Resources

Behind the method: what Exploratory Factor Analysis does, and why it matters for our study on student choice 

Ask a student why they chose their university and they will give you a considered answer.

Ask a student why they chose their university and you will get a thoughtful answer. They will tell you about the course, the campus, the entry requirements. All of it is true, and none of it quite explains why two students with similar grades, similar postcodes and similar offers end up making completely different decisions. 

That gap is where our analysis starts. 

Over the past few months we have run discussion groups and a national survey across our three partner universities, and by the time fieldwork closed we had responses from more than 800 students. We asked about course and career, about cost and location, about belonging and support, and about the people around them and what those people thought. Every question tells you something, but on its own, none of them tells you very much. 

This is where Exploratory Factor Analysis comes in, and it is worth explaining – and not just because I like talking about research methodology! 

EFA looks across a large set of variables and asks what students tend to say together. Not just what they said, but which answers cluster in a way that suggests they are picking up the same underlying thing, even when the questions look quite different on paper. 

So when a student tells us they worried about fitting in, that they looked for people like themselves on campus, and that they were not sure university was really for people like them, those are three separate survey questions. What EFA shows is that they are also three expressions of the same thing. This gives real depth and nuance to the answer – in a quantifiable and robust way, but also groups them into a factor. Do that across the whole dataset and dozens of individual questions resolve into a much smaller set of constructs, which are the real dimensions of how a student decides. 

That matters here for a specific reason. The sector has a great deal of data on what students say influences them. It has much less on which of those influences actually predicts whether a student goes on to apply. EFA is the step that takes us from description to structure. 

The factors then feed the next stage, which is the regression modelling that tests which of them most strongly predicts application intent. That is what produces the Student Decision Index, a ranked framework built from what students told us of the drivers that carry the most weight. 

I should say that EFA is not a neutral tool. The decisions a researcher makes along the way, which rotation to use, how many factors to keep, how to read the loadings, all shape what comes out. We have made those decisions carefully, and we will be open about them at the conference and in the published work. 

What I can say now is that the structure is there in the data, and it is clearer than I expected. I can’t wait to reveal it to the sector. 

The full findings, and the Decision Index, will be shared at the Hidden Drivers conference on 13 October at Nexus, University of Leeds. 

Thank you for reading.