2013. Vol.4, No.7A2, 117-125
Published Online July 2013 in SciRes (http://www.scirp.org/journal/ce) http://dx.doi.org/10.4236/ce.2013.47A2014
Copyright © 2013 SciRes. 117
An Inquiry into the Effectiveness of Student Generated MCQs as a
Method of Assessment to Improve Teaching and Learning
Damien Hutchinson, Jason Wells
School of Information Technology, Deakin University, Melbourne, Australia
Received June 3rd, 2013; revised July 3rd, 2013; accepted July 10th, 2013
Copyright © 2013 Damien Hutchinson, Jason Wells. This is an open access article distributed under the Creative
Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium,
provided the original work is properly cited.
In anticipation of helping students mature from passive to more active learners while engaging with the
issues and concepts surrounding computer security, a student generated Multiple Choice Question (MCQ)
learning strategy was designed and deployed as a replacement for an assessment task that was previously
based on students providing solutions to a series of short-answer questions. To determine whether there
was any educational value in students generating their own MCQs students were required to design
MCQs. Prior to undertaking this assessment activity each participant completed a pre-test which consisted
of 45 MCQs based on the topics of the assessment. Following the assessment activity the participants
completed a post-test which consisted of the same MCQs as the pre-test. The pre and post test results as
well as the post test and assessment activity results were tested for statistical significance. The results in-
dicated that having students generate their own MCQs as a method of assessment did not have a negative
effect on the learning experience. By providing a framework to the students based on the literature to
support their engagement with the learning material, we believe the creation of well-structured MCQs re-
sulted in a more advanced understanding of the relationships between the concepts of the learning mate-
rial as compared with plainly answering a series of short-answer questions from a textbook. Further study
is required to determine to what degree this learning strategy encouraged a deeper approach to learning.
Keywords: Assessment; Multiple-Choice Questions; Student Generated Content; Computer Security;
Assessment refers to the processes used to evaluate student
achievement of the learning objectives of a unit or course (Phye,
1997: p. 525). “Of all the activities associated with teaching and
learning, assessment have the potential to have the most influ-
ence in directing students’ energies and in determining their ap-
proach to learning in a unit of study” (Elton & Johnston, 2002).
In order to assess understanding and measure learning, students
are typically given an assignment that may vary from problem-
solving exercises to project work and essay writing. One of the
critical challenges the teacher faces as part of the assessment
process is how to encourage a deeper understanding of the key
areas of the curriculum (Palmer & Devitt, 2006). Teachers pre-
fer students to undertake their subject with the intention of en-
gaging with it and understanding it (a deep approach) rather
than with the intention of merely reproducing material without
having a good understanding of it (a surface approach) (Elton &
During our teaching experience as an instructor and lecturer
over the past decade, it has become apparent that few students
engage with the learning material on a progressive basis. To
develop quality assessment tasks that prepare students for pro-
fessional practice it is imperative to measure the effectiveness
in terms of student learning that these tasks impart. The student
cohort at the centre of this study was enrolled in a unit called
Introduction to Computer Security (ITCS). This is a first year
unit offered on and off campus at Deakin University in Victoria,
Australia. ITCS is one of eight units offered as part of the IT
security major leading to a Bachelor of IT (IT Security) by the
School of Information Technology, within the Faculty of Sci-
ence, Engineering and Built Environment. The unit may also be
taken as an elective. In ITCS students are required to learn the
principles of computer security for the protection of company
assets and information systems of an organization.
The current assessment tasks for ITCS require the students to
read the textbook and related material then provide solutions to
short answer questions. These solutions are subsequently marked
and a grade is returned to each student with feedback and a
suggested solution. The solutions provided by the majority of
students were largely reproduced from the textbook. This made
it extremely difficult to determine if the students had actually
learned the material or just taken a surface approach with the
intention to complete the task requirements by reproducing ma-
terial (Ramsden, 1988: p. 19).
Applying the advice of Brown, Rust & Gibbs (1994); “if you
want to change student learning then change the method of
assessment”, having students create their own MCQs was imple-
mented as the new method of assessment. The basis for this de-
cision was that student-questioning as a learning strategy plays
a crucial role in students’ active processing of given materials
(Wong, 1985); for students to be active learners and independ-
D. HUTCHINSON, J. WELLS
ent thinkers, they must generate questions that help shape, focus,
and guide their cognition during the learning process” (Singer,
1978). Furthermore, several authors consider the student gener-
ated question as an enabler for participants’ cognitive elabora-
tion and as an effective alternative for achieving meaningful
learning by reinforcing higher-order thinking skills (Wong, 1985;
English, 1998; King, 1995; Leung & Wu, 1999; Silver, 1994).
