New Timetabling Technology and Practices

This report was prepared in collaboration with Clairvaux MacKillop College, to showcase a school’s transformation with education scheduling technology, as well as practices. The adoption of Edval and the associated support has delivered many significant benefits to the College which this report will showcase.

Ability to Model (First full 2015 Draft completed in Term 1 2014)

The introduction of Year 7 students to Secondary School in 2015 is a significant change. Edval technology was shown to be capable of rapid iterative solution variants to timetables, allowing modelling of various aspects. This allowed a proof of concept on the proposed curriculum that included Year 7. This included a complete draft timetable for the entire school, prepared in a single day from electronic requirements, and involved a high degree of the use of algorithms. Auto-scheduling, auto-staffing and auto-rooming were among several tools employed to derive the desired outcome of the modelling process.

Examples of information shown was that the proposed number of new staff were suitable and in fact slightly  above what may actually be required to support the inclusion of Year 7. Or, that the English faculty was a bit short of load and new staff hired would ideally have a secondary teaching area of English. It was also shown that despite the arrival of Year 7, the whole school timetable could be appropriately roomed – including science labs and specialist kitchens, etc.

A basic modelling of structure, or staffing may provide some information, however, a complete, clash-free draft solution scheduled and fully roomed gives a far more accurate picture of resources. Where a staffing only solution may show arrangements are possible with respect to staff ‘loads’, this may not be possible with respect to timetable structures – which inhibit the ability of some teaching load to be appropriately employed.

It was shown to be possible to rapidly adjust parameters such as staffing loads or teachers employed, and dynamically observe the results of these changes in ‘real time’ – allowing a wider range of options to be considered. When modelling changes, there was a reduced bias, as the process was more automated – and discussion on timetable requirements were found to be open and transparent as opposed to determining assignments using traditional methods.

The confidence in seeing a complete draft solution was found to be a positive factor in the process. Also, the fact such a complete draft was able to be produced in Term 1 of the year was impressive. It allowed more visibility on the type of staff to be hired, which was knowledge delivered early on in the year. This was particularly useful given that all schools would be seeking new staff this year for their Year 7 cohorts, so getting in early was helpful.

Australian Curriculum / Nominal hours

The school timetabling software was able to demonstrate a capability to model various different curriculum structures. Clear visibility of impact of changes was able to be shown, including the ability to measure teaching time against the indicative hours purported for each of the Australian Curriculum syllabi. These were measured directly against the timetable and also the school calendar – taking into account holidays, etc.

A registration report was also able to be produced, showing a range of information related to the measuring of teaching time per year and curriculum ‘pie’ breakdown by period, etc.

Chinese / Japanese / HPE

The College asked the Edval representative about the subject ‘Chinese’ last year, indicating they were worried it would not run as demand had been dwindling. This was backed by the limited number of Year 11 student preferences for 2014. Edval was able to model a different timetabling strategy for their elective lines within Year 9, which has resulted in the Chinese class becoming full with students – tremendous demand.

Also, with the support from such a large, active peer-supported class, many students have selected this subject for second semester as well, to carry on for a full year – and appear to consider it much higher than in previous years.

Instead of considering possibly cutting the subject, the College are now confidently running the class twice this year, and students are also far more engaged than in previous years. The College has acknowledged that this is largely as a result of the technology and advice given to them by Edval. The Chinese teacher has expressed significant surprise at this result, saying she has never had so many students in her class, or such an engaged group

–  she simply can’t believe it. Other subject teachers have also inquired, asking if they can get the same ‘timetable magic’ applied to their own subjects to increase numbers or engagement by students

Staffing flexibility

Unlike all previous years, the Edval system has brought a new way of staffing the school. Traditionally a grid  would be prepared, and faculty heads required to staff ‘against’ this structure, which artificially restricted staffing options to the grid. The change to the new way has seen staffing preferences expressed independently against each class, without the traditional overhead of a grid structure. While some were initially apprehensive, the end result speaks for itself. The College has been extremely happy with the staffing assignments, which are perhaps slightly better than may have been achieved via the previous method. These were primarily produced using computer algorithms in the Edval software, with some occasional manual adjustments.

