Cognitive Load Theory was used to redesign a Database Systems course for Information Technology students. The redesign was intended to address poor student performance and low satisfaction and to provide a more relevant foundation in database design and use for subsequent studies and industry. The original course followed the conventional structure for a database course, covering database design first, then database development. Analysis showed the conventional course content was appropriate but the instructional materials used were too complex, especially for novice students. The redesign of instructional materials applied Cognitive Load Theory to remove split attention and redundancy effects, to provide suitable worked examples and sub-goals and included an extensive re-sequencing of content. The approach was primarily directed towards mid to lower-performing students and results showed a significant improvement for this cohort with the exam failure rate reducing by 34% after the redesign on identical final exams.
Student satisfaction also increased and feedback from subsequent study was very positive.
This case study describes the application of Cognitive Load Theory (CLT) to the redesign of an introductory Database Systems course, which had historically received low student satisfaction and relatively high student failure rates. Past students had also displayed a lack of understanding of some key concepts as evidenced by performance in exams as well as student feedback. The course was considered to be highly technical and challenging for the profile of students, who are primarily IT students based in a business school and in their second year of university studies.
The course was (and is) a compulsory course in the Information Technology degree at an Australian regional university. Students are expected to have completed “Systems Analysis and Design” as a prerequisite, and in third year, students use their new skills to complete two capstone “Computing Project” courses. Some Database Systems students also continue to advanced web programming, which includes significant use of databases to build dynamic websites.
Observed problemsHistorically, the conventional Database Systems course had a high failure rate. For example, in the year before the redesign the failure rate for the 36 students who completed the course - excluding those students who withdrew or ceased to engage with the course before the exams - was 33%. Including the students who did not complete the course assessments but did not withdraw, the failure rate was 47.4%. In addition, verbal feedback from academic staff delivering the web programming course and capstone project courses indicated that some students who passed the Database Systems course struggled with putting their skills into practice to actually design and develop databases, and generally did not display self-efficacy in database skills.
Understanding the structure of SQL and the ability to build SQL queries is considered integral to database study (Robbert & Ricardo, 2003) and had been a core focus in the course materials and activities, yet students performed poorly in the SQL section of the end-of-semester exam. Students also showed a lack of understanding as to why they were required to learn SQL. For example, an email from a student at the end of the teaching semester enquired:
“What is the benefit of SQL? I mean to say that in Ass 1B and Ass 2 I made queries in design view and it works correctly in less time, and there is no need to join the tables through code - it joins auto. In SQL we have to write code and it takes more time sometimes for new [learners] like me it’s harder, so why SQL????” (student correspondence)
This observed deficiency in SQL query construction skills was confirmed by informal comments by local industry employers -- which the university maintains relations with for the purpose of internship placement -- who indicated that new graduates were generally poor at using SQL in conjunction with other languages in web-based contexts.
It was determined that no change was required to the actual content of the course. The existing structure and resources were reviewed, along with student feedback. Cognitive Load Theory principles were then applied to minimise extraneous cognitive load.
Relational database design and creation, as for any specialised field, utilises its own terminology or ‘jargon’. For many students, this terminology can be a major barrier to learning, especially when studying new technical topics for which they have limited or no existing schemas. By segmenting the material, concepts can be introduced without jargon first to allow students time to form schemas, only using the correct terminology exclusively once concepts are understood. In the course redesign, concepts were introduced first using layman’s terms and simple examples and the exclusive use of jargon was avoided early in the introductory topics. Wherever possible, topics were introduced by relating it to the students’ existing schemas, allowing them to build on their existing knowledge.
For example, in their first assignment, students were asked to create a database for a university, utilising the students’ existing understanding of degrees, courses, teachers, workshops and assessments. Many other day to day examples of databases and their use were provided during the course, from the very simple (such as shopping lists) to more complex online e-commerce sites (such as Amazon) and search engines, and were used to demonstrate concepts using the students’ pre-existing schemas.
Sequence Issue 1: SQLThe conventional design of the Database Systems course covered SQL commands from basic SELECT statements through to advanced topics, such as sub-queries and inner-joins, in two consecutive topics presented mid-way through the course. It is likely that presenting the SQL content in this format did not give students enough time to consolidate and automate their learning of terminology, definitions and the hierarchical organisation of concepts. It also did not apply the SQL commands in the context of other concepts. This approach is likely to have led to cognitive overload for the less able students and consequently impeded their learning of SQL generally.
In the redesign of the course, SQL was distributed and integrated into all but one of the topics and was, importantly, introduced in small, learner-paced segments to provide students with the opportunity to automate and master segment content before moving on to the next segment. The content provided a consistent theme of building on their SQL knowledge at the end of each topic, including providing relevant worked examples. Online quizzes provided to students were used to consolidate this knowledge.
