Beyond Early Warning Indicators: High School Dropout and Machine Learning

This paper combines machine learning with economic theory in order to analyse high school dropout. It provides an algorithm to predict which students are going to drop out of high school by relying only on information from 9th grade. This analysis emphasizes that using a parsimonious early warning system – as implemented in many schools – leads to poor results. It shows that schools can obtain more precise predictions by exploiting the available high‐dimensional data jointly with machine learning tools such as Support Vector Machine, Boosted Regression and Post‐LASSO. Goodness‐of‐fit criteria are selected based on the context and the underlying theoretical framework: model parameters are calibrated by taking into account the policy goal – minimizing the expected dropout rate ‐ and the school budget constraint. Finally, this study verifies the existence of heterogeneity through unsupervised machine learning by dividing students at risk of dropping out into different clusters.     

The impact of Chicago’s small high school initiative

This project examines the effects of the introduction of new small high schools on student performance in the Chicago Public School (CPS) district. Specifically, we investigate whether students attending small high schools have better graduation/enrollment rates and achievement than similar students who attend regular CPS high schools. We show that students who choose to attend a small school are more disadvantaged on average, including having prior test scores that are about 0.2 standard deviations lower than their elementary school classmates. To address the selection problem, we use an instrumental variables strategy and compare students who live in the same neighborhoods but differ in their residential proximity to a small school. In this approach, one student is more likely to sign up for a small school than another statistically identical student because the small school is located closer to the student’s house and therefore the “cost” of attending the school is lower. The distance-to-small-school variable has strong predictive power to identify who attends a small school. We find that small schools students are substantially more likely to persist in school and eventually graduate. Nonetheless, there is no positive impact on student achievement as measured by test scores.     

Do Higher Real Minimum Wages Lead to More High School Dropouts? Evidence from Maryland Across Races, 1993–2004

We explore whether higher levels of the real minimum wage have differing effects on high school dropout rates across students of various races and ethnicities (whites, African Americans, Hispanics, and Asians). Using a panel of data across Maryland counties and annual observations in 1993–2004, we found higher real minimum wages to be associated with higher dropout rates for Hispanic students, but not for other races and ethnicities. We used a variety of model specifications and explanatory variables, including real income, the unemployment rate, teen pregnancy rates, and educational attainment among adults. Several of our findings are broadly consistent with commonly reported sociological observations regarding how behavioral choices may be affected by different levels across races and ethnicities of cultural integration of recent immigrants, family cohesiveness, the value placed on education, small business ownership, and hourly (vs. salaried) employment.     

Why Do Some Schools Group by Ability?

A bstract .   This article explores the reasons that schools choose to group their students by ability. A mathematical model suggests that high diversity of achievement and a large number of students in a school will cause a school to be more likely to group. Less legitimately, some schools may choose to use ability grouping to segregate students by race or class within a school.
The data used in this study come from the 1992 NAEP 8^th grade Mathematics Assessment. Cross-tabulations suggest that ability grouping is correlated with diversity of student achievement, the number of students, racial diversity, and low levels of poverty. The mean achievement of students and school spending levels do not have much correlation with the schools' tracking decisions. Logit regressions indicate that the number of students in a school and its racial diversity are most important, and that other factors (most notably the diversity of achievement) are not. These results lend credence to those who suggest that grouping is not necessarily done for academic reasons, but rather is a subtle way to segregate by race or class.     

Dropout in secondary education: an application of a multilevel discrete-time hazard model accounting for school changes

For several decades, researchers have focused on dropout in search for an explanation and prevention of this phenomenon. However, past research is characterized by methodological shortcomings. Most of this research was conducted without considering the hierarchical structure of educational data and ignored the longitudinal path towards dropout. Moreover, research that did take into account these shortcomings, did not correct for student mobility between schools, despite the strong correlation with dropout (South et al. 2007). In this study, we attempt to address these shortcoming by implementing a multilevel discrete-time hazard model and exploring the effect of different school classifications on the school effects. Partially analogous to Grady and Beretvas (2010) we compare models with estimated school effects based on the first and on the last school attended and compare these models with multiple membership models and cross-classified models. The results of this comparison indicate that ignoring student mobility can have strong implications on the predictors of dropout. Not only do models which take into account this mobility yield better model fits, models ignoring this mobility tend to miss the effect of school level variables. With respect to the conclusions on dropout research, our models provide evidence for the often cited student characteristics predicting dropout and indicate stronger school effects than generally assumed.     

