People participate in vocational education and training (VET) for a variety of reasons and at different stages of their life. Some undertake VET to gain necessary vocational skills to enter the labour market for the first time, while others enter in order to upgrade existing skills, learn new ones, or simply for personal interest.
The issue of passing and successfully completing a VET qualification may not be the prime objective for all students. This issue, together with the fact that not all people have the same ability to cope with the education and training demands required of some qualifications, suggests VET qualification completion rates may not be enough to determine the full effectiveness of the sector. A number of different performance measures exist; however, little information is available concerning the likelihood of success for individual students and which students are more or less likely to succeed. Consequently, there is a need to identify the various learner groups undertaking VET and determine their likelihood of success.
As such this project will investigate which individual, course and provider characteristics are the most important in determining the likelihood of whether or not a student will complete their training, including interactions between these characteristics.
The overarching research questions to guide this work are:
- Understanding the current state of play, what is the profile of students most likely to complete a VET qualification and least likely to complete a VET qualification?
- How could we improve VET completion?
Using a decision-tree algorithm, a profile analysis could be performed on the VET program enrolment data to identify the current state of play. It involves measuring common characteristics within a population of interest. Demographics such as average age, gender are some typical variables we could include in such profile analysis. In this aspect, we are assessing which categories of students have the most potential to complete a VET qualification, with hope to find a subset of the available inputs that could explain the VET completion.
Logistic regression will also be used for this project. While many other techniques are available, logistic regression is most preferred here because (1) when it is done correctly it is very powerful; (2) it is straightforward; and (3) it has a lower risk of over-fitting the data. Logistic regression is an excellent technique for finding a linear path through the data that minimises the error.
The possible types of characteristics to be included in the modelling are listed below. Note this is not an exhaustive list:
- Provider type and provider size
- Age at time of enrolment (e.g. 15-24/25-34/35-44/>=45years)
- Delivery mode (mode of attendance: internal/external/workplace-based)
- Indigenous status
- Program qualification level
- Disability status
- Program field of education
- Language spoken at home
- Training state of delivery
- Employment status at time of enrolment
- Type of attendance (full-time/part-time)
- Highest educational attainment at time of enrolment
- Accreditation type
- Geographical location (urban/ regional/ remote)
- Socio-economic status