Student attrition is a serious problem, not only in Australia, but all over the world.
Many Australian higher education institutions are trying their best to improve their services in an attempt to retain students. One way of doing this is through the application of Advanced Analytics.
By using Advanced Analytical techniques applied to data such as academic results and demographic information, institutions can create models which predict the likelihood of a particular student to churn. A continuous loop of predicted results compared to actual results allows for improvement in accuracy of these predictive models. Click here to view our infographic and find out how this is possible.
As we all know, human behaviour is complex, and students are leaving their studies for many different reasons. It would be impossible to accurately predict every single permutation with a relative degree of confidence had we only the aforementioned data sources.
In a previous blog post, I also talked about the use of Natural Language Processing in Text Mining to automatise the process of “reading” tweets, and categorising them based on their themes and sentiments.
These techniques can be applied to any data set or sources that contain text. E.g. social media posts, blogs, or survey responses.
Surveys are good mediums for gauging the pulse of the student community, and including open ended questions allows them to fully express their wants, needs, what they are satisfied with, and what they are dissatisfied with. Although the practise of surveying is not a new concept to the educational domain, not everyone is applying text mining to uncover additional insights.
Performing Text Mining on student survey results will not only allow institutions to improve the way in which they currently analyse survey results, but also yield deeper insights into the reasoning.
I have found by looking at the results of international student surveys that many students are complaining about the lack of airport pick up services or welcoming parties; services that are provided in most universities but not promoted enough to have every student aware of them.
The journey to reducing student attrition, however, doesn’t just stop here…
Themes and sentiments extracted from the qualitative survey responses can be used as an additional input to further enrich our student attrition predictive model.
Example: SPSS stream using both qualitative and quantitative data
Other than having an additional input which may result in more accurate predictions, using qualitative data along with quantitative ones can also allow institutions to prioritise areas of improvement directly related to student attrition, such as increasing promotions on existing student services or organising more social events for international students.
Advanced Analytics is data driving decision making. For further information, contact Tridant today.