Project 1: SAT & ACT Data Analysis

Author: Lee Wan Xian

Executive Summary

The purpose of this project is to use exploratory data analysis to advise the admission board of a college in United States (US) on whether SAT or ACT test scores should be a mandatory admission requirement. There are 2 objectives for this analysis. First, it is to understand the relationship between SAT/ACT test scores and the retention rate of undergraduate students. Next, it is to understand the relationship of SAT/ACT participation rate and the general wealth of the respective states. There will also be a brief analysis on colleges who have mandatory SAT/ACT test requirements against those who do not.

This analysis shows that there is no strong correlation between SAT/ACT test scores and the retention rate of undergraduate students as well as between SAT/ACT participation rate and the general wealth of the respective states. To add on, there is no significant impact that mandatory SAT/ACT test requirements leads to a higher or lower college applicants acceptance rate.

In conclusion, we will recommend that SAT/ACT requirement should be left as an optional consideration for admission.

Problem Statement

A college based in United States (US) is accessing the need to make SAT & ACT compulsory for college admissions. We are hired by the college admission board to assess how relevant SAT & ACT scores are in predicting students' ability to cope with the college curriculum & continue on to their sophomore year. As the college has multiple campuses across the country, the board is also interested to know whether poorer states are less likely to take up the standardized test, as compared to wealthier states.

Data Dictionary

SAT Participation Rate, ACT Participation Rate & Annual GDP by State Dataframe

Feature	Type	Description
state	object	U.S. State Name
sat_part_rate_2017	float	Participation rate of 2017 high school graduate who took SAT in their state
sat_part_rate_2018	float	Participation rate of 2018 high school graduate who took SAT in their state
sat_part_rate_2019	float	Participation rate of 2019 high school graduate who took SAT in their state
act_part_rate_2017	float	Participation rate of 2017 high school graduate who took ACT in their state
act_part_rate_2018	float	Participation rate of 2018 high school graduate who took ACT in their state
act_part_rate_2019	float	Participation rate of 2019 high school graduate who took ACT in their state
annual_gdp_2017	float	Real annual GDP of the state for 2017. Measured in millions of chained 2012 dollars
annual_gdp_2018	float	Real annual GDP of the state for 2018. Measured in millions of chained 2012 dollars
annual_gdp_2019	float	Real annual GDP of the state for 2019. Measured in millions of chained 2012 dollars
mean_sat_part_rate	float	Average Participation rate of high school graduate who took SAT in their state across 2017 to 2019
mean_act_part_rate	float	Average Participation rate of high school graduate who took ACT in their state across 2017 to 2019
mean_annual_gdp	float	Average real annual GDP of the state across 2017 to 2019. Measured in millions of chained 2012 dollars

College enrollment data by State Dataframe

Feature	Type	Description
state	object	U.S. State Name
statecode	object	U.S. State Abbreviation based on US Postal Code Standards
retention_rate_ave	float	Average rate of freshman year students by state who continued on to their sophomore year
sat_mid_score	float	Median SAT score of students accepted into college by state
act_mid_score	float	Median ACT score of students accepted into college by state

Acceptance rate for US Colleges Admission Dataframe

Feature	Type	Description
school	object	Name of College
test optional?	object	Yes: SAT/ACT are optional for acceptance into this college No: SAT/ACT are mandatory for acceptance into this college
number of applicants	int	Number of applications received by the college
accept rate	float	Rate of students who applied & got accepted into the college

Summary of Findings

Student Retention Rate

Our analysis shows that ACT scores are moderately good indicators to show how well students can cope and continue on with their sophomore year. However, SAT scores are weak indicators for the same purpose.
It is good to mention that states with a higher median SAT/ACT score tend to have a higher student retention rate in general.

Variable	Correlation with 2019 Retention Rate	Correlation with 2018 Retention Rate	Summary
SAT Score	0.33	0.23	There is a weak positive correlation with Student Retention Rate
ACT Score	0.42	0.30	There is a weak to moderate positive correlation with Student Retention Rate

Annual State GDP

Our analysis shows that the relationship between SAT/ACT participation and the general wealth of the state is relatively weak.
It is good to point out that 13 states with 100% ACT participation rate fall within the lower spectrum of annual GDP by state. This could be explained by the fact that some of these states made ACT mandatory for students to sit prior to applying for college. For instance, the state of Kentucky (source).

Variable	Correlation with Annual State GDP	Summary
SAT Participation Rate	0.24	There is a weak positive correlation with Annual State GDP
ACT Participation Rate	-0.27	There is a weak positive correlation with Annual State GDP

Impact on acceptance rate if SAT/ACT were made mandatory

From our analysis, there is no significant causal relationship on acceptance rate when it comes to SAT/ACT being made mandatory for college admission.
The factor that truly affects applicant acceptance rate will be the number of applications received by the college. The higher the number of applications received, the lower the acceptance rate.

Variable	Correlation with Acceptance rate	Summary
Number of Applicants with Mandatory SAT/ACT	-0.22	There is a weak negative correlation with acceptance rate
Number of Applicants without Mandatory SAT/ACT	-0.33	There is a weak negative correlation with acceptance rate

Conclusion

From our preliminary analysis, SAT & ACT test scores does not have the ability to forecast the likelihood of a student being able to cope with the US college education. There is no strong justification that the retention rate of students will increase due to higher scores that the students had in their SAT or ACT. To add on, the wealth of the state is not a good indicator to gauge the participation rate of SAT/ACT in the state.

Further analysis can be done to assess the magnitude of SAT & ACT scores on predicting the student retention rate. For instance, the analysis can be done at college level instead of state level average. We can also analyse the impact that state regulations have on SAT or ACT participation rate.

Recommendation

From our analysis so far, the college does not need to make SAT & ACT compulsory for college admissions. We recommend that SAT & ACT should be left as an optional admission requirement instead.

Python Libraries

Python = 3.8
Pandas
Numpy
Matplotlib
Seaborn

Credits

College enrollment data by institution level, US Department of Education. (Link).
Annual GDP by State, Bureau of Economic Analysis. (Link).

leewanxian/SAT_ACT_data_analysis