There also seems less effort in the process as there is far less negotiation or need for HOD’s to even consider ‘assignments’ overall – rather just a more simpler expressing of preferences for each class in isolation of others, or staff loads. While this is helpful, the key benefit the College was focussed on was the quality of staffing, and thus education delivery to its students. The staff who are responsible for timetabling feel the new process, supported by new technology, has provided a significant benefit to the College.

It has been realised the new approaches lead to a far better, ‘student driven’ timetable. There is a new focus on timetable requirements in this process, not a fixed arrangement of assignments. In addition, there is almost a new language of timetabling – due to the very different business processes involved.

Year 9 Elective lines

The College runs semester-based Year 9 ‘electives’, two lines per semester. Currently, students make a selection for Semester 1, then mid-year make a selection for Semester 2. So at present, lines for each semester are constructed in isolation from each other. Part of the rationale for this being that students don’t necessarily know what they want to study in second semester before beginning the year, so there was a perceived reduction in administration effort by collecting requests on an ‘as needed’ basis in each semester.

Edval consultants reviewed this existing approach and determined it would be more flexible to generate lines for the entire year together, irrespective of the fact many students will still change their minds on elective subjects mid-year. The reasoning behind this relates to ‘solution space’. The solution space for constructing four lines (whole year) is double that of any one semester solution space. This increase in solution space provides significant additional flexibility for algorithms, which results in these benefits:

• Significant increase in the satisfaction rate of student subject requests
• Ability to reduce the number of classes running, for a direct financial saving in staffing
• Increased flexibility to schedule lines to the grid, where there is one less class, as this also means one less teacher and room needed on these periods
• Increased and improved class sizes, particularly in marginal subjects
• Increased faculty diversity in lines – giving more options for students in each line, and more scheduling flexibility
• Potential ability to run a more diverse curriculum with ‘single semester’ subjects, that may not have otherwise been viable in either semester, under the traditional approach
• Less administrative effort in asking students to re-choose, as fewer have missed units
• More flexibility in staffing across the semester divide, by keeping staff roughly in the same lines in both semesters. This facilitates the later scheduling of lines, which may have otherwise required more timetable change mid-year to accommodate variation in which staff are on those lines
• Ability to maximise class size in (say) Chinese, as a possible single semester subject, but then lock this class to first semester. This gives flexibility to consider not running it in second semester if there is no strong demand from students wanting to carry on, or new entrants. It also provides a psychological benefit to the larger cohort by grouping to one class initially means more students will know others doing that subject – and thus be more inclined to possibly carry on Chinese to a full year in second semester. This outcome would be less likely if there was only a small class in semester one

Year 9 Elective Results

Modelling showed that the College’s ‘The Timetabler’ Year 9 solution, based solely on the semester one preferences, had more than 19 clashes. The exact number was not easily identified, as the use of reserves is not readily accessible in that software without going through viewing student by student.

Edval consultants were able to provide a better result on the same lines, by more efficient allocation of students to the existing lines. This was a higher number of preferences satisfied, but also more important choices satisfied (those missed were lower priority subjects). It was actually provable that Edval’s solution was perfect, as all those missed in the Edval solution were due solely to class size or forced clashes with single subjects – in the lines as provided.

As the school had already captured four student choices, being two primary and two reserves, it was suggested these four be considered ‘whole year’ requests by the students. Edval then generated a new solution, with four lines being a whole year arrangement. The increased flexibility to load balance across semesters provided a dramatic improvement. The result was a much lower level of missed subject choices across the whole year, than the school had achieved in just one semester. This was also done with one less semester class running than the school had proposed – being a financial saving.

Further modelling was then done to identify higher savings. A solution which had 29 missed preferences was achieved across the whole year, but with three less classes – a significant financial saving. Balanced across semesters, this result is 14 missed in one, and 15 missed in the other semester. This was significantly less than the arrangement initially proposed by the school, which was over 20 missed in just the first semester alone.