Sequence Issue 2: Design First, Create SecondThe conventional design of the course introduced the concepts of needs analysis and data modelling (such as Entity Relationship Diagrams) before the students had an understanding of database fundamentals -- such as data types and relationships -- and before good design was discussed. In the redesigned course, the same content was re-sequenced (see Figure 3) so that analysis of data needs and modelling of data were introduced after the students had an understanding of the relational database model and had experience creating databases using controlled examples as templates that slowly built in complexity. It is argued that this sequence better supports an approach that works from simple and concrete to complex and abstract, thus increasing the hierarchical complexity of associated schemas in a suitably controlled and meaningful construction.
Sequence Issue 3: ConsolidationThe spreading of SQL across all topics resulted in ten topics, leaving two weeks available for consolidation. The content was re-sequenced to include the ten topics and two case studies (see Figure 4) with students expected to cover one section of the content per week of the teaching semester. The case studies were also designed to perform as large scale worked examples for the design and use of databases and to reinforce concepts covered in the previous five topics.
Removal of Split Attention EffectsContent in the supplied PDF study guide was presented in a way that integrated the text explanations into screenshots and explanatory figures to reduce cognitive load imposed by split attention (Figure 5). This is a relatively quick and easy modification to materials using a word processor.
Worked Examples:Worked examples were used wherever possible in the supplied study guide.
Student satisfaction with the course is ascertained by a student feedback survey conducted by the University each semester. The survey contains a mixture of 5-point Likert-scale questions and opportunity to comment on each question on the survey in text answers. Six Likert-scale questions about the course itself included “The workload demands of this course are about right” and “This course helped me to develop some valuable skills/attributes” while six further Likert-scale questions on the teaching in the course included such questions as “The staff member/s shows a genuine concern for the quality of my learning” and “The staff member/s presents the subject matter clearly”. On all Likert-scale question a score of 1 indicates “strongly disagree” whilst 5 indicates “strongly agree”.Averaging the scores for the course and the course teaching gives two satisfaction scores - one for the course itself, and one for teaching/staff -- with a maximum score of 5 each. Course and teacher feedback scores were collected for 2 years before the redeveloped course delivery, and again after the delivery of the new course.
The same end of semester exam was given to students in the Semester before redevelopment, and at the end of the first delivery of the redeveloped course. The rest of the assessment was changed, but this allowed us to ascertain the impact of the redesign/redevelopment.
There was a significant improvement in student satisfaction with the course, at the alpha = 0.05 level (Mann-Whitney U Test, Ua = 5318, z = -2.3, p = 0.0107). At the same time, student satisfaction with teaching performance remained high, and in fact improved significantly with the redesign (Mann-Whitney U Test; Ua = 4171, z = -3.14, p = 0.008).
Text comments offered from students as part of the course feedback survey for the redesigned course include such positive comments as “… prepared very, very well.” and “... communicates complex concepts and ideas in a clever and easily understood manner.” The change in attitude of the students towards the importance of SQL query construction skills - in contrast to the conventional course - was demonstrated by the following comment: “I do not think there was enough emphasis on learning the Structured Query Language.”
The only meaningful comparisons that can be made between the student performance in the conventional course and the re-designed course are those comparing the examination results, as the other items of assessment were changed as part of the course re-design. An identical exam was given to both cohorts of students and exams were marked by the same marker, using the same exam marking guide.
The raw exam mark results are shown in Table 2 below. The difference in exam scores was found to be statistically significant (t = 4.8636, p < 0.0001).
Not only did the mean score improve by almost 20 marks (out of a total of 100), but the lower bound of the range was significantly increased. In the conventional course, approximately 14% of students scored < 20 marks in the final exam. In the redesigned course, using the same exam script, the lowest score was 42.5. The reduction in fail rates from the conventional version delivery to the redesigned delivery was significant (Fisher Exact Probability Test: p = 0.0013).
In the past, there have been many concerns expressed by academics teaching courses which follow on from the Database Systems course. After the redevelopment, the coordinator for the advanced web programming course that follows the Database Systems course provided feedback:
In previous years, students have been provided with a limited set of Database permissions to complete the requirements of their assessments – that is a Database driven website. This year those students that completed the Database I course requested additional permissions on the shared server so they could set referential integrity constraints. This has never happened before. During the semester it became clear that these students were undoubtedly more prepared than previously, with an advanced understanding of database design, construction and queries.
There are some limitations to this work. The results are not from a controlled study, and are from one delivery of the re-designed course only, with a relatively small sample size. The course results themselves could not be compared meaningfully, as the assessment regime changed between the two deliveries.The re-designed approach applied several well-defined principles and effects from Cognitive Load Theory - primarily the removal of the split attention effects, utilisation of the sub-goal effect and promotion of the use of worked examples. The materials could be further examined and improved by including some additional effects from Cognitive Load Theory.
