Thoughts on the IB Results — A Dive into the Numbers this year

Since IB results were published on Monday, many students and teachers around the world have been scratching their heads trying to figure out how their scores were calculated. Some of these students and teachers have accused the IB of a lack of transparency and have cried ‘unfair’ in response to the grades awarded. Some of the concerns are justified; some are not. As I like to think of myself as someone who understands how numbers are generated and what they mean, here is my understanding and explanation of why scores look the way they do this year.

Firstly, to avoid potential criticism of being an IB apologist, I run a Secondary School which saw almost all students achieve lower scores than the predicted grades we submitted. We had a number of high-achieving students see a significant drop in their totals (5–7 points) and they will need to reassess their university options over the coming weeks. Having said that, our own ranking of our students from top to bottom academically would be more-or-less in line with the results generated from the IB this year. There wasn’t a big shake-up. Average IB scores for our students are in line with those of previous years.

Here are the key points to consider when looking at your school’s IB results this year:

Predicted Grades

It is frankly absurd to imagine that results were going to be in line with predicted grades (this year or any year). If any school communicated to students that they expected them to be awarded their predicted grades, then they have done their students a huge disservice. Predicted grades are always higher than actual scores when you look at student totals.

The IB very quickly published its global data proving that average scores this year were in line with previous years’ scores, not only in terms of DP totals, but also in each subject group. They have to keep results consistent from year to year. University offers are based on an understanding that the percentages of students achieving each level in a subject will remain constant every year. If there is too much variation, universities will not be able to accurately predict the proportion of offer holders who will meet their conditions; therefore, the system will break down (i.e. too many students will meet the offers)

This point should have been obvious to IB teachers and school leaders ever since results started being published. Don’t forget — grade boundaries are never revealed until after results are published. Why? Because grade boundaries are changed every year to ensure that the percentage of students getting each grade remains consistent. Getting a 7 doesn’t mean that you have reached a certain level of achievement in a subject — it just means that when your work was placed together with the work of all other IB students, you achieved in the top 5–7%. Grades are always comparative in nature. They mean nothing without the context and the competition.

I have seen some teachers shouting that their predicted grades should have been accepted as the final grade. Nonsense. A system like that would result in two ugly consequences: (1) universities would simply increase the scores required for admission (i.e. offers of 36 would have to become offers of 38 or 39), and (2) parents in fee-paying schools would put incredible pressure on teachers to submit inflated predicted grades. Schools should not be in favour of either of these possible outcomes.

The Role of Internal Assessments

The IB stated in February that Internal Assessments were going to play a more important role in the calculation of final grades this year, as in most subjects, these assessments were the only student work available. Yesterday, component grades were released for each school, so teachers can now see the weighting that the Internal Assessments played in the calculation of final grades.

When I look at our students’ scores, for example, I see that in English Language and Literature, the raw score for the IA seems to have been multipled by around 1.6, and the raw score for the written tasks seems to have been multiplied by around 2.5 in SL and about 1.25 in HL. The final score awarded to the student is the sum of these two scores. When I look at the Maths scores, I see that the raw score for the IA has been multiplied by 5, Psychology IA raw scores seem have been multiplied by around 4.5, Biology IA’s by around a factor of 4, etc…. Aside from English Language and Literature, the final totals for other subjects seem to be based on the IA raw score multipled by a given factor and then moderated up or down based on some other factor (more on that below).

We can, therefore, assume that IA’s which used to account for 20–30% of a student’s grade now account for maybe 50–80% of a student’s grade (ball-park figures obviously).

This should have been obvious to all teachers when the IB announced how it would be calculating grades. If a school did not then turn its attention towards providing quality feedback to students on their IA drafts (there was still time in most subjects), then they probably made a poor strategic decision. Teachers should have been reaching out to colleagues around the world to ensure that they had the best possible understanding of IA assessment before providing students with feedback. Hopefully the IB has been able to ensure a reasonable level of standardization with IA grading this year despite needing to use newly trained IA examiners — obviously, we will never know.

Global Factors and School-Based Factors

This leaves us with the portion of the grade that could have been impacted by school-specific moderation factors based on previous years’ scores. This is what most schools and teachers should be most concerned about.

Here is what could have happened:

(a) All raw IA scores in a subject are decreased by a certain number of points to ensure that grade boundaries don’t need to change that much in order to still have the same percentage of students receiving each score. Remember that IA scores are typically the highest component score for students — if IA’s accounted for 100% of a student’s score, grade inflation would occur.

(b) The IB trusts the teachers’ predictions, and raw IA scores are modified up or down based on the predicted grades to calculate the final score.

(c) The IB looks at the scores from exams in recent years for each school and modifies student grades based on the performance in exams of previous cohorts in each school.

Or a combination of all three!

There is nothing unfair about option (a); however, if IA’s accounted for almost all of a student’s grade, it would have been nice to know. If option (b) is true, this is good news for the teachers claiming that their predicted grades should have mattered more, although it brings with it all kinds of concerns when you consider the lack of standardization and agreement amongst IB teachers when assessing work.

Option (c) is the harsh one, as students attending a school that had not performed well in the past will have been unfairly graded down. I can understand the IB needing to do this, however; otherwise, teachers and schools around the world this year would just overpredict scores with impunity. Still, the academically strong student attending a school in which teachers have overpredicted in the past will have suffered here.

Now here’s a closer look at our DP Biology data for this year:

Biology Standard Level

Biology Higher Level

This data strongly suggests that (a) scores were modified to fit into realistic grade boundaries, and (b) predicted grades played a role in final totals. Here’s why.

Final totals for students with very high IA raw scores were decreased significantly. This is presumably because it’s easier to do well in IA’s than in exams. If scores hadn’t been moderated down at the top end, there would have been too many 7’s. This is a shame for students who had low to mid-range 7’s in IA’s, as they would have slipped to 6’s. Only students with top 7’s would have held on to their 7.

Students with lower IA’s who were still predicted a 7 had their grades moderated down by a much lower factor. This suggests that the teacher’s optimistic prediction had a positive effect, even at the top of the scale.

Students with low predicted grades (see students 5,6 and 7 in SL and Student 6 in HL above) had their IA raw total moderated down by a bigger factor than those with similar IA raw scores and higher predicted grades.

Students at the lower end of the scale had their raw scores increased or moderated down only slightly. This, presumably, is to ensure that the number of students failing is consistent with previous years. It happens every year.

If a school-specific factor was introduced here, it would have to be extremely specific to certain levels of achievement and historical predicted grades. The evidence I am looking at suggests that predicted grades (and the ranking of predictions within a school) are likely to have played a much greater role in final scores than any blanket moderating factor applied to a school.

In Conclusion

It is highly unlikely that the IB will release the algorithm used to calculate final grades. This would result in people like me going through it meticulously on behalf of our students in order to advocate for higher scores. The system that has been used this year is not perfect by any means, but under the circumstances, it’s far less random than critics are claiming. (It’s also obviously far more complicated than anything I am suggesting in this article — this is just a layman’s basic understanding of what might have happened). In conclusion:

(a) Final grades were obviously lower than predicted grades. They are every year and have to be in order to play a relevant role in the university admission sorting system.

(b) Grades are comparative in nature. Whatever the calculating system is, averages have to remain similar from year to year. If a student’s predicted 7 dropped to a 6 this year, that’s because other students around the world did better in the IA’s.

(c) IA’s clearly played a major role in the calculation of final grades this year.

(d) Moderating factors seem to have been applied based on predicted grades submitted by teachers.

(e) If a school-based moderating factor has been applied, it is in combination with the spectrum of predicted grades submitted by the school in a subject.