• How initiatives to promote data use affect student outcomes,
• The details of initiatives to promote data use (e.g., district data systems), or
• A description of ways to use data, rather than an analytical examination of how data is used.

Research on the practice of data use, or “what actually happens when people in schools, school districts, and higher education interact with data in the course of their ongoing work in the situated context of their workplaces,” would help practitioners understand when and how to best use data for decision making.

Although there is little empirical research on data-driven decision making, there are numerous examples of how institutions of higher education integrate student learning assessment, evaluation of operations, strategic planning, and budgeting in order to maximize institutional effectiveness. Below are two examples of institutions that best demonstrate these characteristics.

Institution A is a rural community college with fewer than 10,000 students. Fifty-four percent of full-time students graduate or transfer within three years, as compared to the national average of 40 percent.

• Mission Review 
  Institution A invited all faculty and staff to partake in a yearlong review of the College’s mission and goals through 17 task groups that reviewed the core themes of the mission. The task groups developed objectives and performance indicators for each theme, and the entire college community was then invited to provide feedback on the objectives and performance indicators through a web-based Wiki.

• External Factors 
  The institution stays closely attuned to the local economy and labor market to guide its planning and prepare its students to succeed in the workforce. For example, when analysis showed that the region needed more nurses, the institution doubled its program’s capacity.

• Student Outcomes 
  One recommendation resulting from the College’s yearlong review was to report and document measurable outcomes for its A.A. and A.S. transfer degrees. In response, the College expanded and updated its learning outcomes assessment database, with faculty and adjuncts reporting student learning outcomes each quarter for at least one course. A collection of administrators and faculty meet monthly to share best practices and review the program-level learning outcomes and assessments in all departments.

• Budget Development 
  Budget development at the College is an ongoing process that includes faculty, staff, and administrators. The key factor of the process is that all budget proposals are examined within the context of outcomes assessment measures, core themes, unit and divisional strategic plans.

Institution B is an urban community college serving almost 30,000 students – sixty-four percent of whom graduate or transfer within three years.

• Integrating Planning and Resource Allocation 
  After receiving feedback to “more closely integrate the resource allocation process for faculty hiring with program review and other planning processes,” Institution B modified the academic program review process to include requests for any new faculty positions or to fill vacant positions. The College’s Academic Senate ranks all of the position requests and submits the recommendations to the Executive Vice President and Education Programs and Superintendent/President, who decide if the recommendations align with institutional priorities.

• Data and Assessments 
  The College self-identified a goal of increasing access to information necessary for planning and assessment. After much research and evaluation, software was selected to provide expanded access to the College’s data dashboards. A committee formed of staff from Institutional Research and Information Technology meet weekly to create and discuss the reporting systems.

• Student Learning Outcomes 
  Another self-identified goal was to evaluate the effectiveness of the first full year of the College’s Student Learning Outcomes (SLO) implementation program. When departments conduct program reviews, they are expected to use program-level SLO data and analysis to create course improvement plans.

• Resource Allocation 
  The College also identified its planning and resource allocation process as an area for improvement and responded by: 1) Defining the process for prioritizing and processing facilities-related requests, 2) Improving the software used for departmental planning and resource requests by clarifying areas that frequently received questions, and 3) Acting on feedback to include classified staff in the program review process.

The foundations of a data-driven decision making blueprint can be gathered by the similarities shared between Institutions A and B. Strategic planning and other operational processes of both institutions are collaborative processes that involve the entire college community providing feedback at one point or another. Additionally, all planning and self-studies begin with a review of the colleges’ mission and goals – a testament to their key role in linking strategic planning, assessment, operations, and budgeting.

The Skills Employers Want and the Skills Undergraduate Business Programs TeachSource: BizEd

On the other hand, researchers have also found some gaps in the skills graduates are bringing to their first jobs, such as a lack of soft skills and hands-on experience. And regardless of how well they are preparing their graduates today, good business schools are ever-sensitive to the need to adapt their curriculums to the changing needs of employers. To learn how business schools can best meet these challenges, Hanover Research recently examined the practices of the most innovative undergraduate business programs. One of the most important trends that emerged was a tightening integration of the elements of the business curriculum with a liberal arts foundation.