This paper delivers the findings of a study into the use of an
alternative assessment approach involving students creating
MCQs. Previous studies where students created MCQ’s were
found to be successful (Yu & Liu, 2004b; Palmer & Devitt,
2006). Studies indicate that students could potentially be moti-
vated to study a subject in greater depth if provided with the
opportunity to write their own problem-solving exercises (Sir-
car & Tandon, 1999).
By having students create MCQs it is anticipated that stu-
dents would be involved in a deep approach to learning by
having to both engage with and understand the ITCS principles
rather than just reproduce them. It is expected that this ap-
proach will add value to teaching by encouraging the student to
develop a deeper approach to learning by providing a new and
different assessment method that motivates greater interaction
with the unit material (Toohey, 1999: p. 167). It is also ex-
pected that this approach will provide formative feedback
quickly to the students as well as provide quantitative data
(Censeo, 2007) and insights into those areas on the unit mate-
rial that require more emphasis or improvement. A repository
of MCQ’s will also be created by the students that can be pro-
vided back to them as an additional learning resource. The aim
of the study was to measure the effect of students constructing
their own MCQs on learning. The hypothesis to be tested was
that when given a topic to research and write on, student under-
standing would be enhanced if they were asked to construct
MCQs based on the material under study.
This section provides an overview of the literature surround-
ing the use of MCQs for assessment. The premise for using the
MCQ as a method of assessment is outlined together with the
associated advantages and disadvantages in relation to teaching
ITCS. Second details of how the MCQ can be used to achieve
active learning is described within the context of two other
studies that used the construction of student generated MCQs as
the method for learning.
The 21st century has brought new challenges for teaching
and learning in higher education. In particular assessment needs
to be innovative in practice, responsive to individual needs and
relevant to real life. Both Parry (2004) and Parker (2004) em-
phasize the need for a learning paradigm that uses more forma-
tive, technology-mediated assessment. Moving in this direction,
Table 1 provides an analysis of the relationship between key
elements of assessment and the use of MCQs for assessment.
With the intention of using student generated MCQs and soft-
ware to support the assessment process, the third column of Ta-
ble 1 presents how this relationship is applied to teaching com-
puter security by highlighting the advantages and disadvantages
of using MCQs as an assessment method.
Using Student Generated MCQs to Achieve Active
The literature reviewed indicates some work has been under-
taken in the area of student generated questions for learning and
assessment purposes. This section presents two examples of
studies that applied the concept of student generated MCQs for
the purpose of achieving active learning.
Palmer and Devitt (2006) conducted a study about construct-
ing MCQs as a method for learning. Their study was performed
to measure the influence of student-based construction of MCQs
as a stimulus for the learning and understanding of topics in
clinical surgery (Palmer & Devitt, 2006). The hypothesis of this
study was that in order to create high quality MCQs, deeper
understanding of the material involved in the question would be
created (Palmer & Devitt, 2006). If the students can understand
the method of construction of a good MCQ, “it will teach the
difference between mere knowledge acquisition and how that
knowledge can be used in terms of comprehension, application,
analysis and evaluation” (Palmer & Devitt, 2006).
At the completion of the study it was found that the MCQs
created were of a high standard and the students looked more
favorably at the exercise, although many found that it did not
replace traditional teaching methods (Palmer & Devitt, 2006).
In terms of assessment the students “produced a large bank of
potentially viable MCQs which are perfectly suitable for both
formative and summative assessment purposes” (Palmer & De-
vitt, 2006). The following outlines some of the other significant
findings of this study (Palmer & Devitt, 2006):
one quarter of the questions created were evaluated to be
capable of testing higher order cognitive skills;
the majority of the MCQs created tested knowledge and
both teachers and students need to be educated about the
benefits of untried learning methods for them to be more
receptive to new learning initiatives;
an unanticipated weakness of the learning strategy in this
study (which is often viewed as a strength of an MCQ as-
sessment) was the inability to assess a wide range of mate-
rial; “the students may have gained a deep understanding of
their particular area of study, but they obtained only a su-
perficial understanding of other areas, thus showing no sub-
stantial net increase in their overall understanding”.
The authors suggest that this problem may be rectified if
students are required to create a larger number of MCQs. How-
ever we would suggest that this would only create a larger bot-
tleneck in terms of marking. Instead we would recommend an
alternate approach which would involve peer review of the
created MCQs. This would be a topic for future research.
Yu & Liu (2004a) conducted their study into the perceived
potential value of student MCQ construction in the introductory
physics laboratory. The study was performed to determine whether
a more active learning atmosphere in physics labs would be
cultivated by having students construct MCQs pertaining to
interacting physical phenomena during the learning process (Yu
& Liu, 2004a).
The hypothesis of this study was that the creation of MCQs
would stimulate students to be more attentive to their tasks and
be more reflective on their own thinking and learning (Yu &
Liu, 2004a). Similar to the study performed by Palmer and De-
vitt (2006), in the process of constructing the MCQs, students
would be more likely to be intellectually active in order to gen-
erate a question-stem, provide the correct answer to the posted
question and ponder three plausible distracters that can essen-
tially distinguish the students who have learned the concepts
from those who have not (Yu & Liu, 2004a). In terms of active
Copyright © 2013 SciRes.
D. HUTCHINSON, J. WELLS
Copyright © 2013 SciRes. 119
Relationship between key elements of assessment and the use of MCQs for assessment in relation to our teaching.
Teaching Computer Security
Key Elements of Assessment Assessme nt Using MCQs Advantages Disadvantages
3 contemporary assessment issues in
Australian higher education (Lublin,
potential of online assessment;
assessment design for large classes;
Immune to the influence of bluffing
(not guessing) (Burton, Sudweeks, Merrill,
& Wood, 1991).
test knowledge quickly with large
students own work;
Formative assessment (improving):
tasks that form developmental or
ongoing teaching/learning process;
provide ongoing feedback to the
student to improve their learning &
decision making (Lublin, 2000;
Phye, 1997: p. 514).
Excellent opportunity to offer efficient
Comments usually limited to indicate
Need explanation for wrong answer and
marking key & explanation for correct
answer (Ramsden, 2003: p. 188);
Total score—how the test taker
compares to others.
prompt feedback on student
(Phye, 1997, pp. 16-17; Lublin,
Difficult and time consuming to
design and develop e.g. 1 hour per
question (ACVIM, 1997);
Feedback often withheld to save
expense of constantly designing new
tests (Lublin, 2000).
tasks that often occur at the end of a
used primarily to indicate how much
student has learned e.g. assign
grades (Lublin, 2000).
Excellent opportunity to offer
Section score—strengths and
weaknesses can be understood;
Summative MCQ tests work best when
question bank built up over time:
from questions subjected to item
used to indicate its degree of difficulty.
Students’ tend to revert to
rote-learning mode when they know
they will be tested by MCQs (Lublin,
Reliability of assessment—extent to
which results of assessment can be
trusted (Miller, 1987: p. 73; Toohey,
1999: p. 180; Lublin, 2000).
(Burton et al., 1991):
Reliable; no argument about the right
Not affected by scorer inconsistencies
(like essay questions).
(Phye, 1997: pp. 16-17; Lublin, 2000):
Marking is reliable, cheap, &
Can be reused/item banked.
High reliability & motivating
regular study (Toohey, 1999: p.
Validity of assessment—extent to
which the assessment reflects the
learning objectives of the unit (Miller
1987: p. 73; Toohey, 1999: p. 180;
Can be invalid (Burton et al., 1991).
Group of people should be involved in
Generating the questions and trialling
(Phye, 1997: pp. 16-17; Lublin, 2000):
surface approach to learning i.e.
swotting for recall.
At mercy of one person’s way of
communicating the discipline.
Good assessment starts with author
determining objectives & use of
results (Phye, 1997: p. 186; Lublin,
2000; Censeo, 2007).
Student understanding of assessment
process facilitated by (Lublin, 2000):
explaining assessment approaches,
how approaches relate to the unit’s
objectives & criteria by which
students will be assessed.
e.g. problem-solving abilities; vs.
recall of information only.
A standard multiple-choice test item
consists of two basic parts:
1. a problem (stem); question or an
2. list of suggested solutions
(alternatives); contains one correct or
best alternative (answer) and a
number of incorrect or
inferior alternatives (distractors)
(Burton et al., 1991).
Select and write down the general
Then write down the testing point.