Furthermore, this solution was running both Chinese and Japanese in second semester, though these were optional. At present, 100% of students who had requested this subject were satisfied in Semester 1, meaning the school could actually achieve a solution with five less classes if there was no ‘change’ in student requests mid- year, to study these subjects for a second semester. It is also possible to consider linking these subjects to those in other years, such as having Year 9 students attend a similar Year 10 language class for Semester 2, without paying for the additional staffing of the second Year 9 class.

While the initial desire to allow students to reselect mid-year for Semester 2 is understandable, it should be noted this is still entirely possible. The consideration of generating lines as a whole year process means we plan whole year, but students should be entitled to CHANGE their selections for Semester 2. Many students will not need to, but those who want to change should not be prevented. Obviously the classes running in Semester 2 will be partly guided by the first year process, but this doesn’t mean it must somehow be ‘locked in stone’.

The Timetabler had one student missing Chinese. This was already a small class, so ideally any missed preference would occur in larger classes. This is to reduce workload in the large class, and also to facilitate peer support work in the small class.

The new approach had the Edval algorithms determining the suggested numbers of classes to run, or which subjects to drop. This seemed to result in a more impressive end solution, where the suggested arrangement of classes to run was seen to lead to a higher satisfaction rate, for a lower overall staffing cost. This feature does not appear to be found in the previous timetabling software.

Timetabled Professional Development Groups

The former Principal, Ms Laura Keating, requested that the timetable facilitate groups of teachers be made available for professional development within school hours. As is the case for many schools, professional development is either run outside the normal school hours, or requires staff to be relieved from classes. This is due to the fact the timetable structure has not been pre-planned to accommodate such an initiative. This area has been managed as a post-construction phase (seeking opportunities where identified and interested staff are free), where better results can be achieved if the requirement was built in to ‘make’ staff available to undertake professional learning experiences that are directly related to the strategic vision for teaching and learning.

It was advised the ideal scenario would be to allow the timetabler scope to determine not only where the Professional Development Groups can be scheduled, but also the composition of the groups.

Allowing the timetable itself to guide Professional Development grouping of staff meant that it can more readily accommodate such needs.

A simple example is to group staff who all teach a given line together, so when that line is not running, they are more ‘potentially’ free. In reality though, grouping has more complex considerations.

The Edval technology supports algorithmic assignment of staff to classes, or modification of staffing where permitted by the user. As such, the Professional Development groups can be scheduled WITHIN the timetable, and with staff assigned as ‘best fits’ the timetable. Where there are requested staff be assigned together for Professional Development, this could also be catered for to a degree. No guarantee can be made that specific staff will be able to fit together in the timetable, due to various restrictions. As the school runs to a whole school line template, there are fewer opportunities to group staff specifically, where this can be more easily managed if the structure was more flexible with junior/middle/senior year groups.

The concept of bringing in more school activity to occur within school time is important, and we actively encourage this approach. It was easy for the Edval representative to demonstrate that this was achievable.

Note also that good timetabling will aggregate free periods together for staff, so they have more block time off. This facilitates teachers getting immersed in lesson planning or other tasks, where more sporadic periods off may be less helpful. This is also good for Professional Development, as it can possibly be best scheduled when they have a light teaching day, so they miss fewer classes on that day if engaged in a full day Professional Development activity. The current whole school template was found to be a bit restrictive in accommodating this need.

We seek further specific details of the Professional Development requirements to better understand this need, such as:

• Are there any specific days or times that Professional Development is best suited for?
• Are there any specific staff who should be ‘naturally’ grouped together by preference? (of either staff themselves, or the school)
• Are there any semi-regular Professional Development activities? Or is it more ad-hoc?
• Is  there  a  desire  for  faculty  meetings  to  be  scheduled  in-timetable  as  well,  similar  to  Professional Development groups? If so, further details on this aspect would also be helpful.

Student Management – Changes to Subject Selections

The process of assigning students to classes, and changing their subject selections was found to be not just much better and faster, but also significantly different. The new workflow and tools provides more information in an easier to interpret fashion. All changes made are done more simply. A high degree of automation is applied, including algorithms that determine and prompt the user in ways not seen before.

