**Methodology Overview**

**Data Set Characteristics:**

All data used to produce PayScale's Return on Investment (ROI) Package were collected from employees who successfully completed PayScale's employee survey.

**Bachelors Only: **Only employees who possess a Bachelor's Degree and no higher degrees are included. This means Bachelor graduates who go on to earn a Master's degree, MBA, MD, JD, PhD, or other advanced degree are not included.

For some Liberal Arts, Ivy League, and highly selective schools, graduates with degrees higher than a bachelor's degree can represent a significant fraction of all graduates.

Careers that require advanced degrees, such as law or medicine, are not included.

**U.S. Only: **All reports are for graduates of schools from the United States who work in the United States. This sample does not include U.S. territories, such as Puerto Rico or Guam.

**Full-Time Employees Only:** Only graduates who are employed full-time and paid with either an hourly wage or an annual salary are included.

Self-Employed, project-based, and contract employees are not included. For example, project-based graphic designers and architects, and nearly all small business owners and novelists, are not included.

**Selection Criteria for Schools:** To select the schools to include in the 2010 College ROI Report, PayScale started with the schools included in the 2009 College Salary Report.

Due to the more stringent data requirements to calculate accurate earnings over 30 years (graduates from 1980 through 2009), some smaller colleges included in the 2009 College Salary Report had insufficient data to be included in the 2010 College ROI Report.

In addition, schools that grew substantially or added a bachelor's degree in the last 30 years may not have had sufficient data over the full time period to be included.

Finally, schools that do not have on-campus room and board, or that did not have current data in IPEDS for tuition, room, board, fees and graduation rate, were excluded.

Of the approximately 600 schools in the 2009 College Salary Report, about 50 were not included in the 2010 College ROI Report.

PayScale plans to expand the number of schools for both the College Salary and College ROI Reports in the future. With further graduate salary data and resources, we hope eventually to report on nearly all of the approximately 2,500 bachelor's degree granting institutions in the US.

A school's inclusion in or exclusion from either the PayScale College Salary or ROI Reports is not based on school quality, typical graduate earnings, selectivity, or location in the US.

**Definitions:**

**Overall Graduation Rate:** Using the Integrated Postsecondary Education Data System (IPEDS) produced by the National Center for Education Statistics (NCES), we find the overall graduation rate according to the cohort of students entering the specific school in the fall of 2002. This overall graduation rate is the percentage of students who graduate within 6 years.

**Typical Years to Graduate:** Using IPEDS, we find the 4, 5, and 6 year graduation rates for each institution. Using these graduation rates, we determined whether an undergraduate education was typically completed in 4, 5 or 6 years according to when the majority of full-time, first-time undergraduates graduated relative to the overall graduation rate.

Total Cost for a Graduate in 2009: Using IPEDS, we find the total cost to attend the institution and graduate in 2009. Depending on the number of typical years it takes to graduate, we either calculate the cost for 4 years (2005-2009), 5 years (2004-2009), or 6 years (2003-2009) of attendance. This cost is comprised of the sum of Tuition and Fees, Room and Board, as well as Books and Supplies.

For public schools, we calculate the cost for both an in-state student and an out-of-state student. We only include the cost for those who live on campus.

Total Cash Compensation (TCC): Combines base annual salary or hourly wage, bonuses, profit sharing, tips, commissions, and other forms of cash earnings, as applicable. It does not include equity (stock) compensation, cash value of retirement benefits, or the value of other non-cash benefits (e.g. healthcare)

**Median Pay:** The median pay is the national median (50th percentile) annual total cash compensation. Half of a school's employed graduates earn more than the median, while half earn less.

**30 Year Median Pay for a 2009 Bachelor's Graduate:** Using PayScale's database, we calculate the current 30 year median pay for a bachelor's graduate of 2009 from a specific school by summing up the median pay for bachelor graduates who graduate between 1980 and 2009 from that school. We are using data over the last year so these earnings figures are in current dollars.

By using this method we are effectively taking future potential earnings and deflating them down to current dollars by wage inflation. In other words, this amount represents a present value of future earnings discounted by wage inflation.

We are effectively assuming the (wage inflation adjusted) earnings 30 years from now for a 2009 graduate is the same as the current earnings of a 1980 graduate. If the character of a school's graduates has changed substantially in the last 30 years, this measure may be inaccurate.

For example, a school that has added or substantially expanded an engineering school in the last 30 years, such that the mix of graduates today are much more likely to be engineers than 30 years ago, would tend to have a 30 year median pay that underestimates the future earnings of the typical 2009 graduate.

**34-36 Year Median Pay for a 2009 High School Graduate:** This is calculated by summing up the national median pay for high school graduates who graduate between 1974/1975/1976 and 2009. Similar to above, we are using data over the last year so these earnings are in current dollars.

Given the long term decline in earnings of high school graduates over the last 40+ years, it is likely we are overestimating the wage inflation adjusted future earnings of 2009 high school graduates in 30 years time by using current earnings of 1974/5/6 graduates. While this does not change the ranking of schools for return on investment, it may mean the dollar value of bachelor's degree from all schools is underestimated.

