Main Article
From a young age we’re told to do what makes us happy. So why shouldn’t that apply to college? The answer is it should. Of course we should go to the college that makes us happy. The only problem is, how does one go about finding the place that will be the best for four years of their life? Many start their search geographically. Do I want to stay close to home? Is there a place for me out West? Another big factor is type of school. Do I want to go to a big party school? How about a small liberal arts college? Maybe you decide to apply to liberal arts schools out West and what do you know! You got into Claremont McKenna and it’s a no brainer. You’re going. Congratulations! But what does that mean for life after college? By analyzing data concerning career salary per year by the region and type of college, we attempted to determine how these two commonly considered attributes correlate with students’ potential earnings.
We began with two datasets. The first compared school type with starting and mid-career median salary, and the second compared school region with starting and mid-career median salary. We joined these dataframes by school name, which output a table containing 268 entries with information on school type, region, starting median salary, and mid-career median salary. To this table we added a column that compares median starting salary to mid career salary for each school by giving the average increase in pay as a percentage of starting career salary. This outputs a table containing all the information we neede to begin our analysis.
| school_name | school_type | region | starting_median_salary | mid_career_median_salary | pchange_start_mid |
|---|---|---|---|---|---|
| University of Pennsylvania | Ivy League | Northeastern | 60900 | 120000 | 97.044 |
| University of Delaware | State | Southern | 45900 | 84500 | 84.096 |
| Oklahoma State University | State | Southern | 42800 | 80700 | 88.551 |
| Black Hills State University | State | Midwestern | 35300 | 43900 | 24.363 |
| University of Alabama at Huntsville (UAH) | State | Southern | 43100 | 82700 | 91.879 |
We first looked at the relationship between school type and starting median salary:
We can see right away that gradutes from Engineering and Ivy League schools have significantly higher starting median salaries. So much so that the middle 50% of engineering and ivy league school don’t even intersect with the next highest school type’s middle 50% (liberal arts). If we take a look at the regression table for this data:
| term | estimate | std.error | statistic | p.value |
|---|---|---|---|---|
| (Intercept) | 59411.111 | 1072.400 | 55.400 | 0.000 |
| school_typeIvy League | 1063.889 | 1933.296 | 0.550 | 0.583 |
| school_typeLiberal Arts | -13664.303 | 1261.143 | -10.835 | 0.000 |
| school_typeParty | -13696.111 | 1478.201 | -9.265 | 0.000 |
| school_typeState | -15284.825 | 1126.202 | -13.572 | 0.000 |
“Engineering” is used as a baseline for comparison, so it is represented by the “(Intercept)” row. We can see that the median value for starting salaries of graduates from Engineering schools rests at 59411.11. Each subsequent row represents the difference between this value and the median starting salary from that particular type of school.
Next we looked at the relationship between region and starting median salary:
Here we see that there is no region which has a significantly higher starting median salary. It is interesting to note that the large spread of the Northeastern region is likely shaped this way due to the Ivy League schools, which are only in the Northeast, raising the value of the 3rd quartile. If we take a look at the regression table for this data:
| term | estimate | std.error | statistic | p.value |
|---|---|---|---|---|
| (Intercept) | 50155.556 | 1133.384 | 44.253 | 0.000 |
| regionMidwestern | -6353.993 | 1351.475 | -4.702 | 0.000 |
| regionNortheastern | -888.413 | 1334.179 | -0.666 | 0.506 |
| regionSouthern | -5867.320 | 1339.629 | -4.380 | 0.000 |
| regionWestern | -6004.274 | 1474.405 | -4.072 | 0.000 |
“California” is used as the baseline for comparison here, so it is represented by the “(Intercept)” row. We can see that the median value for starting salaries of graduates from schools in California is 50155.56. Each subsequent row represents the difference between this value and the median starting salary from that particular region
After seeing both the correlation between region and starting median salary and school type and starting median salary we decided to look at the two of them together in a multiple regression. Below is the faceted boxplot which shows this:
Here we see that the highest median starting salary correlates with engineering schools in california. On the lower side we saw Midwestern and Western liberal arts schools along with Midwestern and Southern state schools. If we look at the regression table for this:
| term | estimate | std.error | statistic | p.value |
|---|---|---|---|---|
| (Intercept) | 73650.000 | 2874.115 | 25.625 | 0.000 |
| regionMidwestern | -17750.000 | 4064.613 | -4.367 | 0.000 |
| regionNortheastern | -13294.444 | 3177.454 | -4.184 | 0.000 |
| regionSouthern | -20983.333 | 3710.467 | -5.655 | 0.000 |
| regionWestern | -19100.000 | 4064.613 | -4.699 | 0.000 |
| school_typeIvy League | 119.444 | 1975.047 | 0.060 | 0.952 |
| school_typeLiberal Arts | -26316.667 | 3710.467 | -7.093 | 0.000 |
| school_typeParty | -23150.000 | 4978.113 | -4.650 | 0.000 |
| school_typeState | -25345.238 | 3007.866 | -8.426 | 0.000 |
| regionMidwestern:school_typeLiberal Arts | 13654.167 | 4908.486 | 2.782 | 0.006 |
| regionNortheastern:school_typeLiberal Arts | 13233.111 | 4032.875 | 3.281 | 0.001 |
| regionSouthern:school_typeLiberal Arts | 21375.000 | 4837.857 | 4.418 | 0.000 |
| regionWestern:school_typeLiberal Arts | 13123.810 | 4938.447 | 2.657 | 0.008 |
| regionMidwestern:school_typeParty | 13775.000 | 6096.919 | 2.259 | 0.025 |
| regionNortheastern:school_typeParty | 8194.444 | 5667.831 | 1.446 | 0.150 |
| regionSouthern:school_typeParty | 15401.515 | 5638.311 | 2.732 | 0.007 |
| regionWestern:school_typeParty | 16000.000 | 7040.115 | 2.273 | 0.024 |
| regionMidwestern:school_typeState | 12635.238 | 4199.787 | 3.009 | 0.003 |
| regionNortheastern:school_typeState | 9137.683 | 3397.613 | 2.689 | 0.008 |
| regionSouthern:school_typeState | 16050.571 | 3858.070 | 4.160 | 0.000 |
| regionWestern:school_typeState | 14791.790 | 4228.178 | 3.498 | 0.001 |
“California Engineering schools” are used as a base case, so for this table, the “(Intercept)” row represents the median starting salary for these schools. We can see that the median value for starting salaries of graduates from engineering schools in California is 73650. The following rows represent different types of schools in different regions and the difference between starting median salary for those schools and the baseline case.
