EXAM SCORES

---

INTRODUCTION

---

About Dataset

The dataset Students Performance in Exams is taken from the platform Kaggle.

Context

Marks secured by the students

Content

This data set consists of the marks secured by the students in various subjects.

Acknowledgements

http://roycekimmons.com/tools/generated_data/exams

Inspiration

To understand the influence of the parents background, test preparation etc on students performance

About Research

This study has five Research Questions:

1. Which exam is the hardest?

2. Male or female?

3. How effective is the test preparation course?

4. Which major factors contribute to test outcomes?

5. What would be the best way to improve student scores on each test?

In this study you will see all the steps:

1. Data cleaning

2. Data analysis

3. Conclusions

---

Data Cleaning

---

About data cleanliness

This dataset is completely clean. There are no NULL values or values that contradict the logic. For example in the columns math score, reading score, writing score there are no values that exceed 100 or are negative. So we were relieved of important work.

math score

reading score

writing score

---

Data Analysis

---

Which exam is the hardest?

Logically speaking, it's a math exam. Let's see if that's actually true.

SELECT ROUND(AVG(math_score),
		2) AS avg_math,
		ROUND(AVG(reading_score),
		2) AS avg_reading,
		ROUND(AVG(writing_score),
		2) AS avg_writing
FROM students 

avg_math	avg_reading	avg_writing
66.09	69.17	68.05

Apparently he is. The math exam is the hardest, followed by the writing exam and the easiest is the reading exam.

Male or female?

Now look who's doing better on which exams.

SELECT gender,
		ROUND(AVG(math_score),
		2) AS avg_math,
		ROUND(AVG(reading_score),
		2) AS avg_reading,
		ROUND(AVG(writing_score),
		2) AS avg_writing
FROM students
GROUP BY  gender 

gender	avg_math	avg_reading	avg_writing
male	68.73	65.47	63.31
female	63.63	72.61	72.47

As we see women do much better on exams like wirting and reading, and men do much better on math

How effective is the test preparation course?

Let's see how useful the course has been in preparing for the exam for students

SELECT test_preparation_course,
		ROUND(AVG(math_score),
		2) AS avg_math,
		ROUND(AVG(reading_score),
		2) AS avg_reading,
		ROUND(AVG(writing_score),
		2) AS avg_writing
FROM students
GROUP BY  test_preparation_course

test_preparation_course	avg_math	avg_reading	avg_writing
none	64.08	66.53	64.50
completed	69.70	73.89	74.42

Look at this! The courses for exam preparation proved to be quite effective. Students who took the preparation courses did 8.77% better in math, 11.06% better in reading, and 15.38% better in writing.

How effective is the test preparation course? Definitely yes. Courses to prepare for the exam have shown their effectiveness.

Which major factors contribute to test outcomes?

Let's analyze how race, parental education, and lunch(standard or free/reduced) affect test scores. To do this, we need to write three queries

SELECT race_ethnicity,
		ROUND(AVG(math_score),
		2) AS avg_math,
		ROUND(AVG(reading_score),
		2) AS avg_reading,
		ROUND(AVG(writing_score),
		2) AS avg_writing
FROM students
GROUP BY  race_ethnicity
ORDER BY  race_ethnicity

race_ethnicity	avg_math	avg_reading	avg_writing
group A	61.63	64.67	62.67
group B	63.45	67.35	65.60
group C	64.46	69.10	67.83
group D	67.36	70.03	70.15
group E	73.82	73.03	71.41

As we can see group E passes the best exams in all subjects, and group A on the contrary is the worst and passes exams much worse in all subjects

But let's not jump to conclusions and say that group E is just better than everyone else and very smart. Let's take a look at how many percent of each group took courses to prepare for the exams. Maybe that's it?

SELECT race_ethnicity,
		 ROUND(COUNT(*) FILTER(WHERE test_preparation_course = 'completed')::DECIMAL
         / COUNT(*) * 100,2) AS completed
FROM students
GROUP BY  race_ethnicity
ORDER BY  race_ethnicity

race_ethnicity	course_completed
group A	34.83
group B	35.79
group C	36.68
group D	31.30
group E	42.86

It is true that Group E did best in the exams because a large number of students took exam preparation courses

Now let's see how parents' education affects their children's exam results

SELECT parental_level_of_education,
		 ROUND(AVG(math_score),
		2) AS avg_math,
		 ROUND(AVG(reading_score),
		2) AS avg_reading,
		ROUND(AVG(writing_score),
		2) AS avg_writing,
		 ROUND(COUNT(*) FILTER(WHERE test_preparation_course = 'completed')::DECIMAL / COUNT(*) * 100,2) AS completed
FROM students
GROUP BY  parental_level_of_education
ORDER BY  parental_level_of_education

parental_level_of_education	avg_math	avg_reading	avg_writing	completed
associate's degree	67.88	70.93	69.90	36.94
bachelor's degree	69.39	73.00	73.38	38.98
high school	62.14	64.70	62.45	28.57
master's degree	69.75	75.37	75.68	33.90
some college	67.13	69.46	68.84	34.07
some high school	63.50	66.94	64.89	43.02

As we can see it does make a big difference. Students whose parents have a master's degree pass best of all with only 33.90% having taken exam preparation courses with them.

Now look at how lunch (standard or free/reduced) is affected. Lunch shows the income of the household. For poor families it is free/reduced.

SELECT lunch,
	   ROUND(AVG(math_score),
	   2) AS avg_math,
	   ROUND(AVG(reading_score),
	   2) AS avg_reading,
	   ROUND(AVG(writing_score),
	   2) AS avg_writing,
	   ROUND(COUNT(*) FILTER(WHERE test_preparation_course = 'completed')::DECIMAL / COUNT(*) * 100,2) AS completed
FROM students
GROUP BY  lunch
ORDER BY  lunch

lunch	avg_math	avg_reading	avg_writing	completed
free/reduced	58.92	64.65	63.02	36.90
standard	70.03	71.65	70.82	35.19

And it turned out to be a very important factor to passing the exam even the most important one. Students with standard lunch had better results by about 10%

What would be the best way to improve student scores on each test?

To raise your exam results in all subjects the best way is to take an exam preparation course and your results will increase by 10%. But what positively improves the results is the family income and their education, but this is something the student cannot directly influence :(

---

Conclusions

---

It is very important to read more in order to improve in writing exams. I found this out with the help of correlations and graphs that I made with the pandas library and matplotlib.

df = pd.read_csv('D:/university/VSCODE/VSCODE/Markdown/Exam Scores/StudentsPerformance.csv')
dfscore = df[['math score','reading score','writing score']]
dfscore.corr()

	math score	reading score	writing score
math score	1.000000	0.817580	0.802642
reading score	0.817580	1.000000	0.954598
writing score	0.802642	0.954598	1.000000

df.plot(kind = 'scatter', x= 'reading score', y = 'writing score')
plt.show()

Another thing to note is that family income, parental education, and taking a course to prepare for the exams are very influential.

Goodbye !!!