According to (Burton et al., 1991;
ACVIM, 1997; Phye, 1997: p. 521;
Lublin, 2000; Censeo, 2007) criteria for
writing a “good” MCQ include keeping the:
alternatives homogeneous in content;
alternatives free from clues as to which
response is correct;
grammar of each alternative consistent
with the stem;
alternatives parallel in form;
alternatives similar in length.
(Burton et al., 1991): Low
(knowledge of terms, facts, methods,
and principles) vs. high-level
(comprehension, application, and
D. HUTCHINSON, J. WELLS
must specify observable, preferably
measurable, changes in the learner’s
behaviour at the end of the course
(Miller, 1987: p. 10);
help teachers plan more efficiently
for students’ learning experiences
(Miller, 1987: p. 14).
Suggestions for recommended actions
to take based on test results e.g. self-study
plan (Censeo, 2007).
Students appreciate feedback that
assists them to improve future efforts
(James, McInnes, & Devlin, 2002).
Students’ tend to revert to
rote-learning mode when they know
they will be tested by MCQs (Lublin,
Category of educational objective for
Cognitive—objectives are those
dealing with the acquisition of
knowledge; recall, processing,
stating information (Bloom, 1956;
cited in Miller, 1987: p. 51).
Promotes the development of deep
approaches to learning (Ramsden, 2003).
Designed for the cognitive levels they
can test (ACVIM, 1997);
e.g. Teaching IT Security; securing a
small business information system.
MCQ: Knowledge (+);
Comprehension (+); Application (+);
Analysis (?); Synthesis (−); Evaluation (?).
Key: (+) = usually; (−) = sometimes; (?) =
rarely (Bloom, 1956, cited in Miller, 1987:
(Burton et al., 1991):
les to new situations
& solve problems.
Allow broad coverage, higher
order thinking (Lublin, 2000;
Censeo, 2007; Phye, 1997).
(Burton et al., 1991; Lublin, 2000):
unique thinking processes;
progressive problem solving;
learning they found that MCQ construction helped make stu-
dents monitor consciously and actively their own learning’ and
concluded that MCQ construction “is an instructional strategy
with great potential” (Yu & Liu, 2004a). The following outlines
some of the other significant findings of this study (Yu & Liu,
by identifying important principles or concepts that were
difficult to comprehend students engaged in cognitive strate-
gies skills indicative of active, self-regulative learners, in-
cluding rehearsing, organizing and continuously monitoring
their state of cognition;
it made students “think and reflect more on physics-related
questions and phenomena”, and “discuss more frequently
and intensely with group members”.
“it seemed to transform the classroom into a more reflective
and inquisitive learning atmosphere that are more active
Further studies conducted by Yu & Liu into active learning
through student generated questions in physics experimentation
classrooms continued to support the positive potential of stu-
dents creating their own MCQs. Their findings were that “MCQ
generation activity qualitatively changed students learning be-
havior, and helped students become more active learners” (Yu
& Liu, 2004b).
From these studies it is evident that using student generated
MCQs as an alternative form of assessment offers potential for
enabling active learning. However as indicated there are several
issues that must be considered for success of this learning
method to be realized.
The aim of this study was to show that a student’s knowledge
on a given topic will increase after they create MCQs relating
to that topic, and that by following predefined criteria and
guidelines, students would create high quality MCQs which can
be then redelivered in an online environment as either assess-
ment or self-review tests for revision. This section presents a
description of the methodology undertaken for this study.
Method for Collecting Evidence
The method for collecting evidence involved performing an
experiment. The aim of the experiment was to compare the re-
sults of a pre-test set of MCQs with a post-test set of MCQs
completed by the participants. Part of the experiment involved
analyzing the quality of students’ formal assessment and using
this as evidence of improved learning. The following is a de-
scription of the procedure undertaken.
Procedure for the Experiment
Students enrolled in ITSC were invited to take part in a three
part experiment to determine if creating MCQs has an educa-
tional value. One activity undertaken as part of the experiment
was an assessment task. However the students had the choice to
opt in or out with regard to being part of the experiment. In this
case the students’ results were excluded from the sample group
and in no way had an adverse affect on their course result.
Part 1: A paper based MCQ quiz (pre-test) with a mixed
level of difficulty associated with each question was provided
to participating students to individually complete at the start of
their tutorial class in the second week of semester. The hy-
pothesis made here was that they have not learned or only have
limited knowledge of the material to be covered at this stage.