The Need for Soft Skills

In the years since the global financial crisis of 2008, employers have begun to emphasize a need for employees with both specific business know-how and broader problem-solving skills. In other words, organizations are actively seeking out workers with both technical proficiency and a range of soft skills, as indicated by the Association of American Colleges and Universities’ most recent survey of employers, which found that employers are far more likely to prefer a combination of hard and soft skills than only one or the other in isolation, as shown in the chart below.

Which is more important for recent college graduates who want to pursue advancement and long-term career success at your company?Source: Hart Research Associates

Unfortunately, a major study from the Carnegie Foundation recently found that undergraduate business education is often “too narrow, fails to challenge students to question assumptions, think creatively, or understand the place of business in larger institutional contexts.” The problem is less one of coverage – most undergraduate programs require at least some liberal arts or general education foundation – than of integrating the soft skills learned in liberal arts, the general organizational knowledge covered in fundamental business courses, and the technical skills students pick up in functional business courses.

Experimenting With the Curriculum

Some institutions are already tackling these problems. The undergraduate business curriculum at Santa Clara University is commonly cited in discussions such as these, but other colleges have taken similarly innovative approaches to weaving a curriculum together that relates business, technical, and liberal arts skills to each other. At Franklin & Marshall College, students interested in business choose a major in Business, Organizations, and Society, an integrated curriculum in which students “learn ‘how to think’ about alternative approaches to resolving issues, not simply ‘how to do’ problems,” although skill components also form an important part of the program. Similarly, Philadelphia University’s business major provides a “liberal arts-infused, professional education program” based on the University’s practice-oriented Nexus Learning model. Collectively, colleges and universities such as these have recently banded together in the Aspen Undergraduate Business Education Consortium, which is seeking to “further the integration of liberal learning and business education.”

As this movement gathers force, institutions seeking better integration of the business curriculum should foster a model that explicitly and continually links students’ learning of business skills to their use of various arts-and-sciences disciplines that provide a larger, complementary view of the world. Institutions will have to change how courses are designed and what is actually delivered in the classroom to produce a more fully integrated undergraduate business curriculum. Transitioning away from function‐based, stand-alone courses to blended learning is a challenge. Faculty will have to modify course material to incorporate the many different factors that go into real world business decisions, such as ethics, technology, diversity, and organizational behavior. A better integrated business curriculum will also require instructors, many of whom have specific areas of expertise, to bridge divides that exist in courses like accounting and organizational leadership. Despite these challenges, however, the benefits to students – and society – of producing more well-rounded business graduates seem inescapably clear.

Almost uniquely among academic fields, education hosts two kinds of doctorates: the Ph.D., and the Doctor of Education, or Ed.D. Although the Ed.D. has a distinguished lineage, originating at Harvard University in the early 20th century, angst over the legitimacy or necessity of the degree has an equally venerable history, dating from a study published in 1931, just ten years after the granting of the first Ed.D. The conventional explanation for the maintenance of two doctorates in the field of education has been that the Ed.D. is for practitioners, the Ph.D. for researchers. Thus, those attempting to tease out the difference between the two doctorates look to distinctions such as whether candidates aim to work in administrative leadership (Ed.D.) or to conduct research or teach at the university level (Ph.D.).

The actual history of the Ed.D., however, suggests that Harvard only created the degree because its Arts & Sciences faculty at the time resisted granting a Ph.D. in a professional field; although the Ed.D. appeared to be a distinct degree, it was “developed within the format of the Ph.D.,” essentially replicating many of the features of the research degree, including the three key components of graduate coursework, a qualifying exam, and a dissertation. In another telling, the Ed.D.’s founders wanted a more distinctly practice-oriented terminal degree, but were forced to include research elements, such as the dissertation, to “placate” the university’s research-focused president.

