Thanks to my partner, A, and Jess, Dean and Michael for workshopping this post!
I have a huge sweet tooth. If there’s one thing I love as much as justice, spreadsheets and pajamas (ok… and a hot plate of pad see ew), it’s probably desserts. Just yesterday, I had the world’s most delicious affogatto (decaf espresso poured over nutella ice cream). I am staring at a table full of fillings to make a batch of Hamantaschen cookies in time for Purim. And I’ve never met a homemade, raisin-free cookie I haven’t wanted to devour.
This week, a special data analysis request came my way. My delightful partner (who’s high school students have become accustomed to this awesome seating chart tool we built together) came up with another data challenge. His school has ~650 entries in the restorative justice discipline log (powered by Google Forms). His principal asked for a report on discipline trends and themes. If you try to go at it the old fashioned way (skimming the log, jotting notes on a napkin or perhaps in an email draft…), well, you might just need an Old Fashioned! (I take mine without the whiskey). I took on the project because I knew Excel tools could make the analysis quicker and more comprehensive, and help my partner and his colleagues achieve their school culture goals.
What did we learn?
- 120 students had at least one entry in the restorative justice discipline log
- 31% of the disciplinary write-ups come from the top 10 most challenging students
- February and December had the highest rate of write-ups per school day
- The two staff members with the most write ups each had more than twice as many as any other staff member
- When I summarized write-ups per student, the range was 46, the average was 5.8 write ups per student, the median was 3 and the mode (most commonly repeated number) was 1. Huzzah! Data a la mode!
What tools did I use?
- Mean, Median, Mode: these classic statistical calculations are easy to do in Excel and tell us valuable information about the dataset! Don’t be afraid to try simple solutions first — then move onto the tougher stuff.
- Pivot Tables! Pivot tables were my go-to tool for this data project because they let me summarize the data by student name, teacher name, type of infraction, etc. I’ll save a Pivot Tables tutorial for another post, but for now, just remember, if you need to sort and add up categories (how many times does this name appear in this dataset?), Pivot Tables are a good tool for that! (Now in Google Sheets too!)
- Charts: Once I summarized my data, I made a lot of bar graphs, line graphs and pie charts because they were useful to show relationships. Then, I shared the charts with the school administrators so they wouldn’t distracted by extraneous calculations. There’s a lot of power in making charts! To mitigate some of that power, I always write detailed captions explaining the data sources and conclusions.
- Pulling in additional data: One of the first things I did when I got the dataset was summarize the data by month (thankfully, each entry in the restorative justice discipline log had a time stamp #TheDataAreAlright!). But that doesn’t tell us much if there is a drastically different number of school days in each month (for example, winter break in December!) So I found a datatable from the local school district website that showed scheduled school days per month. Then, I divided (Number of Write Ups per Month) / (Number of School Days per Month) so that we could compare the rate month-to-month.
I talk a lot about making meaning out of data– that’s what I want to support changemakers to be able to do! Looking more closely at these student write-up records can help guide a strategy toward pursuing restorative justice and student growth and accountability. My goal was to leap from 650 entries in the restorative justice discipline log to identifying trends and outliers, and using that information to recommend actions.
There is a LOT of research out there about data-driven decision making. Summarizing that research is outside the scope of this post, but I did want to introduce a model (thanks to my mentor, Dean) that offers a helpful framework for these types of questions.
Let’s take a look at DIKW, also known as Data -> Information -> Knowledge -> Wisdom (this version (photo below) adds Decisions to the model – which I applaud!)
Here’s another way of looking at the process I used to make meaning of the data. I went backwards and applied the DIKW(D) model to my work – with two different examples.
|Principle||Application 1||Application 2|
|Data||There are 650 entries in the Restorative Justice Discipline Log.||There are 650 entries in the Restorative Justice Discipline Log.|
|Information||We can summarize these data by Student (pivot table!)||We can summarize these data by Month (pivot table!)|
|Knowledge||Some students have more write ups than others.||Some months have more write ups than other months.|
|Wisdom||31% of write ups come from 10 students. 40 students each only have 1 write up.||February and December have the highest rates of write ups per school day.|
|Decisions||Will we have more impact by focusing on improving behavior from the top 10 students (31% of write ups) or the rest of the students?||Why do some months have a higher write up rate than other months?|