More research content to follow!

Spatial Skills

My primary research is in spatial skills and their relationship with computing science success. Spatial skills are cognitive skills relating to conceptualising and consolidating space and structure mentally. In my first ever research paper, which was a version of my Masters dissertation, I described various factors of spatial skills for a CS education audience, provided a conceptual model for the relationship between spatial skills and CS success and detailed the results of an experiment demonstrating that spatial skills tend to be higher in CS students and faculty further in their education or careers. This paper won the Chair's Award at ICER in 2018.

Investigating the Relationship Between Spatial Skills and Computer Science
Jack Parkinson, Quintin Cutts

ICER '18: Proceedings of the 2018 ACM Conference on International Computing Education Research, August 2018

Following this research I began my PhD to determine the nature of the relationship. My first step was to determine whether the relationship was causal. This involved deploying a spatial skills intervention developed by Sheryl Sorby with first year CS students at my institution. I discovered that actively improving spatial skills can have positive outcomes for CS students, particularly less experienced students. This is compelling for two reasons: firstly, spatial skills training is relatively - compared to introductory programming - easy to deliver for significant gains in CS outcomes; and secondly, it is a rare example of a cognitive skill which is transferable, where brain training in one context results in tangible improvements in another.

The Effect of a Spatial Skills Training Course in Introductory Computing
Jack Parkinson, Quintin Cutts

ITiCSE '20: Proceedings of the 2020 ACM Conference on Innovation and Technology in Computer Science Education, June 2020

As a result of this work, spatial skills training became a standard part of the introductory CS curriculum at my institution. The programme was also moved online during the COVID 19 pandemic, which opened doors for adoption of the course at another institution. Changing context led to less substantial results, though some interesting interactions with gender were discovered in collaborative work with the University of Toronto Mississauga, particularly that spatial skills training across a whole CS1 cohort narrows the spatial gender gap but does not close it entirely.

Spatial Skills and Demographic Factors in CS1
Anna Ly, Jack Parkinson, Quintin Cutts, Michael Liut, Andrew Petersen

Koli Calling '21: 21st Koli Calling International Conference on Computing Education Research, November 2021

I collated this work, along with other deliveries at the University of Glasgow and a contribution from Ryan Bockmon, who had run a spatial intervention at three universities in the USA, into an experience report of several deliveries in multiple contexts. We ended this paper with recommendations on how best to deliver training, which appears to be in-person with paper-based instruments, but noted that all the different training contexts saw gains in spatial skills, and generally positive interactions with computing grades.

A challenge that could have been reasonably raised against any one of the intervention studies is that we notice improvements in students' grades because any intervention may yield improved CS outcomes. By taking students aside, paying them extra attention and giving them another reason to be on campus, they might improve in their regular modules as a side effect rather. This may be true, and doesn't necessarily mean that spatial skills training is therefore irrelevant in and of itself, but my original model stipulated that the relationship was cognitive.

Practice report: six studies of spatial skills training in introductory computer science
Jack Parkinson, Ryan Bockmon, Quintin Cutts, Michael Liut, Andrew Petersen, Sheryl Sorby

ACM Inroads, Volume 12, Issue 4, December 2021

In order to explore this more deeply, I attempted to explore the interaction between spatial skills and a much more granular and limited measure of CS aptitude than general module grades. I developed an expression evaluation test using a very limited set of Python constructs - ones the students should all know - with 30 questions of increasing complexity. This replicates the structure of the PSVT:R, a spatial skills test which a pool of student volunteers took alongside the expression evaluation test. Just like with assessment grades, there was a positive correlation between expression evaluation and mental rotation, which was stronger for students who self identified as being less experienced. This study demonstrated that spatial skills still correlate with success in a very limited subskill of early computing aptitude, removed from any other factors associated with their course or assessment.

Relating Spatial Skills and Expression Evaluation
Jack Parkinson, Quintin Cutts, Steve Draper

UKICER '20: United Kingdom & Ireland Computing Education Research conference, September 2020

An opportunity arose for some longitudinal research in 2021. During the summer of 2021 the students who had originally been enrolled in CS0 in 2017, most of whom had taken a spatial skills test as I began my research, reached the end of their degree and graduated. Just prior to their final results being published, I reached out to the cohort to ask to review their academic career with respect to their original spatial skills score in their first year. I had several issues concerning drop-out and non-response, leaving only a small proportion of the original cohort available to examine, but I was able to make some statistically significant observations: there is a strong correlation between students' final GPA and their spatial skills as measured in their first year of study, which was stronger for STEM students. The correlation also appears to grow year-on-year. The students' spatial skills also correlated more strongly with more computationally complex module results than they did with modules which didn't require the construction of complex mental models.

Relationships between an Early-Stage Spatial Skills Test and Final CS Degree Outcomes
Jack Parkinson, Quintin Cutts

SIGCSE 2022: Proceedings of the 53rd ACM Technical Symposium on Computer Science Education V.1, February 2022