word can be less spatially addressable (i.e. containing fewer opportunities for visuospatial, perceptuomotor encoding) than the analogue word, thereby removing an important and useful aspect of the external memory field.

The text-centric nature of typed note-taking may be a disadvantage when studying concepts, where techniques more easily accomplished by the analogue word such as drawing diagrams are helpful. Given existing evidence of the importance of visuo-spatial processing and spatial elements found in conceptual processing (cf. Chapter 2.3.5), the more stable page layout of the printed text in the paper condition may also have contributed to participants in this condition performing better for conceptual questions (by allowing for greater degrees of perceptuomotor encoding).

This evidence, although interesting and helpful, does not clearly address to what extent the ontological disconnectedness of text and substrate played in this learning process – it is too heavily masked by potential confounding variables. Similarly to the previous research question, more research is needed to more carefully untangle potential confounding variables which in this case include note-taking proficiency and experience and general study (method) preferences.