Regardless of what actually happened at Harvard 90 years ago, the Ed.D.’s role has remained confused throughout its history, as some institutions have used it to train researchers and others to prepare practitioners. And when Harvard itself recently announced that it would finally begin offering a Ph.D. in education, while phasing out its esteemed Ed.D. program, a familiar question arose anew: is the Ed.D. necessary?

Reform or Obsolescence?

At least one expert has answered this question in the negative. In the wake of Harvard’s decision, Arthur Levine, former president of Teachers College at Columbia University, expressed a desire to see the Ed.D. “fade away” altogether, possibly to be replaced by a terminal degree for educational administrators modeled on the M.B.A. Levine originally floated this idea in his extensively researched 2005 report on educational leadership for the Education Schools Project, which proposed the complete elimination of the Ed.D. for school leaders, calling the degree a “back door for weak education schools to gain doctoral granting authority” and a “meaningless and burdensome obstacle” for educators seeking to take on school leadership roles. The proposal was complemented by recommendations to introduce an M.B.A.-style Master’s of Educational Administration (or, M.E.A.) for practitioners and to reserve the Ph.D. for those intending careers in education research.

The suggestion that the Ed.D. be eliminated, however, proved to be the “most controversial element” of the Levine report, meeting resistance from education schools, school superintendents (many of whom hold the Ed.D.), and current students in Ed.D. programs. And other researchers, including Lee Shulman, former president of the Carnegie Foundation, have responded to Levine’s proposal by reaffirming that the “highest professional degree in education deserves to be a doctorate,” while granting that such a degree needs to be more clearly distinguished from the Ph.D. Towards that end, a group of researchers and education schools has been working since 2007 under the aegis of the Carnegie Project on the Education Doctorate (CPED) to develop a more distinct and effective education doctorate for practitioners.

The CPED’s principal aim is “to reclaim the education doctorate and to transform it into the degree of choice for the next generation of school and college leaders.” The initiative first brought together faculty from over 20 participating education schools, including Penn State University, the University of Southern California, and others, to discuss key concepts for the reform of the Ed.D., such as the use of capstone projects in place of the dissertation. The faculty then returned to their campuses to experiment with reform measures, before reconvening to share insights. Some of the key outcomes of the project have included the development of concepts such as the capstone course or “laboratories of practice” that allow doctoral students to develop the applied skills they need to become what the CPED calls “Scholarly Practitioners.” In 2010, the consortium won grant funding to study the effects on member institutions of the CPED’s work, an assessment that remains ongoing.

Whether the CPED’s work will be able to qualm the concerns about the Ed.D., however, remains to be seen. For one, the decision by Harvard, as a top-tier university, to eliminate its own Ed.D. has raised the question of whether other institutions might not follow suit, and either abandon the Ed.D. or convert it outright to a Ph.D. Regardless of whether the Ed.D. takes the direction of reform or obsolescence, however, one thing seems sure: the education doctorate of this century will look very different from that of the last.

Challenges Faced by Low-Income Students

A number of factors can increase a student’s likelihood of not persisting to degree completion, including working full-time while studying, having dependent children, living off-campus, entering college with inadequate academic preparation, and having insufficient resources to pay for college – and low-income students are more likely than other students to face many of these challenges.

In order to help these low-income students not only enroll in but also complete their degree programs, some institutions have targeted retention efforts specifically at helping them overcome these obstacles. In particular, these efforts have focused on providing students with academic support to help them overcome any inadequate preparation and with financial assistance to ease the burdens these students may bear outside of the classroom.

The Importance of Financial Aid

Research from Noel-Levitz and the American Institutes for Research suggests that targeted financial aid can improve the retention rates of low-income students versus students from higher income brackets with similar academic preparation. Specifically, the research found that raising financial aid to the threshold of between 55 and 60 percent of students’ financial need (cost of attendance minus expected family contribution) could significantly increase retention rates.

In practice, some institutions have already recognized the importance of targeted financial aid to improving retention among at-risk students. Georgia State University, which has a demonstrated track record as a leader in improving retention and graduation rates for its students, recognized that, for many students, “the biggest factor in graduating is funding.” Accordingly, the University has used grants of $500 per semester to assist students in danger of losing their state HOPE scholarships because of low GPAs (known as Keep HOPE Alive grants).

The Need for Academic and Social Support

Money alone, however, cannot solve the problems faced by low-income students. One scholar suggests that, to improve enrollment and retention rates, “universities need to pay attention to the practical and emotional barriers to college attendance low-income students face.” This is borne out by Georgia State’s program, which does not simply dispense money to at-risk students and send them on their way. Its Keep HOPE Alive grants are accompanied by mandatory workshops in time management, financial literacy, and academic skills to help students regain their footing.

Similarly, research from the Pell Institute has identified a number of specific interventions from the federal Student Support Services program that have increased retention rates among low-income and first-generation students, including a structured first-year experience that helps students integrate into college life; an emphasis on academic support such as developmental courses or tutoring; and “an active and intrusive approach to advising” that helps catch problems before they become irreparable and can refer students to available support resources.

Access to college for all students is an important goal. However, it is only the first step in completing the promise of higher education for low-income and at-risk students, and institutions must pay attention to their ability to retain and graduate these students, using financial, academic, or other interventions where necessary to help them achieve their goals.

How is it that community colleges, so often associated with technical and vocational training, have come to be such important providers of liberal arts degrees? In fact, a liberal arts emphasis is bound up with the origins of the community college, which originally began as an extension of the high school curriculum to a fifth and sixth year. And though two-year colleges increasingly turned to job training during the 20^th century, the liberal arts associate’s degree has remained an integral part of the community college mission for at least two reasons. First, it often provides a gateway for students to go on for a bachelor’s degree from a four-year institution. And second, the liberal arts degree itself can provide valuable vocational training, conferring a number of skills highly valued by employers.

Gateway to the Bachelor’s Degree

Community colleges offer a range of credentials, including terminal associate’s degrees and certificates in technical fields such as health care or the trades. However, they also play a crucial role as a “less expensive pathway to bachelor’s degree attainment” for students who often come from “traditionally underrepresented” groups, including minority, first-generation, non-traditional, and low-income students. Increasingly, more affluent students are turning to this option as well, with 22 percent of all college students with family income over $100,000 attending two-year colleges in 2010-2011. Regardless of their background, transfer to a four-year institution is the goal for a majority of community college students, with surveys finding that at least 50 percent and as many as 80 percent intend to go on for a bachelor’s degree. And because it covers the range of general education courses required to do upper-division coursework at a four-year institution, the liberal arts associate’s degree is commonly pitched as the vehicle for these transfers.

Liberal Arts as Vocational Preparation

In addition to preparing students for transfer to four-year institutions, the liberal arts associate’s curriculum can be just as valuable for career purposes as more technical credentials. Though oft-disparaged as impractical, liberal arts degrees confer the kind of general skills, such as oral and written communication, that transfer well between different fields, providing graduates with greater flexibility over the course of their careers. Further, some business leaders argue that companies are actually in desperate need of employees with the creativity and analytical skills that liberal arts programs cultivate. Surveys of employers have confirmed this hypothesis, with significant majorities of respondents looking for colleges and universities to place greater emphasis on skills “developed through a liberal education,” such as effective communication skills and the ability to solve complex problems. 

The Role of the Liberal Arts Associate’s Degree

Although a liberal arts curriculum cultivates valuable career skills, it is less clear what value the liberal arts associate’s has as a terminal degree, without a transfer to a bachelor’s program; high-demand careers requiring only an associate’s degree tend to be in technical fields, such as health care. But given the increasing use of community colleges as gateways to the bachelor’s, for students of all backgrounds, it seems likely that the liberal arts associate’s degree is destined to become an even more important part of the community college mission.

Reporting Structures

Offices of Student-Athlete Academic Services (SAAS) are common at Division I universities, and some Division II institutions also maintain such offices, such as the Office of Academic Support for Athletes at Kentucky State University. One key issue in operating a SAAS office, however, is who controls it – the coaches or the professors? Research from the University of Washington has identified three basic types of reporting structures among SAAS offices, including those that report solely to the athletics department; those that report solely to an academic officer (e.g., a vice-provost); and those that report jointly to athletics and academic administrators. The first two models have counter-balancing strengths and weaknesses, but the UW research found no compelling weaknesses in the joint reporting structure. Nonetheless, there was no clear consensus among the institutions surveyed, with each structure being employed by equal numbers of respondents.