Due to the theory based nature of ITCS the question and learn-
ing style employed was more knowledge based rather than
problem solving. The pre-test required each participant to recall
knowledge on a given topic and apply that knowledge to find
the correct answer. The pre-test was collected and marked by a
third party. For the purpose of feedback only a result was pro-
vided to the sample group at this stage.
Part 2: All enrolled students were required to complete the
assessment task of individually creating 15 MCQs on a given
topic relating to their studies. The students were given four
weeks to complete the task. The students were given a brief
description of the topics which were framed as short answer
questions as an impetus to create the MCQs. They were re-
quired to research the topic using the prescribed textbook to
create high quality MCQs. “As most successful student gener-
ated question instructional studies appeared to involve either
direct instruction or explicit written instruction on question
generation” (Wong, 1985; cited in Yu & Liu, 2004a), two me-
thods of training were provided before having students con-
struct MCQs on their own. First participants were provided
Copyright © 2013 SciRes.
D. HUTCHINSON, J. WELLS
with a set of criteria for creating good quality MCQs (Burton et
al.). Second a complete example consisting of a sample MCQ
and its elements was provided as a guide to what was expected
for the assessment.
One of the key challenges with using MCQs is the bottleneck
associated with creation, formatting and managing of the test
bank and conversion for use in an online learning system e.g.
Blackboard (Blackboard, 2007). With this in mind, a software
application called “Test Monkey” was used as a tool to enable
students to enter their MCQs and manage the test bank. This is
represented in Figure 1.
MCQs are well suited for use in a software solution because
answers are provided for the student to choose from, thus the
computer system is not required to attempt to interpret a stu-
dent’s answer, it instead matches the student’s response to the
correct answer, giving a final result. All students were provided
access to and were required to use this software to enter their
questions. The questions the students created were marked based
on the criteria for creating “good” MCQs as displayed in Fig-
The students were required to access the software from a
Web site. The intention was to enable uploading of the test
bank into an online learning system for future use.
Part 3: A second MCQ quiz (post-test) comprising the same
questions as the pre-test was provided to the sample group to
complete 6 weeks later following completion of the assessment
task. The reasoning behind using the same quiz MCQs was to
later enable the direct comparison of where the participants’
knowledge increased the most. Again the paper based post-test
was completed by the participants in their tutorial classes. The
post-test was collected and marked by a third party. For the
purpose of providing meaningful and quality feedback the par-
ticipants were provided with: the correct answers to the MCQs
and the relationship between the Test Monkey rating criteria,
the MCQ creation guidelines and the assignment general re-
The results from the pre-test and the post test were compared
in an attempt to show that allowing students to create assess-
ment content is beneficial from a learning perspective. Since
the students were not provided with the correct answers for the
pre-test, the assumption was that learning would take place
which can be measured because the sample group would have
to engage with the learning activity on a deeper level to pro-
duce “good” MCQs.
A summary of the method of inquiry is provided below.
Sample group—participants enrolled in ITCS.
Part 1: Pre-Test: Students completed a MCQ Quiz on the
Test monkey software application.
MCQ rating criteria in test monkey.
Organizational Security and Human Factors;
Cryptography and Public Key Infrastructure;
Standards and Protocols and Network Security;
Infrastructure and Remote Access Security.
Part 2: Students researched, created and submitted a set of
MCQs on the same topics:
The MCQs were graded by the teacher based on defined
Part 3: Post-Test:
The sample group completed a post-test using the same
questions as the pre-test.
Analysis of results:
Results from the pre and post tests were compared to de-
termine whether student generated assessment does encour-
Analysis of student learning:
Determine if learning has occurred;
Compare scores from pre and post tests to measure im-
Analyze if questions of high quality from part 2 produce
high results in post-test from part 3.
In this experiment any improvement in learning was meas-
ured by comparing the results of the pre-test and post-test MCQ
quiz before and after the creation of the MCQs in the assess-
ment task. As indicated the same ITCS topics were used in
order to measure the difference between what was known be-
fore and following the creation of MCQs.
Results and Discussion
The focus of the analysis was on the improvement of a stu-
dent’s marks after creating MCQs, to justify the educational
benefits which MCQ creation provides. It was anticipated to see
students improve their test scores after creating MCQs and that
Copyright © 2013 SciRes. 121
D. HUTCHINSON, J. WELLS
there would be a correlation between the performance on the
post-test and the mark awarded for the assessment task. This
section delivers the findings and discussion of the results re-
flecting on what the findings mean in relation to the research
A series of tables and charts were used to display the results
from the experiment. The results and basis for discussion have
been taken from these displays.