The data we have seen so far is all based off of median salaries of graduates at the start of their careers. However, it is important to note that different careers have different rates of salary growth. Graduates working in specialized fields may start off earning upwards of fifty thousand dollars, but what if the median salary cap in their field is sixty thousand dollars? To account for this, we decided to take a look at the percent increase from median starting salary to median mid-career salary:
As we can see, starting salary is not a great indicator of mid career salary. While engineering and ivy league schools may get you a head start, salary growth is dominated by liberal arts schools. Looking at the regression for this data, we can get exact numbers on the correlation of school type and region with salary change:
| term | estimate | std.error | statistic | p.value |
|---|---|---|---|---|
| (Intercept) | 66.415 | 8.738 | 7.601 | 0.000 |
| regionMidwestern | 4.598 | 12.357 | 0.372 | 0.710 |
| regionNortheastern | 13.112 | 9.660 | 1.357 | 0.176 |
| regionSouthern | 11.350 | 11.280 | 1.006 | 0.315 |
| regionWestern | 16.375 | 12.357 | 1.325 | 0.186 |
| school_typeIvy League | 19.268 | 6.004 | 3.209 | 0.002 |
| school_typeLiberal Arts | 24.512 | 11.280 | 2.173 | 0.031 |
| school_typeParty | 21.704 | 15.134 | 1.434 | 0.153 |
| school_typeState | 16.833 | 9.144 | 1.841 | 0.067 |
| regionMidwestern:school_typeLiberal Arts | 0.403 | 14.922 | 0.027 | 0.978 |
| regionNortheastern:school_typeLiberal Arts | -6.924 | 12.260 | -0.565 | 0.573 |
| regionSouthern:school_typeLiberal Arts | 0.353 | 14.708 | 0.024 | 0.981 |
| regionWestern:school_typeLiberal Arts | -23.688 | 15.013 | -1.578 | 0.116 |
| regionMidwestern:school_typeParty | -11.537 | 18.535 | -0.622 | 0.534 |
| regionNortheastern:school_typeParty | -12.479 | 17.231 | -0.724 | 0.470 |
| regionSouthern:school_typeParty | -12.992 | 17.141 | -0.758 | 0.449 |
| regionWestern:school_typeParty | -27.068 | 21.403 | -1.265 | 0.207 |
| regionMidwestern:school_typeState | -13.955 | 12.768 | -1.093 | 0.275 |
| regionNortheastern:school_typeState | -13.015 | 10.329 | -1.260 | 0.209 |
| regionSouthern:school_typeState | -16.375 | 11.729 | -1.396 | 0.164 |
| regionWestern:school_typeState | -24.790 | 12.854 | -1.929 | 0.055 |
Our baseline case here is, once again, California Engineering schools, with subsequent rows corresponding to differences in average percent increase for each category. The correlation between region and salary growth is not as significant as that of school type; the table shows that the differences in salary growths, when school type is held constant, is similar for each region. Generally, it appears that the correlation from starting career to mid career salary is smallest among engineering and party schools, and greatest for liberal arts and (of course) Ivy League schools.
Through analysis of the regression tables and models from the data frame we created, we found that both region and school type correlates with starting median salary and mid-career salary. However, school type seems to be more strongly correlated with these metrics, especially the percentage change of median salary between the two time periods. Where engineering and party schools fell short in median salary percentage growth, liberal arts and Ivy League schools thrived. Though the region and type of school one attends factors into potential career salaries, there are many other factors that contribute to how much money someone will make. Nonetheless, this data is legitimate and shows a distinction between the region and type of school and median salaries. Ultimately, this should be used not to determine the limitations of one’s salary based on their school, but see where most people of similar education are on the socioeconomic scale at the beginning and middle of their career as well as the change between those time periods.
