Introduction

My research question is: Is the cost of attending a college related to its completion rate? I chose this question because the cost of college can vary greatly between institutions, and I wanted to see whether there is a relationship between the amount students pay and the percentage of students who complete their programs.

For this project, I used data from the U.S. Department of Education’s College Scorecard. The dataset contains information about colleges and universities in the United States. For my analysis, I focused on three variables: INSTNM, which identifies the institution; COSTT4_A, which represents the average annual cost of attendance; and C150_4, which represents the completion rate for students at four-year institutions. After removing observations with missing cost or completion-rate information, my dataset contained 2,138 institutions. The College Scorecard provides publicly available information that can be used to compare characteristics and outcomes across colleges.

college_data <- read_csv("~/Downloads/Most-Recent-Cohorts-Institution.csv")
## Rows: 6273 Columns: 3308
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (2368): OPEID, OPEID6, INSTNM, CITY, STABBR, ZIP, ACCREDAGENCY, INSTURL,...
## dbl  (851): UNITID, SCH_DEG, HCM2, MAIN, NUMBRANCH, PREDDEG, HIGHDEG, CONTRO...
## lgl   (89): LOCALE2, UG, UGDS_WHITENH, UGDS_BLACKNH, UGDS_API, UGDS_AIANOLD,...
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
college_analysis <- college_data %>%
  select(INSTNM, COSTT4_A, C150_4) %>%
  filter(!is.na(COSTT4_A), !is.na(C150_4))
dim(college_analysis)
## [1] 2138    3
summary(college_analysis)
##        INSTNM        COSTT4_A         C150_4      
##  Length   :2138   Min.   : 6024   Min.   :0.0000  
##  N.unique :2123   1st Qu.:22233   1st Qu.:0.3767  
##  N.blank  :   0   Median :31067   Median :0.5209  
##  Min.nchar:   6   Mean   :37587   Mean   :0.5182  
##  Max.nchar:  75   3rd Qu.:50814   3rd Qu.:0.6667  
##                   Max.   :93512   Max.   :1.0000
average_cost <- mean(college_analysis$COSTT4_A)
average_completion <- mean(college_analysis$C150_4)

average_cost
## [1] 37587.15
average_completion
## [1] 0.5182494
college_analysis <- college_analysis %>%
  mutate(completion_percent = C150_4 * 100)
ggplot(college_analysis, aes(x = COSTT4_A, y = completion_percent)) +
  geom_point(alpha = 0.4) +
  labs(
    title = "College Cost and Completion Rate",
    x = "Average Annual Cost ($)",
    y = "Completion Rate (%)"
  )

correlation <- cor(
  college_analysis$COSTT4_A,
  college_analysis$C150_4
)

correlation
## [1] 0.5772045
college_analysis %>%
  filter(COSTT4_A == min(COSTT4_A) | COSTT4_A == max(COSTT4_A)) %>%
  select(INSTNM, COSTT4_A, C150_4)
## # A tibble: 2 × 3
##   INSTNM                            COSTT4_A C150_4
##   <chr>                                <dbl>  <dbl>
## 1 Pepperdine University                93512  0.834
## 2 Colegio Universitario de San Juan     6024  0.323
college_summary <- college_analysis %>%
  summarise(
    Number_of_Colleges = n(),
    Average_Cost = mean(COSTT4_A),
    Average_Completion_Rate = mean(completion_percent),
    Lowest_Cost = min(COSTT4_A),
    Highest_Cost = max(COSTT4_A),
    Correlation = cor(COSTT4_A, C150_4)
  )

college_summary
## # A tibble: 1 × 6
##   Number_of_Colleges Average_Cost Average_Completion_Rate Lowest_Cost
##                <int>        <dbl>                   <dbl>       <dbl>
## 1               2138       37587.                    51.8        6024
## # ℹ 2 more variables: Highest_Cost <dbl>, Correlation <dbl>

Conclusion and Future Directions

The results suggest that there is a moderate positive relationship between college cost and completion rate. The average cost of attendance among the 2,138 institutions in the cleaned dataset was $37,587.15, while the average completion rate was 51.82%. The correlation between cost and completion rate was 0.5772, which indicates that colleges with higher costs generally tended to have higher completion rates in this dataset. The scatter plot also showed an overall upward pattern. For example, Pepperdine University had a cost of $93,512 and a completion rate of 83.35%, while Colegio Universitario de San Juan had a cost of $6,024 and a completion rate of 32.26%.

These results help answer my research question by showing that college cost and completion rate are related in the data. However, the analysis does not show that higher college costs cause students to have higher completion rates. Other factors, such as student demographics, admission requirements, financial aid, school size, and academic programs, could also affect completion rates. For future research, I would like to examine some of these factors to better understand why completion rates differ between colleges. It would also be useful to compare public and private institutions or examine whether the relationship is different across states.

References

U.S. Department of Education. (2026). College Scorecard. https://collegescorecard.ed.gov/data/