A sample of 30 students accepted the invitation to partake in
the experiment. The students first completed the pre-test and
the results were collated. The pre-test determined the student’s
knowledge of the material before they commenced the assess-
ment task of creating MCQs. As indicated by the percentages
and grades from the students’ pre-test results it was clear that
the student’s knowledge was not of a high standard. This,
however was expected because the students had not been ex-
posed to the topics covered in the pre-test. The pre-test was
designed in this manner so it could be determined if creating
MCQs would enhance the student’s knowledge on the given
The analysis conducted showed the average mark obtained
on the pre-test was 23.27 (51.71%) out of a possible 45 marks.
This means on average students only just passed the pre-test by
a margin of .77 or 1.71%. The maximum mark achieved was 32
(71.11%) and the minimum was 14 (31.11%). The median
score was recorded at 23.5 sitting one mark above the pass
An analysis of the students grades was also displayed and it
is interesting to observe that 10 students or 33.33% failed the
pre-test (which is rather high when a normal bell shape curve is
considered), a further 14 students achieved a pass grade and
only 5 of the students were able to obtain credits. Finally only 1
distinction was awarded and no high distinctions were allo-
A range of statistics in relation to the pre-test results, includ-
ing the mean, mode and standard deviation were all recorded
around the pass mark of 22.5, with the results being 23.3, 23.9
and 4.59 respectively. The negative skew of the data confirmed
that the students understanding of the topics tested in the pre-
test was low to non-existent. From these findings it was ex-
pected with some confidence that any knowledge obtained to
improve the students’ results in the post-test would be obtained
in the assessment task.
The histogram presented in Figure 3 shows the frequency
distribution of the results obtained in the pre-test.
In summary the results of the pre-test were low, as expected.
The data collected had a normal distribution so statistical sig-
nificant testing was determined to be appropriate.
The next step in the experiment was to conduct the assess-
ment task which represented the learning process. Students were
required to create MCQs and those questions were rated by the
teacher to determine if they were of a high quality. The criteria
against which the MCQs were rated are displayed in Figure 2.
To show the level of correlation between a student achieving
highly in the assessment task and highly on the post-test, the
raw scores the students achieved on the assessment task as well
as the percentages were calculated. As displayed in Figure 4
the level of correlation is high with two outliers.
The final step in the experiment was having the students take
the post-test to determine if by creating MCQs they were able
to increase their level of knowledge on the tested topics. An
increase in post-test score would indicate that the assessment
Histogram of pre-test results.
Correlation of assessment task vs. pre-test.
task did have a positive affect on the students learning and
would contribute to validating the research question.
The results showed that the post-test score increased from the
pre-test after the students undertook the process of creating
MCQs. The highest improvement was a 33.33% increase; with
this test score increasing 15 marks from a 21 (Fail) to 35 (High
Distinction) marks. This was the highest improvement. The
average improvement was an increase of 4.57 marks represent-
ing 4 more questions being answered correctly on the post-test
than the pre-test. In general all students were able to achieve a
higher post-test score. In fact the fail rate in the post-test fell by
13.33% down to a more reasonable 20% when a normal bell
shaped curve is considered.
Further comparison between the results of the pre and post
tests indicated that the average post-test score increased by 4.56
to 27.83 out of 45, moving the average result to a low credit, an
increase of a whole grade. The maximum mark on the pre-test
achieved after the completion of the assessment task was 39,
just 6 marks off 100%, a difference of 7 marks from the pre-test.
The minimum mark was raised to 17, a difference of 3; the
median showed an increase of 4.5 moving to 28 and four less
Copyright © 2013 SciRes.
D. HUTCHINSON, J. WELLS
people failed the post-test.
One of the more significant findings was the change in grades.
Three participants achieved high distinctions and there was a
major increase in students moving from a pass to a credit grade.
In relation to the research question this shows that it was possi-
ble to significantly raise the level of knowledge that students
were able to recall after they undertook the creation of MCQs.
The statistics calculated for the post-test, including the mean,
mode and standard deviation all increased to, 27.83, 28 (18.4%
of the sample achieving this score) and 28 respectively. The
data shifted from a negative to a positive skew and overall the
results of the post-test were higher than what was previously
achieved on the pre-test.