Academic Support Services

Whether services are organized in a dedicated SAAS office or not, it is imperative for institutions to proactively monitor and support athletes’ academic progress. At Division II Slippery Rock University, for instance, an Athletic Academic Coordinator oversees a range of activities intended to keep athletes on track academically, including a mandatory academic orientation and college-skills workshops for first-year athletes and grade reporting, including mid-term grades, to coaches and athletic administrators. Additionally, student-athletes in any year may be required to attend a weekly Study Table (i.e., study hall), with the number of hours of required attendance pegged to the individual’s GPA (more hours are required for those with lower GPAs). In that regard, Slippery Rock’s policy is congruent with the views of one expert, who suggests that “the extent to which the use of [academic support] programs is required by student-athletes should be dependent on their at-risk status and demonstrated academic performance.” Regardless of who is required to use these services, however, colleges will want to make sure that services such as study skills training, subject-matter tutoring, and study hall are available to all student-athletes.

In an era when “institutional reputation risk related to academic integrity is at an all-time high,” colleges and universities will want to ensure that these basic structures and services are in place to support their student-athletes, regardless of whether they compete in national spotlight or in the quieter confines of sub-Division I athletics.  

A Longstanding Practice for Graduate Programs

At the graduate level, however, differential tuition is a longstanding practice. The graphs below illustrate the reality of differential tuition at the graduate level. They plot the average per year tuition charges for various types of programs among all public and private institutions, as reported to the National Center for Education Statistics for 2008-2009 (the most recent year for which NCES calculated this statistic).

Source: NCES Digest of Education Statistics 2011, Table 352

As the graphs show, at both public and private institutions first professional degrees tend to be significantly more expensive than graduate degrees (e.g., in the humanities or social sciences). Further, some first professional degrees are more expensive than others, with medicine and dentistry topping the field at both public and private institutions.

Justified for Graduate Students?

Given the concerns about differential tuition for undergraduates, particularly in its more extreme forms, it may be wondered why graduate program tuition differentials are so uncontroversial. The explanation is in the nature of graduate programs themselves, which tend to be much more “self-contained” than the undergraduate curriculum. The notion of setting “different prices for fundamentally different academic programs or types of students” seems less inequitable than charging different rates for similar programs, and graduate programs and graduate students are more easily segregated than undergraduate programs. Undergraduate programs not only share resources, such as general education course offerings, common facilities, or student services, but may also share students over the course of their careers – one college abandoned differential tuition because the policy was requiring it to collect back charges from students who switched to more expensive majors. Graduate students, on the other hand, cannot switch programs so easily, and make less use of university-wide resources.

Because graduate programs are so self-contained, their tuition can be more directly tied to the program and student characteristics, the most influential of which are two: the costs of offering the program, and expected earnings of program graduates. As two analysts from the Center for Measuring University Performance described medical education, for instance,

Medical school tuition is appreciably higher than that for other graduate [and] professional degrees for two reasons—medical education is very expensive to provide, and recipients of medical degrees are guaranteed a substantial income after graduation.

Similarly, the University of Wyoming allows tuition differentials to be charged only for graduate programs that face “unusually high costs of program delivery” and whose students are “training for relatively lucrative professions.”

Differential tuition will likely remain controversial at the undergraduate level; however, charging differential rates for graduate programs is not only well-entrenched, it makes sense. Although some general principles must be observed, such as aligning the differentials to actual program costs and to the likely future earnings of program graduates, institutions cannot – and likely will not – allow this revenue source to remain untapped.

21st century theological education providers are challenged to adapt to rapidly evolving “cultural norms, educational models, international tensions, business practices, and religious presence.” A review of literature on the subject uncovered four key methods that U.S. theological institutions are using to adapt to the religious, educational, economic, technological and social realities of the 21st century.