Optimizing high school curriculum assignments

A quantitative model is described for the optimal assignment of students to courses for their complete high school careers. Values of curriculum, assigned by counselor-judges for types of students, are found to be highly reliable (r_tt = 0.65). Model-generated student schedules are found to be unfailing in their conformity to all of the constraints of the test high school's curriculum. ACPP (the Automatic Curriculum Planning Program) incorporates a two-stage dynamic programming algorithm with curricular sequence constraints resolved by Boolean reduction techniques.     

Latent Class approach to analyze children’s nutritional trajectory and school dropout. A longitudinal population-based application

The study of the nutritional status is relevant during the entire life course, but in children it is relevant as malnutrition may be a marker of underlying functional and mental health deficits. Evidence of the association between malnutrition and school dropout is not conclusive. Our aim was to analyze children’s nutritional trajectory measured using their Body Mass Index (BMI) of a Uruguayan cohort and its association with school dropout. With this purpose, Latent Class and Joint Latent Class Mixed Models were fitted to children’s cohort study (N?=?1392 girls and 1492 boys) in sex-stratified analyses adjusting for sociodemographic characteristics. We identified latent classes of boys and girls with similar BMI trajectories during school years and differences in relevant socioeconomic and anthropometric characteristics. Results indicated that boys dropped out at younger ages than girls. No association between age of school dropout and nutritional trajectory classes was found. None of the classes exhibited a deficit or decrease in BMI trajectories during school ages, although the obesity and overweight classes could be of concern. Results suggested no significant association between obesity or overweight and age of school dropout for children up to 14 years old. Future research on other samples may inform about trajectories in higher educational levels.

     

Credit rationing, race, and the mortgage market

This study applies microdata from the 1983 Survey of Consumer Finances to evaluate the effects of borrower race and default risk in mortgage lending. The empirical analysis is based on a probit model of whether borrowers obtain FHA or conventional mortgages; the former are fully insured and are characterized by easier downpayment constraints, but are typically more expensive. Hence, households borrowing through the FHA will tend to be credit constrained in the conventional market. Results of the analysis indicate that variables which proxy lender concerns about default risk and cost have an important effect on the type of loan borrowers obtain. Empirical estimates also suggest that minority households are significantly less likely to obtain conventional financing than whites, even after controlling for various proxies of default risk. These results suggest that race effects in mortgage lending may persist for reasons unrelated to borrower default risk.     

Constrained mobility and the evolution of efficient outcomes

We study an evolutionary model akin to the one studied in Anwar (2002) where a set of agents use myopic best response learning to i) determine their action in a 2 × 2 coordination game and ii) to choose on which of multiple islands to interact. We focus on the case where the number of agents maximally allowed on each islands is constrained. We extend Anwar’s original analysis by considering the case when there may be more than two islands. We find that if the constraints are such that one island may be empty, universal coordination on the payoff dominant action is possible in the long run. If the constraints are such that all islands will be full, then for relatively mild constraints, and apart from special cases, the coexistence of conventions will occur, with one island coordinating on the risk dominant action and all remaining islands coordinating on the payoff dominant action. For relatively stringent constraints all agents will play the risk dominant action.     

Financing constraints and the use of performance-sensitive debt

Performance-sensitive debt (PSD) is a popular financial instrument in the corporate private debt market. In a real-options setting, this paper aims to clarify how PSD impacts on investment policy, capital structure, and agency cost of financing constraints when the firm faces the upper limit of debt issuance. We show that the constrained leverage hardly depends on the performance sensitivity. In particular, our conclusions predict that PSD can decrease the severity of financing constraints relative to the fixed-coupon debt case and the loss of firm value arising from investment and financing distortions due to the presence of financing constraints. The higher the performance sensitivity, the less likely that the firm is financially constrained. These findings provide a novel investment-based explanation for issuance of PSD.     

Does attending predominately-female schools make a difference? Labor market outcomes for women

This study explores the effects of attending predominately-female high schools on labor market outcomes. The existing literature about these schools is quite limited, and most research focuses on role-model effects at coeducational schools. Since returns to predominately-female high school attendance are likely to be upward biased due to selection, data from the National Longitudinal Survey of Youth are used to explore the determinants of such attendance. Girls who are raised Catholic, who are nonwhite, or who live in urban areas are more likely to enroll in predominately-female schools. Though women who attended these schools are no more or less likely to enter the workforce, they do earn a 19.7% higher wage than women who attended coeducational high schools. Controlling for personal characteristics as well as selection into predominately-female schools and into the workforce, the estimated wage differential falls to 12.6%.     

