Effect of modeling pedagogy on IB Chemistry HL scores

Three years ago was a very busy time in my teaching career.  I had just attended a three week long workshop on teaching chemistry using modeling pedagogy.  I walked away ready to completely overhaul my teaching methods and even institute standards based grading in my chemistry class.  At the same time I was teaching IB chemistry HL for the first time ever. I had a group of seniors in their second year and a group of juniors starting the course.  Today we received our third set of data and the 2018 group is our first group of students to have been taught using modeling pedagogy prior to taking IB Chemistry HL. The first two groups (2016, 2017) took chemistry prior to IB Chemistry HL, but they took chemistry before I had attended the workshop and were taught using traditional teaching methods.  

Figure 1: IB score distributions from 2016-2018

Figure 2:  IB average scores (3.96, 3.74 and 4.42) from 2016-2018

The data is limited because it only represents 25, 27 and 31 students for the three groups.  There is also likely variation in my teaching abilities as I have become more familiar with teaching IB Chemistry HL.  But the data does match my experiences while teaching the three groups. The modeling pedagogy increased the ability of the students to question the material we learned.  As the content became more difficult there was a presence of mind to push for evidence behind the conclusions I would present. While we did continue to use modeling teaching for some lessons (equilibrium, crystal field splitting) the biggest difference to me was the level of analysis that students applied in their learning was at a deeper level.  They made more connections between content areas, they were able to visualize systems at the particle level and change back to symbols quicker, they had more experience and context to learn in a more permanent fashion.  

None of these comments are an indictment on the first two groups.  All three cohorts of students had remarkably intelligent and successful students that did brilliant work in class.  They all had strong metacognition and did deep levels of analysis of chemistry. I also do not feel that I taught these first two groups poorly.  We did a lot of engaging activities and reflection for all three groups. The modeling approach in the chemistry prerequisite course I believe to have elevated the 3rd cohort to an improvement of that learning.  There is a frequent concern amongst teachers that altering teaching methods will lead to students not being prepared for the most challenging coursework that they will take in the future. While there is a lot of uncertainty and unreliability in the data in Figure 1 and Figure 2 based on student and teacher variation, there is evidence to support modeling having a positive impact on top students that move on to take a course such as IB Chemistry HL.  To get the IB Diploma students need to average a 4 score on their classes with a few technicalities mixed in. The % of students achieving a 4 or better improved from 52% to 59% to 68% over the three years. Many students that get a 3 in HL chemistry still get their diploma and students getting a 3 or better went from 80% to 93% to 93%. Our diploma rates for the three years were steady at 79%, 79% and 82%. It will be another two years before I have a group test as our program has grown and I now only teach every other cohort.

Update with 2020 scores: