Master's Dissertation First Full Draft

5.5. Research question 3

(3) What differences exist, if any, between the results of research questions 1 and 2 and what is the nature of these differences?

(a) Between-subjects (one-week delay)

For factual questions, participants in the reading-only condition performed better when using paper compared to when tablets were used, with a statistical significance of p = 0.017 and a large effect size of 0.69. This result was not repeated in the note-taking condition. Instead participants using paper performed marginally worse than those using tablets. This difference is, however, not statistically significant (p = 0.45) and the effect size is effectively non-existent (-0.07).

For conceptual questions, participants in the reading-only question performed approximately equally regardless of the device used, with a highly non-significant result (p = 0.658) and an effectively non-existent effect size (0.04). Participants in the note-taking condition who used paper performed marginally better when compared to participants who used tablets, although this result is also not significant with p = 0.68 and a near non-existent effect size (0.03).

(b) Within-subjects

For factual questions, participants in the reading-only condition performed substantially better when using paper compared to when they used tablets, with a significant result (p = 0.01) and a moderate-to-large effect size of 0.65 being observed. Participants in the note-taking condition, however, performed slightly worse (for factual questions) when using paper compared to when using tablets, although the non-significant result and very small effect size suggest that performances were in fact approximately equal (p = .443, E.S. = -.158).

For conceptual questions, participants in the reading only condition performed better when using paper compared to when they used tablets. However, these differences are non-significant (p = 0.054) and a small-to-moderate effect size (0.33) was observed. It is possible that a real, but weak effect does exist, but these results do not provide adequate evidence to support such an assertion. Further research is required. Participants in the note-taking condition also performed better (for conceptual questions) when using paper compared to tablets, with these differences returning a significant result (p = 0.026) and moderate effect size of 0.48 suggesting real, but moderate effects.

5.6. Research question 4

(4) Are there differences in learning (as measured by test performance) if the delay between reading task and test (research question 2) is reduced from approximately one week to several days?

The relevant parametric assumptions were met and independent-samples t-tests were conducted to compare the obtained means for participants in the between-subjects (one-week delay) and between-subjects (two-to-three-day delay) samples. A total of four tests were conducted, comparing means for questions which tested factual recall (both for paper and for tablet conditions) as well as for questions which tested conceptual understanding (both for paper and for tablet conditions). The results for the paper condition are detailed in Table 13 below.

Table 13. Between-subjects (one-week vs two-to-three-day delay) – Paper condition

Question Group	Delay time	N	Mean	SD	95% CI of Difference		t	df	Sig. (2-tailed)	Effect Size (Cohen’s d)
					Lower	Upper
Factual	2-to-3	16	.10	.57	-.29	.54	.598	73	.552	.18
	One-week	59	-.02	.77
Conceptual	2-to-3	16	-.27	.73	-.66	.12	-1.37	73	.176	.37
	One-week	59	-.01	.69

No statistically significant differences were detected between participants in the paper condition for factual or conceptual questions. The results for the tablet condition are detailed in Table 14 below.

Table 14. Between-subjects (one-week vs two-to-three-day delay) – Tablet condition

Question Group	Delay time	N	Mean	SD	95% CI of Difference		t	df	Sig. (2-tailed)	Effect Size (Cohen’s d)
					Lower	Upper
Factual	2-to-3	9	.64	.46	.10	.25	2.4	33	.02	1.04
	One-week	26	.03	.69
Conceptual	2-to-3	9	.02	.44	-.34	.44	.24	33	.808	0.11
	One-week	26	-.028	.51

A statistically significant difference between participants in the tablet condition was detected for questions which tested factual recall, with p = .02 and a very large effect size of 1.04, implying that the difference which exists is greater than one standard deviation. No difference was detected for questions which tested conceptual understanding, with a highly non-significant result and a very weak effect size.