Forging new pathways of research and innovation in open and distance learning
RW_2016_Oldenburg_Proceedings
RW_2016_Oldenburg_Proceedings
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Investigat<strong>in</strong>g the Effectiveness <strong>of</strong> the ‘Onl<strong>in</strong>e Learner Pr<strong>of</strong>il<strong>in</strong>g Questionnaire’<br />
<strong>in</strong> Generat<strong>in</strong>g a Pr<strong>of</strong>ile <strong>of</strong> Learners Based on Learner Dispositions: A Pilot Study<br />
Pengiran Shaiffadzillah Pengiran Omarali<br />
sampl<strong>in</strong>g strategy employed was convenient sampl<strong>in</strong>g whereby a l<strong>in</strong>k to the survey was<br />
distributed to Brunei student societies resid<strong>in</strong>g <strong>in</strong> various countries as well as to the three<br />
higher <strong>in</strong>stitutions <strong>in</strong> Brunei viz. Universiti Brunei Darussalam, Institut Teknologi Brunei <strong>and</strong><br />
Politeknik Brunei; thus target<strong>in</strong>g undergraduate <strong>and</strong> postgraduate students who share similar<br />
characteristics with the target population <strong>of</strong> the ma<strong>in</strong> study. The pilot considered necessary<br />
measures to ensure that the data it generates was protected from threats <strong>and</strong> thus be valid <strong>and</strong><br />
reliable. For example, pilot<strong>in</strong>g a questionnaire with a larger list <strong>of</strong> items provided an<br />
opportunity to amend the <strong>in</strong>strument <strong>and</strong> remove items that exhibit coll<strong>in</strong>earity, as well as<br />
ensur<strong>in</strong>g <strong>in</strong>strument reliability through correlat<strong>in</strong>g items to check for <strong>in</strong>ternal consistency<br />
(Cohen et al., 2011; p.201).<br />
Analysis<br />
Accord<strong>in</strong>g to Cohen et al. (2011; p.402), a pilot study serves two functions <strong>and</strong> this study<br />
attempted to address both. The pilot study <strong>in</strong>volves (a) the analysis on the <strong>in</strong>strument design<br />
to <strong>in</strong>vestigate on response rates <strong>and</strong> types (Kgaile & Morrison, 2006), <strong>and</strong> (b) the analysis on<br />
the generated data to <strong>in</strong>vestigate the collective robustness (i.e. reliability, validity <strong>and</strong><br />
practicality) <strong>of</strong> the items <strong>in</strong> generat<strong>in</strong>g statistically significant data (Oppenheim, 1992).<br />
F<strong>in</strong>d<strong>in</strong>gs from both perspectives will ultimately establish the reliability <strong>of</strong> the questionnaire to<br />
be used <strong>in</strong> the ma<strong>in</strong> study.<br />
The approximate population for this pilot given the sampl<strong>in</strong>g strategy is roughly N = 4,000,<br />
with n = 149 respond<strong>in</strong>g to the questionnaire (3.8%). However, what is more important is that<br />
the response managed to exceed the m<strong>in</strong>imum n = 100 requirement for factor analysis<br />
(MacCallum, Widaman, Zhang, & Hong, 1999) <strong>and</strong> lean<strong>in</strong>g towards the Rule <strong>of</strong> 150<br />
m<strong>in</strong>imum requirement (Hutcheson & S<strong>of</strong>roniou, 1999). In addition, there is a significant<br />
improvement <strong>in</strong> completion rate as this pilot had 15.4% partial response rate, compared to<br />
42.8% <strong>in</strong> OLP V.1. The respondent sample has a gender distribution <strong>of</strong> 76 females (60.3%)<br />
<strong>and</strong> 50 males (39.7%). The mean age is 21.8, with median = 21.0, <strong>and</strong> range = 19.0<br />
(m<strong>in</strong>imum = 17.0, maximum = 36.0). The academic background ranges from PND n = 3, ND<br />
n = 11, HND/AD n = 5, First Degree n = 86, Masters Degree n = 9 <strong>and</strong> PhD n = 2. On<br />
experience with formal onl<strong>in</strong>e learn<strong>in</strong>g, n = 64 have experienced onl<strong>in</strong>e learn<strong>in</strong>g while n = 62<br />
have not. Accord<strong>in</strong>g to Morrison, “an unrepresentative, skewed sample, one that is too small,<br />
can easily distort the data, <strong>and</strong> <strong>in</strong>deed, <strong>in</strong> the case <strong>of</strong> very small samples, prohibit statistical<br />
analysis” (1993; cited <strong>in</strong> Cohen et al, 2011; p.209). In all three demographic pr<strong>of</strong>il<strong>in</strong>g, the data<br />
<strong>in</strong>dicated acceptable ratio <strong>and</strong> spread reflect<strong>in</strong>g the general population, albeit with the clear<br />
dist<strong>in</strong>ction that experience <strong>in</strong> onl<strong>in</strong>e learn<strong>in</strong>g is attributed to either a student study<strong>in</strong>g <strong>in</strong><br />
Brunei (hence no onl<strong>in</strong>e learn<strong>in</strong>g) or abroad (hence possible onl<strong>in</strong>e learn<strong>in</strong>g). Overall, the<br />
sample was adequate for statistical analysis.<br />
The OLP V.2 data set compris<strong>in</strong>g <strong>of</strong> 149 responses were imported to SPSS. The extent <strong>of</strong><br />
completion among the 23 <strong>in</strong>complete responses widely varied <strong>and</strong> therefore these cases were<br />
omitted from the actual data set <strong>in</strong>stead <strong>of</strong> perform<strong>in</strong>g multiple <strong>and</strong>/or fractional imputation<br />
102 Reach<strong>in</strong>g from the roots – 9 th EDEN Research Workshop Proceed<strong>in</strong>gs, 2016, Oldenburg<br />
ISBN 978-615-5511-12-7