1) Distance Learning and Technology

Although tuition costs for online programs are often the same as residential programs, distance learning has the potential to significantly reduce the costs associated with a theological education. By taking courses online at their own convenience and through infrequent short-term residencies (ranging from several days to several weeks in length), students are able to avoid quitting their jobs and moving to a home on or near campus. For example, the Bethel Seminary InMinistry M.Div. program is designed to support rather than interfere with participants’ current ministries and communities. In fact, a growing number of new theological institutions are opening their doors entirely online, such as Rockbridge Seminary. Other institutions facing financial difficulties are closing the doors of their brick-and-mortar campuses (or regional branches) and maintaining only their online program offerings.

2) Student Debt Assistance

Students at theological institutions face rising tuition costs and steep interest rates on student loans. Many enrolled in ministry-related degree programs expect to earn a relatively low salary upon graduation, making it that much harder to repay mounting student loan debts. Some institutions, like Huntington University, are creating loan repayment assistance programs for graduates demonstrating significant financial need. Institutions can also direct students to external organizations, such as MedSend, which helps medical missionaries re-pay student loans.

3) Dual-Degree Programs and Bi-Vocational Ministry

While some graduate programs result in a certificate or degree in lay ministry, dual degree programs have the potential to attract students interested in bi-vocational or “tentmaking” ministry pathways. These programs also appeal to students who may be concerned about repaying student loan debt or being unable to find a well-paying job in full-time ministry upon graduation. A dual degree program featuring a Master of Divinity and Master of Business Administration, for example, facilitates a broad variety of career options, as well as a firm foundation in theological education. North Park University is one of many institutions that offer this opportunity to graduate students. Theology degrees can also be combined with degrees in economics or political science, such as Wesley Seminary’s dual master’s degree program in Theological Studies and International Development.

Social work is a common degree program for pairing with theological programs, but with a rising number of job openings in the field of health care, there is also significant potential for dual seminary degree programs in fields such as bioethics and nursing. Yale Divinity School and Yale School of Nursing are among several institutions to offer a joint degree resulting in an M.A. in Religion or an M.Div. as well as an M.S. in Nursing. Wake Forest University offers a joint program resulting in both an M.A. in Bioethics and an M.Div.

4) Bilingual Degree Programs

As ethnic minority and immigrant populations grow in the United States, some seminaries have developed special programs for specific groups to accommodate cultural and linguistic differences. Azusa Pacific University is one such institution, having developed bilingual graduate theological programs for both Hispanic and Korean student populations. The Programa Hispano results in either an M.A. in Pastoral Studies or an M.Div. The Asian Program offers various courses within the Graduate School of Theology using a bilingual delivery method in English and Korean. Similarly, Dallas Theological Seminary has combined distance learning and bilingual education to offer a graduate certificate to Chinese-language students, including those studying from China.

The major ethical dilemmas in international education include:

1. Funding shifts, giving rise to both escalating student tuition fees and research funding from private corporations and government sources that may involve questionable practices and outcomes
2. Political agenda for foreign student recruitment, including democratization and imperialism
3. Domestic and transnational competition for students, particularly foreign students who are able to pay higher rates of tuition and who are able to contribute to research in science and technology
4. The perspective of the student as customer, reflecting a philosophical shift in the purpose of higher education and its public mission

Using UK HE institutions as an example, the chart below demonstrates the relative growth of non-EU students from 2008/09 to 2009/10.

Growth has been strongest from Asian countries, as detailed in the chart below:

Within the EU, most international students attending UK schools come from Ireland, Germany, and France.