The histogram displayed in Figure 5 clearly indicates an in-
crease in the post-test scores in comparison to the pre-test
scores. The highest achieved results are now located around the
27.5 mark, compared with previously identified 25 mark. As
displayed the frequency has increased for higher marks and the
graph is plotting higher test scores, starting from 17.5 compared
with 15 from the pre-test.
The t-test performed was a Paired Sample t-test and was the
most important part of the results in terms of validating the
research question. The paired samples t-test compares the means
of the two variables and computes the difference for each case,
testing to see if the average difference is significantly different
from zero. This test determined whether the assessment task
had a statistically significant affect on students test results.
First this test needs to have paired data which means that the
pre-test and post-test scores were taken from the same individ-
ual and second the data must be normally distributed which was
The null hypothesis tested was that there is no significant
difference between the means of the pre and post test; the al-
ternate hypothesis which was hoped to be achieved was that
there is a significant difference between the means of the pre
and post test, meaning that the students did gain knowledge
from the creation of MCQs.
The Paired Sample Statistics showed the statistics for both
sample groups and were used to calculate the significance. The
Paired Sample Correlations showed that the data sets had a
positive correlation of .692 meaning that students who did well
on the pre-test generally did well on the post-test, confirming
that the data is consistent and that statistically students did not
perform worse on the post-test.
The final test was the Paired Samples Test and this indicated
if the differences in the results were statistically significant. The
analysis indicated an increase in the mean of 4.5667 and a large
improvement in the lower and upper bounds of the 95% confi-
dence interval, recording −5.9730 and −3.1603 respectively.
The T value was −6.641 and the degrees of freedom equaled 29.
The final figure that this test presents is the significance of
the 2-Tailed-t-test. The significance value for the paired sample
(pre & post test) record was .000 which means that the null
hypothesis can be rejected. It can be concluded that the differ-
ence between the pre and post test was a significant difference,
validating the educational value of students undertaking the
process of creating MCQs.
Figure 6 displays a Sequence Diagram mapping the results
of the pre and post test against each other. From this it is clear
that the majority of post-test results were higher than the pre-
Figure 7 displays a plot of the assessment task results against
Histogram of post-test.
Sequence diagram pre-test and post-test.
Assessment task vs. post-test results.
Copyright © 2013 SciRes. 123
D. HUTCHINSON, J. WELLS
the post test results. It is clear to see that the majority of as-
sessment results were higher than the post-test results. This
provides further justification that the creation of MCQs resulted
in effective learning.
All the analysis presented in this section demonstrated that
there were statistical significant differences in the pre and post
test results and that having students create MCQs in ITCS as
part of their assessment does have a valid learning benefit.
This inquiry set out to implement a new learning strategy to
facilitate a more active and inquisitive learning atmosphere than
what was currently being experienced. It was contemplated by
incorporating an assessment requiring students to construct MCQs
which could have a positive impact on students’ approach to
learning. By quantitatively analyzing the results from the pre
and post tests and comparing the results of the post test to the
assessment task, the desired outcome of improving student
knowledge and understanding in relation to the concepts and
practices of computer security was realized. The significance of
the method was proven in raising the level of knowledge gained
by these students.
Using “Test Monkey” enabled a unique style of feedback al-
lowing students to review the expert rating for each of their
MCQs as marked by the teacher. This process provided each
student with a customized view of their results, and importantly
not just a mark. It is our belief that this provided opportunity
for self-reflection to enhance their learning experience and
motivated students to study and keep up to date by encouraging
a more interactive approach to learning.
The results suggested that students may have been involved
in a deeper approach to learning and the development of cogni-
tive skills through a more intimate interaction with the unit
material. However the degree of deeper learning that this in-
quiry achieved is not conclusive at this stage. This is under-
standable as it is “not a realistic expectation for students to
produce MCQs testing higher order cognitive skills at their first
attempt”. Also “exercises of this nature are not likely to be
greeted with much enthusiasm as they involve learning methods
unfamiliar to many students” (Palmer & Devitt, 2006). Further
analysis of the data is required, for instance to determine and
rank the overall quality of the MCQs created.
From a teaching perspective the experience was certainly
valuable on several levels. It provided the opportunity to try
something new. It exposed what students did not know and
enabled informed adjustments to the teaching to target those
areas earlier on in the semester. It has provided a resource for
future use and most importantly feedback about ITCS for re-
flection and improvement.