If a student wants to drop a subject, the Edval software shows all alternative possibilities instantly, using complex algorithms. For example a student may drop Chemistry, but if they want to pick up Biology – they can, even if this is in another line. Edval may show this change is possible by rearranging the students other classes, to pick up Biology on an entirely different line to the Chemistry class they dropped.

The inclusion of class size, social links, teacher relationship and other aspects also aid this process – which is a one click event. Click the Biology class and Edval has automatically swapped the student from their Legal Studies and Music classes across lines to facilitate this change to Biology.

Mischievous Pairs – Noticed by Teachers

The Edval system allows for registering social links within class lists, known as ‘mischievous pairs’ or ‘buddy pairs’. These links are used when auto-allocating core class lists by the use of algorithms, which consider gender and class size balance as well as social and others. While it seems unusual to have computerised assistance in setting class lists, this has proved beneficial to the school. Apart from the increased visibility of such issues – even when making manual moves, the end result has been very positive. Teachers have actively commented that the

composition of class lists seemed better this year (after done with Edval’s assistance), as there were noticeably less groups of students who should ideally not be together in classes. A comment from one such teacher is:

“Some bad pairings were not as common as in previous years. (To the timetabler) It must have taken time to change their classes, but the teachers are happy.”

Daily Supervisions – Covering absent staff

Whilst the Deputy Principal was offsite attending a conference, he received a call about the absence of a staff member. Not having ready access to a computer at the time, he began talking the School Officer (Jenny) responsible for student management through the process of registering teachers absent, and then covering all their classes with suitable substitutes. The system was so user friendly that he was able to easily walk his staff through an otherwise complex process, over the telephone. He has indicated that he does not believe this would be possible if using the previous software used to do daily lesson covers.

This new approach also achieved a quality result in the cover assignments made. The system’s intuitive nature of prompting the user for likely selections based on fuzzy logic, means the end user need not be aware of many critical aspects of the process. Much of this is covered internally through the system, and appeared to use knowledge coded into the system by the Deputy Principal, using a series of weighted rules for various aspects such as “Does the suggested substitute teacher have a cancelled class that day”, or “Do they usually teach that subject” and so on.

In general, it was estimated that the process of doing daily covers has been reduced in time by well over half, resulting in a significant saving of time.

The ease of use and high degree of automation has also given rise to the possible outcome of daily covers being reassigned as a role from the Deputy Principal to a Senior School Officer. This means the staff would prepare covers but under supervision of the Deputy Principal, who can then be released for other tasks. By acting as more of an oversight to the process and approving the results – the work can be effectively delegated by the Deputy Principal, without losing control or responsibility for the process. It is estimated this approach may well be adopted progressively within the next few months.

The delegation of the daily covers to office staff may initially seem slightly controversial, but note the Deputy Principal’s intentions have already been coded electronically to a degree, in setting thresholds, weights and rules. While he would usually be involved to a degree – he may not need to be at all times, such as if he was sick.

Currently, the process of doing morning covers in schools without using Edval is time consuming and complex, and a process that is more done by a single user. This puts strain on this user, who feels obliged to come in, even if they are sick, as they have a mission critical position, and the school community relies on them. Delegation means less reliance on a single person – removing stress and aiding succession planning.

Exam Scheduling

The Deputy Principal, Damian Bottaccio, has previously been experienced in ‘The Timetabler’ product, including when generating exam schedules. He has found the entire process within Edval to be much easier, much faster, and with a significant improvement in end quality of the examination solution. This includes better equity for staff assigned to exam covers, as well as more appropriate staff matched to exams of their own faculty. The quality of student exam timetables has improved, as the Edval system has significant features and algorithms designed to reduce exam clumping – and spread them out.

Apart from the ease of generation, the Edval system supports a wider range of output formats, making delivery of the schedules clearer and easier. This includes visibility of student totals for groups of exams to aid in rooming, as well as being able to directly associate teachers and rooms to exams in a more automated, optional manner – again using algorithms (auto-cover exam week). This appeared better than if the system was solely doing an auto-cover ‘day by day’, as equity across the entire exam week was being managed.

t was estimated that the process of setting exam timetables using Edval involved a third of the time than that of the current timetabling system.