The Institute of International Education (IIE) pinpoints six factors likely to affect enrollment shifts of international students:

1. Expand capacity in home country higher education sectors of major sending countries 
2. Increased recruitment by other countries 
3. Domestic (political, economic, social and educational) shifts within key sending countries 
4. Transnational Education (TNE, CBE) and alternative modes of educational delivery 
5. Increased role of private education 
6. From “Brain Drain” to “Brain Circulation (the circular movement of skilled labor)”

When analyzing the determinants of success in a data-driven framework, a typical regression framework chooses variables for inclusion in a model based on a combination of the variables’ statistical significance and whether they help to maximize some overall measure of how well the model fits the data (often the adjusted R-squared value). Such models have widespread use, in both higher education and for-profit settings. Typical subjects investigated through such an analysis may include the determinants of student graduation rates, factors driving sales volumes, or key drivers of profit for a firm. Using the approach described above, it is possible to obtain a group of statistically significant predictor variables. However, there is one associated challenge: If the model yields more than one statistically significant predictor, how can they be compared? Which variable is the most impactful predictor?

Incomplete Solution 1: Look at statistical significance

Typically, multiple variables will be significant at 1% levels or lower. Moreover, statistical significance only tells us that the estimated coefficient is different from zero—nothing else.

Incomplete Solution 2: Interpret the coefficient

The problem here is two-fold. First, the coefficients may not be comparable. For example, a one unit increase in GPA and a one unit increase in LSAT scores are quite different measures, since GPA ranges from 0-4, while LSAT ranges from 120-180. We could potentially measure a 10% or a one standard deviation change in the independent variables. This approach is reasonable for continuous variables, but problems arise when interpreting categorical variables. For instance, how does one become 10%, or one standard deviation, more female?

In short, examining statistical significance and interpreting the coefficients are of limited utility when it comes to answering the central question: How can we rank multiple competing predictor variables in order of importance?

The Solution

An easy solution for examining which independent variable has the most impact is to look at each variable’s contribution towards the explained variance in a model. A numeric example demonstrates this quite easily:

For a model such as:

(eq1):  Y = β₀ + β₁X₁ + β₂ X ₂ + β₃ X ₃

If the R-squared of this model is estimated to be 0.5, then to measure X₁’s contribution to the R-squared, re-estimate the model by constraining β₁ to zero.  That is, estimate the model below:

(eq2):  Y = β₀                + β₂ X ₂ + β₃ X ₃

Let’s suppose the R-squared of this model is 0.4.  Then the share of X₁ towards the explained variance is:

=    R squared of the full model (eq1)-R squared of the constrained model (eq2)
                                                        R squared of the full model (eq2) 

=    (0.5 – 0.4)
              0.5

=    0.2 or 20%

We can do this for every variable, X₁ through X₃, and then compare this number to each other. The 20% in this example can be interpreted as the contribution of X₁ towards the explained variance in this model, or the relative importance or impact of X₁.

Example Application

Suppose a hypothetical business school wants to identify the determinants of success for their first year students, as measured by first-year GPA. Through standard regression analysis, we find that the statistically significant predictors of student success are GMAT scores, undergraduate GPA, and residential status (in-state vs. out-of-state). Assume all three of these variables are significant at the 1% level.

The estimated regression equation can be:

Predicted end of year GPA = 1.5 + 0.4*undergrad GPA + 0.01*GMAT Score + 0.25*Out of State Indicator

Model Interpretation:

• Undergraduate GPA is a statistically significant predictor of first year Business School GPA. For each point increase in GPA (for example from 2.5 to 3.5), the end of first year GPA is expected to increase by 0.4 points.
• GMAT is a statistically significant predictor of first year Business School GPA. For a 10 point increase in GMAT (for example from 600 to 610), the end of first year GPA is expected to increase by 0.1 point.
• Out of state students are expected to have an end of year GPA 0.25 points higher than in-state students.

However, since all of these are statistically significant at the 1% level, all we know is that each variable matters.  But which matters most?  Is it the undergraduate GPA, since the magnitude of the coefficient is the largest?  Is it the GMAT score, since the range for this variable is the greatest?  Using the contribution toward explained variance methodology, an impact table can be developed:

The table tells us that the GMAT score is the most important variable when determining the end-of-first-year GPA. GMAT scores contribute 30.6% toward the explained variance in the model. The second most important variable is undergraduate GPA, and both of these variables are more impactful than the student’s residential status. This information offers significant value for institutions: by determining the relative importance of individual variables, admissions officers can utilize their knowledge of these factors to admit students with the highest chance of success.