Using changes in pupil dilation to assess cognitive load in complex problem solving remains a challenge in chemistry education research. Previously, we presented the first results of our study, in which we verified our method against Klingner’s vigilance and memory studies and measured pupil dilation while undergraduate chemistry students solved problems from the Chemical Concept Inventory (CCI). Our goal was to examine the extent to which the difficulty of questions from the CCI correlated with mental effort indicated by pupil diameter. This next stage of work aims to develop a comprehensive and robust statistical method to analyze pupillometry data when students are performing actual assessment tasks that require multiple steps of thinking over minutes rather than over a few seconds. Additionally, a measurement issue with looking at problem solving over minutes is that pupil diameter changes with any light changes in the testing environment as well as with movement. To address both these challenges, we use multimodel inference to quantitatively assess the goodness of fit of different models, including a generalized least squares time series stationary analysis and a G-ANOVA. We also use student interviews to qualitatively assess interaction effects, such as identifying whether a student perceived a question to be more or less difficult than the reported difficulty in the CCI. By using a multimodel approach, we hope to identify a) trends in cognitive load during problem solving and b) the parameters that need to be altered or controlled for in future experiments. Our goal is to design a robust and standard approach to analyzing pupillometry data. We report results of taking into account change in pupil diameter over time by looking at maximum pupil diameter in smaller time steps and performing a time series analysis over the entire problem solving episode.

Assessing pupillometry data using multimodel inference

253rd National Meeting (2017)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.