Financial savings

It was unexpected to realise actual dollar savings from adopting a product that costs the College, being a separate timetable system to the one provided. Despite this, several aspects of the Edval software were found to lead to direct financial savings for the College – including significant assistance to reduce casual teacher expenses.

In just the first two months of school, the College has already estimated direct financial savings of double the cost of the cover software module licence. An estimated saving on casual teacher expenses of approximately $3000 has shown it is quite certain the new technology will be saving the College significantly more than it costs in license fees.

Savings in several other areas are clearly evident as well, such as significantly reduced administration time, but this is harder to quantify. Completing Daily Supervisions and undertaking Exam Scheduling are areas where an inordinate amount of time can be spent if using other products. The College’s experience has shown that valuable time has not be wasted working with Edval. Savings achieved from the unexpected reduction in elective classes has also been shown to lead to a very large saving to the College, and one that also exceeds the product’s fees.

Playground Duty Roster

The allocation of Playground Duties is a task that is usually left towards the end of the timetabling process and would involve staff identifying a preferred location after having looked at their teaching timetable. The end result has teachers negotiating changes to their allocation and not always with the outcome that they had hoped for. This is a common approach that schools would apply.

Edval’s solution allowed for aspects of this process to be continued, but offers a complete solution that is more equitable and respectful of the teaching load that staff are required to fulfil. As an automated process, with very clear steps to follow, staff are still able to identify their preferred area for duty, but this is done at any time prior to the timetable being finalised. The auto staff feature then considers through a series of pre-determined requirements, such as not applying a duty on a day when a teacher has classes throughout the day, the result being a duty roster that has staff allocated in an equitable and efficient manner.

The previous process of staff changing their duties can still be applied and requires only a couple of clicks to make possible. In fact, you could still manage the duty roster in this ‘tried and tested’ manner and find the process of adding the details into Edval in very little time.

eMinerva

A number of technical observations were made of the new eMinerva system adopted by the College. This does not directly have a timetabling module, as this is currently supplied separately. Nevertheless, it was suggested that some areas may be causing difficulty in efficiently integrating and managing timetables together with eMinerva at  present. These include:

• Limitations on naming days, and limitations on naming course / class codes and years.
• Inability to correctly manage Learning Support classes properly, or as desired – for example the system can’t cope with these class codes being stored in a “Learning Support” year – to allow them to be more independently and efficiently managed. When cancelling classes for example, if Year 12 academic classes are cancelled, Year 12 learning support classes should continue to run. This approach is directly and easily supported by Edval, but not when managed to cater to limitations to the eMinerva system which does not appear as effective with non-standard, or more complex timetable elements like learning support, study classes, priority classes, rotation classes, etc.
• Difficulties importing timetable data, as the system does not allow staged imports, takes a long time to process, and may have limited error messages. It seems a laborious process to import the required data, taking up to 20 minutes with each attempt.

Support

The support provided by the Edval representatives has been outstanding, and far beyond what is delivered by many other technology companies. In addition, this support extends to areas well beyond ‘timetabling’ – such as major changes, and restructuring the school in areas, such as how elective classes are constructed, how staffing is managed, and even new roles and other changes within the school, such as the possible movement of morning covers to a School Officer.

Summary

Clairvaux MacKillop College does not endorse commercial vendor products as a general rule, and has no commercial relationship with Edval, other than as a customer. We trust that others will understand that our desire to properly report on the significant and beneficial outcomes of our recent timetabling project are to allow other schools to similarly consider enjoying these benefits.

The company representatives were keen to promote a view that it is the quality of the timetable that matters, not the administrative effort, or cost, to actually produce it. We have found this to ring true. They appear to have had a number of articles published in educational magazines which outlines various new approaches in this field – with the founders having apparently published a number of scientific papers in journals about timetabling. We have been surprised to learn there was actually such a different way to do every aspect of timetabling, including staffing, electives, scheduling, rooming and more. Should any school be interested, we would be happy.