Formative Assessment thorough Pupil Dilation Tracking

During the Summer of 2018, I, along with my colleague and two undergraduate interns at the Computer Science department at KITE developed a piece software, using Open Face library and simple PC Webcams, to track pupils (as in pupil of the eye) of students while they were watching videos in class, listening to the instructors during lectures, or taking various tests on the computer. The basic idea is to gather Big Data of all students, every few minutes, throughout the semester, about students’ attention to lectures and/or quality of lectures of instructors and/or whether the student is learning /interested in lectures in class or not. Once the software is fully fine-tuned and the Artificial Intelligence (AI) piece of the software has learned various aspects of analyses, we intend to mass deploy this software to help us with formative assessments and grade prediction of students, before it is too late. I.e. at the end of semester, it is too late to make corrections; we want to be able to intervene, if necessary, within the first couple of weeks. This we believe will significantly enhance quality of teaching in our traditional classroom setup at our University.

The Theory - The primary job of the pupil is to control the amount of light that enters the eye. What most people already know is that pupils dilate when it is dark, to allow more light into the eye. However, pupils also dilate due to problem solving, interesting visual stimuli, and load on memory ([1], [2] and [3]). Moreover, experiments also show that people with low test scores on the SAT examination, had larger average pupil size [4]. The reason is that the pupils also dilate under stress. In short: “Anything that activates the mind, or anything that increases the mind’s processing load, also causes the pupils to dilate” [5].

My personal use case, as an instructor, is to use this software and camera to track eye-pupils of my students in real-time as a tool for formative assessment. I.e. while I am teaching in class, the software is giving me live feed for each student attention and/or learning taking place with a 30 second sampling. I.e., is the student getting it? If I see that any student is not getting it, then I intervene with specific questions for that specific student and ensure that a back and forth dialogue takes place, until the concept has been clarified. I usually do that in my teaching practice, but normally, I am making guesses about the student’s learning and/or interest in the subject matter by looking at their faces. With this piece of software, on my computer screen, I am getting a live feed as to what is happening in their heads, almost. This greatly helps in recursive and formative assessment. [7]

This entire experiment is generating a lot of Big Data at our institution. I.e. data about each student’s attention span during lectures and processing load on memory throughout the semester for each course they take. We eventually plan on using Artificial Intelligence to help us analyze this data during Summer of 2019 to answer the following questions and many more:

a. Are the students finding the lectures interesting or boring? And why? Are they finding the material too hard? Does the material have enough visuals to keep their attention?

b. Is the student class attendance directly correlated with his grades?

c. Is formative assessment helping the student keep his interest for the next one?

d. With the help of this technology, can the instructor ensure that NO student is left behind and deploy strategies during each class session for specific students?

REFERENCES

[1] Hess, E. H., & Polt, J. M. (1960). Pupil size as related to interest value of visual stimuli. Science, 132(3423), 349–350. DOI: https://doi.org/10.1126/science.132.3423.349

[2] Hess, E. H., & Polt, J. M. (1964). Pupil size in relation to mental activity during simple problem-solving. Science, 143(3611), 1190–1192. DOI: https://doi.org/10.1126/science.143.3611.1190

[3] Kahneman, D., & Beatty, J. (1966). Pupil diameter and load on memory. Science, 154(3756), 1583–1585. DOI: https://doi.org/10.1126/science.154.3756.1583

[4] Ahern, S., & Beatty, J. (1979). Pupillary responses during information processing vary with Scholastic Aptitude Test scores. Science, 205(4412), 1289–1292. DOI: https://doi.org/10.1126/science.472746

[5] Beatty, J. (1982). Task-evoked pupillary responses, processing load, and the structure of processing resources. Psychological Bulletin, 91(2), 276–292. DOI: https://doi.org/10.1037/0033-2909.91.2.276

[6] OpenFace library: https://cmusatyalab.github.io/openface/

[7] Cope and Kalantzis, (2018), Formative and Recursive Assessment: https://www.coursera.org/learn/ubiquitouslearning/lecture/wApUR/assessment-as-recursive-feedback