The major dilemma experienced was the marking of ap-
proximately 2000 MCQs all at one time. This created a bottle-
neck in the assessment process. To overcome this dilemma the
assessment model should be adjusted so that students would
create MCQs on a weekly basis; thus the assessment would be
spread over the duration of the semester. This would be ex-
pected to encourage a more progressive and active approach to
learning allowing students to receive feedback on a regular
basis to help them improve their learning strategy. This again
would provide direction for teaching; more towards those spe-
cific areas on what our students don’t know or understand.
In terms of future work it is planned to perform additional
inquiries across multiple disciplines to assess and validate the
effectiveness of this assessment approach on a wider scale.
Also it is intended to introduce another feature into the software
to allow a peer review process of the MCQs. In this way stu-
dents would be able to learn from each other which is thought
to add additional value to the learning experience.
ACVIM (1997). Question-writing guidelines. American College of Ve-
terinary Internal Medicine.
Blackboard (2007). Blackboard homepage, Bl ack boa rd Inc.
Brown, S., Rust, C., & Gibbs, G. (1994) Strategies for diversifying
assessment in higher education. Oxford: Oxford Centre for Staff De-
Burton, S. J., Sudweeks, R. R., Merrill, P. F., & Wood, B. (1991). How
to prepare better multiple-choice test items: Guidelines for university
faculty. Bringham Young University Testing Services and the De-
partment of Instructional Science.
Censeo (2007). Guidelines for writing effective tests: A practical “short
course” for test auth ors. Censeo Corporation.
Elton, L., & Johnston, B. (2002) “Setting the scene”, assessment in
universities: A critical review of research. London: Generic Centre
Learning and Teaching Support Network (LTSN).
English, L. D. (1998). Children’s problem posing within formal and in
formal context. Journal for Research in Mathematics Education, 29,
James, R., McInnes, C., & Devlin, M. (2002). Assessing learning in
Australian universities. Canberra: Australian Universities Teaching
King, A. (1995). Inquiring minds really do want to know: Using ques-
tioning to teach critical thinking. Teaching of Psychology, 22, 13-17.
Leung, S. S., & Wu, R. X. (1999). Problem posing with middle grades
mathematics: Two real classroom examples. Mathematics teaching
in the middle school. Reston, VA: National Council of Teachers of
Yu, F., & Liu, Y. (2004a). Perceived potential value of student multi-
ple-choice question-construction in the introductory physics labora-
tory. Proceedings of International Conference on Engineering Edu-
cation ICEE-2004, Gainesville.
Yu, F., & Liu, Y. (2004b). Active learning through student generated
question in physics experimentation classrooms. Proceedings of In-
ternational Conference on Engineering Education ICEE-2004, Gai-
Lublin, J. (2000). Guidelines for good practice in assessment. Univer-
sity of Western Sydney, Campbelltown: Centre for Enhancement of
Learning and Teaching.
Miller, A. H. (1987). Course design for university lecturers. New York:
Nichols Publishing Company.
Palmer, E., & Devitt, P. (2006). Constructing multiple choice questions
as a method for learning. Singapore: Annals Academy of Medicine.
Parker, L. (2004). Intersecting sets: Carrick, assessment and evaluation.
Key Note Address Presented at the Evaluations and Assessment Con-
ference. Melbourne: RMIT.
Parry, S. (2004). Student assessment. Key Note Address Presented at
the Evaluations and Assessment Conference. Melbourne: RMIT.
Phye, G. D. (1997). Handbook of academic learning—Construction of
knowledge. California: Academic Press.
Ramsden, P. (1988). Improving learning—New perspectives. London:
Copyright © 2013 SciRes.
D. HUTCHINSON, J. WELLS
Copyright © 2013 SciRes. 125
Ramsden, P. (2003). Learning to teach in higher education (2nd ed.).
London & New York: Routledge Falmer.
Silver, E. A. (1994). On mathematical problem posing. For the Learn-
ing of Mathematics, 14, 19-28.
Singer, H. (1978). Active comprehension: From answering to asking
questions. Reading Teacher, 31, 901-908.
Sircar, S. S., & Tandon, O. P. (1999). Involving students in question
writing: A unique feedback with fringe benefits. Advances in Physi-
ology Education, 22, 84-92.
Toohey, S. (1999). Designing courses for higher education. Bucking-
ham, UK: SRHE and Open University Press.
Wong, B. Y. L. (1985). Self-questioning instructional research: A re-
view. Review of Educational Research, 55 , 227-268.