Article 1
Mostly students are borrowing to underwrite their investments in college education. Increased borrowing is the result of both rising tuition costs and relaxed guidelines for participating in federally subsidized student loan programs. How much debt is too much is a question that has been debated by legislators and higher education policymakers. However, the literature in this direction is inad- equate to assess the degree to which students have incurred debt to pay for their studies in certain disciplines at particular types of institutions and how this debt relates to beginning salaries of these graduates. Mostly studies in the literature has studied the influence of various variables on initial earnings but no research to date has explored their systematic impact on indebtedness or the ways in which indebtedness and initial earnings are linked.
In this article, the author plan to address this gap in literature by examining both earnings and indebtedness of college graduates. The study conducted by the authors to examine mix of initial earnings, indebtedness related to college, and students choices of institution and major advances one’s understanding of the economic returns to the baccalaureate degree by first, providing more refined estimates of these returns.
In this article, the author first uses the data from Baccalaureate and Beyond (National Center for Education Statistics, 1996), a national sample of recent college graduates, to provide an updated description of the earnings distribution of full-time workers from a number of academic major areas.
Second, the author develops a multilevel earnings model incorporating demographic, family back- ground, education, and labor market variables in order to examine the impacts of each of these factors on postgraduate earnings.
Third, using the subset of students who report indebtedness as a result of the costs of their under- graduate education, the author describes indebtedness distributions as well as the distributions of a debt to earnings ratio of these graduates.
Finally, the author examines the effects of a number of individual level and institutional level variables on this debt to earnings ratio.
The author plans to test the effects of individual-level (graduates) and college level predictors on earnings and debt to earnings ratios. The null hypothesis is that these variables are not significant to predict the earnings and debt to earning ratios.
3. Variables:
Therefore, individual and college level variables are independent variables and earnings and debt
to earning ratios are dependent variables.
Four types of individual-level variables are used:
• demographic: consisted of dummy variables capturing graduates gender (FEMALE) and minority status

1. What is the null hypothesis when using ANOVA a. All the varianc...