Using ClassDojo Data To Spot And Correct Bias
3 min read
According to the ClassDojo home page, over 35 million teachers, students, and parents now use ClassDojo. That's a lot of data, and we should be mindful about the ways in which the rights to the that data are assigned.
The goals of ClassDojo - also as stated on their home page - are to:
Improve behavior
Improve specific student behaviors and engagement by awarding and recording real-time feedback.
Share data
Print or email beautiful behavior reports to easily engage parents and staff.
Save time
Save time by recording behaviors and accomplishments right in class, with just one click: NO extra data entry required.
However, while teachers are recording data on students, and at times using that data to create behavior walls for public display in their classroom, the app is also capturing huge amounts of data about teacher behavior. What students get rewarded? For what do students get rewarded? Are students in a particular class more likely to get penalized for behavior during a specific classroom activity? Are students more likely to be penalized for behavior during a specific time of day? Do boys get rewarded or penalized more than girls? Is there any variability across gender or race based on subject? What patterns exist among teacher communication to parents?
From what I understand, ClassDojo doesn't specifically collect information on race, but because data analysis works, some of these details can be inferred with general accuracy over a data set. If ClassDojo worked with school districts within FERPA to get this data back to schools and teachers, they could be close to 100% accurate around demographic details.
If ClassDojo chose to, they could give incredibly valuable feedback to teachers. Their data could demonstrate if teachers showed bias in how they gave feedback in multiple ways. Given the size of their data set, they could probably detect patterns of bias - or lack of bias - within schools and districts. This would be incredibly valuable information for teachers, schools, and districts to have as they planned ongoing teacher professional development, and would also help get models of more successful school cultures. This would also be incredibly valuable for parents to know when they were working with teachers: if a teacher has a track record of bias, a parent has a right to know, and ClassDojo's dataset almost certainly has the components to analyze this.
Of course, it's highly unlikely that ClassDojo would ever reveal this information, as their business model is predicated on marketing to teachers. If ClassDojo were to turn around and hold up a mirror that reflected an unflattering picture, it would not go over well. But, the data is almost certainly there - teachers have given it to ClassDojo with every mouse click and behavioral record entered.
And, nothing in the ClassDojo terms of service prohibits this type of analysis or use. As I said earlier, if ClassDojo offered this type of analysis as a service, it would be incredibly valuable. Teachers could use the data they entered to spot when their lessons fall flat, if there are specific times of day where they have difficulty with their class, and/or if they have different responses based on race and gender.
