Baccalaureate Data Visualization

How did Romanian high schoolers perform at the exam?

The Romanian Baccalaureate is the final exam that a student must pass in order to graduate high school. Despite being an important news topic each year, there is no in-depth analysis of the results in the media. Most of the time, we only see the national pass rate and a per-county map. Therefore, I decided to write an article with figures to give some insights about the subject.

Data is provided by the Government’s open datasets website. The figures are interactive, hover over them to get more information. Be aware that I will use Romanian language for figures since it would be confusing to translate terms such as profilul uman.

Past results

In recent years, the national pass rate has been around 65-70%. Out of 127.001 non-absent students at the 2021 Summer Baccalaureate, 69% passed the exam. Note that the Ministry of Education formula of calculating the rate does not include absent students, even if they signed up for the exam.

The urban-rural divide

Urban high school students are far more likely to pass the exam (71%) than their rural counterparts (56%). Nevertheless, this isn’t the only worrying statistic. According to the National Institute of Statistics, 46% of Romania’s population lived in rural areas in 2017. However, students that go to rural high schools account for only 32% of all Baccalaureate participants. This can be a sign of rural families that prefer sending their children to urban high schools and a sign of high dropout rates in rural areas.

Girls get higher grades

Girls are more likely to pass the Baccalaureate (71%) than boys are (60%). The latter are prone to getting eliminated for trying to cheat the exam (115 compared to 32 girls eliminated) despite there being fewer boys participating in the national exam. This can be explained by their risk-taking behavior, or maybe girls are better at not getting caught by the teachers. Boys are also more likely to not show up at the exam (6%) than girls (4%).

School specializations

Romanian highs chools have a wide variety of specializations to choose from. We can view them on the following treemap chart, which is composed of rectangles ordered from largest to smallest. The area of each rectangle is determined by the number of students. The most popular specialization is Matematică-Informatică (24.275 students).

Although it is known that students from certain specializations perform better, there are few official sources or mainstream news that show this gap statistically. Coloring the treemap proportionally with the pass rate, we can observe significant differences. For example, 85% of Matematică-Informatică students passed, whereas only 10% of Tehnician în agricultură students passed.

Students can also pick the subject of their third test. The most popular choice is anatomy (72% pass rate). Almost all students that picked chemistry and physics have passed (96% and 94%).

Biased grading

In order to pass the exam, the student needs a grade higher than 5 in each test and an average of 6. Out of 127.001 non-absent candidates, 146 managed to get a perfect final grade.

Rather than look at the final grade histogram, we can break it down into the multiple exams to observe some interesting patterns. There are three big numerical spikes:

  • grade 1.00: students that didn’t prepare at all for the exam
  • grade 5.00: students that barely passed
  • grade 10.00: students with maximum grade

Notice the big dropoff just below the pass grade threshold followed by a huge spike. There are barely any students in the 4.90-4.99 range, while the next one has a hundred times more students. The histogram should be smoother, but some teachers would rather artificially increase the grade to 5 than to fail a 4.9 student.

The second exam (mathemathics or history) has the highest number of students that couldn’t even get 0.1 points (2822 students).

The third exam, which has a selectable subject, is the easiest one for getting a high grade: 7223 students managed to obtain the maximum grade, compared to 2723 and 1354 at the other exams.

Conclusion

I hope you found this article helpful in understanding the students’ background and results. The figures were made in Python Plotly. The source code is on my GitHub repository and data on the Government’s open datasets website.