Behavioural diagnosis to reduce dropouts in apprenticeship programmes





The Behaviouralist partnered with ACT to expand the organisation’s understanding of learner motivation in apprenticeship programmes and embed behavioural science methods across the organisation.

Apprenticeship programmes are a great option for people looking to gain experience and new skills for a specific role, whilst being paid at the same time. ACT, the largest apprenticeship provider in Wales, matches learners with employers, and provides assessors that support learners throughout their apprenticeship. While the take-up of apprenticeships has been on the rise in the UK, so have dropout rates across the same programmes (Gov.UK, 2023). For the first phase of our ongoing work with ACT, we conducted an initial behavioural diagnostic to learn about ACT learners and determine how to predict and address learner dropouts.

 A multifaceted behavioural diagnostic to learn who is most at risk of dropping out

We began by conducting a data audit and interviewing key stakeholders at ACT, to understand the different steps that learners must take to complete an apprenticeship programme and to scope the data collected by ACT across such steps. This enabled us to create a learner journey map that visualises the steps and actions that a learner takes from start to finish. The learner journey map details six key steps that start before the learner applies to the programme and concludes after they graduate. In the map, we identify key actions for each step, as well as potential barriers that can arise and contribute to a learner dropping out of their programme.

Next, we conducted an exploratory analysis using data on learners across two academic years. The analysis aimed to gain a deeper understanding of learners, quantify the impact of different characteristics on completing (or dropping out of) the programme, and to assess the value added of assessors, employers, and providers on learner achievement.  

As part of embedding a behavioural science approach within the ACT organisation, we designed and delivered three workshops for senior members that taught them the principles of behavioural science, walked them through designing behavioural interventions using behavioural science frameworks, and introduced them to evidence-based decision making and the importance of testing and evaluating. 

The exploratory data analysis revealed key information about learners and dropouts, and some potential explanations for why learners drop out

We found that on average, the majority of learners complete their program, but there are significant differences in completion rates between sectors. For example, learners in Education & IT are approximately 50% more likely to successfully complete their program than learners in Health. Completion rates also vary within sectors: for example, migrant workers in the Health sector are nearly 20 percentage points more likely to complete their programme than non-migrant workers.

The drop out analysis revealed that learner, employer, and programme characteristics affect dropout rates. Some examples of factors that influence dropout rates include:

  • Age: learners aged 30-39 are more likely to complete their programmes than those aged 19 or younger.
  • Employer type: Learners in public sector organisations are more likely to achieve than those in private companies.
  • Employer size: Learners in large organisations (over 250 employees) are more likely to complete their programmes than those in micro-organisations (fewer than ten employees).
  • Programme level: Higher-level learners are less likely to complete their programmes compared to lower and mid-level learners

Assessors are the learner's primary point of contact during their apprenticeship and are responsible for supporting them throughout their programme. Our analysis revealed that assessors have a significant impact on whether their learners complete their apprenticeship program or not. We also investigated the impact of employers and providers on learner completion. The impact of assessors was much greater than that of employers and providers.

Building on the findings to design improved methods to predict and address learner dropouts

ACT undertake substantial research and evaluation of the success of its programmes through its annual self-evaluation process. By working with the Behaviouralist, however, ACT was able to gain new insights as the analysis and findings controlled for sociodemographic, employment and programme characteristics.

The second phase of our work with ACT will build on the learnings from the first phase and will focus on conducting a more in-depth diagnostic to understand the reasons for learner dropouts in order to develop new approaches to overcome these barriers to achievement. This will involve collecting new quantitative data, as well as conducting qualitative research, through interviews and surveys with learners and assessors.

Explore more solutions