**Type of School:** We separate schools into three categories: Public (In-State), Public (Out-of-State), and Private. Schools are deemed either Public or Private by IPEDS, while In-State or Out-of-State refers to the tuition used to calculate the Return on Investment.

**Return on Investment (ROI) Calculations:**

**In calculating the return on investment, we must first determine the investment in college and the return from attending college.** The investment is the cost of college as determined by the actual cost of attending college. The return (gain) is the additional expected future income stream received for being a college graduate.

**Investment in College:** The investment in college is the cost of attending college, as calculated by the Total Cost for a Graduate in 2009 as defined above.

**Return from attending College:** The main financial benefit of attending college is the gain in income received by a college graduate over a high school graduate. However, by choosing to attend college, one is giving up 4-6 years of income one could have received if one went straight to work after high school. Therefore, we calculate the gain in median pay over a high school graduate (Earnings Differential) as the difference between the 30 Year Median Pay for a 2009 Bachelor's Graduate and the 34-36 Year Median Pay for a High School Graduate.

Using the Return and Investments above, we calculate two measures of ROI:

**30 Year Net Return on Investment (2010 Dollars): **To calculate the 30 year Net Return on Investment, we start with the Earnings Differential less the Total Cost for a Graduate of 2009.

Those who graduate earn an income premium over their classmates who do not graduate. We assume those who don't graduate earn slightly more than high school graduates, enough to compensate for their education costs and years spent in college, giving them a zero 30 Year Net Return on Investment.

Since not all people who attend college graduate, we then multiply the Earnings Differential less the Total Cost for a Graduate of 2009 by the overall graduation rate, giving the 30 Year Net Return on Investment.

This is an expected net return as it factors in the school's overall graduation rate. For example, a school with an 80% graduation rate, $1,000,000 Earnings Differential for graduates, and $200,000 in Total Costs would have a $640,000 30 Year Net Return on Investment.

Note: This figure is in 2010 dollars and thus represents a real return rather than a nominal return.

This ROI represents a net return on investment after the opportunity cost (High School Graduate Earnings), cost of investment (tuition, room, board, books, etc.), and probability of graduating have been taken into account.

This measure is useful for high school seniors evaluating their likely financial return from attending college.

**Annualized ROI: **This is the expected value of the Earnings Differential divided by the Total Cost for a 2009 Graduate, annualized to represent the percent of expected ROI received each year after graduation.

Once again, we calculate an expected value of the annual gain by assuming no net gain or loss for people who attend college and drop out relative to high school graduates.

For example, a school with an 80% graduation rate, $1,000,000 Earnings Differential for graduates, and $200,000 in Total Costs would have annualized return of 4.9%.

In order to compare this percentage to the rates of return on other investments (e.g., return on the S&P 500 and interest rates on bank accounts), we convert this number to a nominal rate using the Social Security Administration's National Average Wage Index, which gave wage inflation between 1979 and 2008 of 4.3%. Therefore we convert the annualized ROI to a nominal rate by multiplying the real annualized rate by this rate of wage inflation.

This means schools with graduates whose expected value of the Earnings Differential is just enough to pay back their investment in college will have an annualized return of 4.3%.

For comparison, as of June 1st 2010, 30 year U.S. Treasury bonds were yielding an annual return of 4.2%. The lowest annualized ROI of the included schools is just above 4.2%. Therefore, attending the school with the lowest annualized return is a wash compared to taking tuition money, investing it in a 30-year Treasury bond and working straight out of high school.

90% Confidence Interval on the 30 Year Median Pay: For all schools, the 90% confidence interval on the 30 year median pay is +/-5%, with the following exceptions:

- Due to large variation in the pay of graduates from elite schools (Ivy Leagues, Stanford, Caltech, etc.) the 90% confidence interval is +/-10%.
- Due to smaller data sets, the 90% confidence interval for small liberal arts schools or schools where a majority of undergraduates complete a graduate degree is also +/-10%.

To translate the impact this error has on net ROI, we need to examine how 30 year median pay is included in net ROI. For example, a 5% error on a 30-year median pay of $1.8 million is +/-$90,000. If we are dealing with a small school, or a school where a majority of undergraduates go on to a graduate program, the error increases to 10%, or +/-$180,000.

Since net ROI includes the sum of the 30 year median pay, high school earnings and cost of education, both of which have much smaller errors, the uncertainty in the 30 year net ROI is about the same as the uncertainty in the 30 year median pay. For an example school with $1.8 million 30 year median pay, $100,000 in education costs, and given $1.1 million high school graduate earnings, the net ROI would be $600,000 +/-$90,000 for most schools and +/-$180,000 for the aforementioned exceptions (90% confidence interval).

Note: The above example assumes a 100% graduation rate. If instead the graduation rate was 60%, then the net ROI becomes $360,000 +/-$54,000 for most schools and +/-$108,000 for the exceptions.