Cash Position, Credit Risk and Hedging

In choosing between forward and spot hedging, cash constrained and/or high credit risk firms are more likely to hedge foreign currency transactions forward than firms of greater quality. This arises because the cost of the levered component of a spot hedge is greater than the cost of the unlevered component and this premium increases with higher credit risk. For given cash and credit characteristics, importers are more (less) likely to hedge forward than exporters if transactions costs in the home security market are less (more) than the corresponding costs in the foreign security market.     

The “Bologna Process” and college enrollment decisions

We use survey data on high school graduates before and after the Italian reform of tertiary education implementing the “Bologna Process” to estimate its impact on college enrollment. Individuals leaving school after the reform have a probability of enrollment that is 15% higher compared to otherwise identical individuals. This increase is concentrated among individuals with good school performance and low parental background. We interpret this result as an indication of the existence of constraints – for good students from less advantaged households – on the optimal schooling decision. We also find a small negative impact of the reform on university drop-out.     

Capital investment and determinants of financial constraints in Estonia

Unlike previous empirical work concerning investment behavior and the determinants of liquidity constraints, we use a switching regression framework when sample separation is unknown and endogenous and firms are assumed to operate either in the financially constrained or in the financially unconstrained regime. By using new panel data for Estonian companies during 1993–2002 we find that: (i) investment behavior is characterized by two distinct regimes; (ii) the likelihood of being financially constrained is higher in firms that are recently privatized, small and where ownership is concentrated in the hands of insiders and the state; (iii) the actual probabilities of operating in the financially constrained regime are quite high and essentially stable during the whole period under consideration; (iv) ownership structure affects investment beyond its indirect effects through financial constraints.     

Barriers to skill acquisition in Brazil: Public and private school students performance in a public university entrance exam

This paper uses a novel data set to quantify the difference in performance of public and private school students in an entrance test exam of the major public university in Brazilian Northeast (Universidade Federal de Pernambuco – UFPE). Although there are many public universities in Brazil, from our knowledge, there is no study that uses data on entrance test scores at such universities to evaluate the determinants of students’ performance and the barriers for public school students to get in the good universities. The data set has detailed information on individual and school characteristics, and family background. We found that test scores of public school students are on average about 4.2–17% lower than those taken by private school students, depending on the set of controls. This result is robust when we address problems related to attrition, omitted variables (e.g., cognitive ability), and unobservable selectivity. We also show that once students get into the university, those from public schools perform as well as those from private schools. In addition, the proportion of public school students that gets into the university is roughly the same as the proportion of students doing the entrance exam. However, there is a strong barrier for public school students to get into high competitive majors. The fraction of students from public schools that gets into high competitive majors such as law, medicine, and electronic engineering is almost null. Our findings provide quantitative evidence to the common view that the Brazilian elitist high education system is an important channel for inequality persistence.     

信息不对称理论下中小企业信用担保风险分析

信用担保是国际公认的高风险行业,信用担保市场上的信息不对称是信用担保风险客观存在且居高不下的主要原因。本文从信息不对称理论的视角,分析了中小企业信用担保风险的表现形式,并提出了解决信用担保过程中信息不对称问题的对策。     

Does joining social media groups help to reduce students’ dropout within the first university year?

Using observational data, the aim of our study paper was to investigate whether university students’ dropout within the first year is influenced by participation in social media groups such as Facebook pages created and run by other students. Specifically, in this paper such participation is considered as a treatment and represents a means to help promote and strengthen social relationships amongst students but also to help share information on courses and other material useful for studying and preparing for exams. For this purpose, data from a sample survey of students enrolled in a major Italian university were used. Given a non-random treatment assignment, analysis was carried out using propensity score matching (PSM) in order to correct for selection bias due to a set of observable pretreatment covariates. Several matching techniques and sensitivity analyses suggested that the results were robust for estimating an average treatment effect on the treated group. The estimated effect indicated that participation in social media groups is effective for lowering the dropout rate.     

Do dropouts suffer from dropping out? Estimation and prediction of outcome gains in generalized selection models

In this paper we describe methods for predicting distributions of outcome gains in the framework of a latent variable selection model. We describe such procedures for Student‐t selection models and a finite mixture of Gaussian selection models. Importantly, our algorithms for fitting these models are simple to implement in practice, and also permit learning to take place about the non‐identified cross‐regime correlation parameter. Using data from High School and Beyond, we apply our methods to determine the impact of dropping out of high school on a math test score taken at the senior year of high school. Our results show that selection bias is an important feature of this data, that our beliefs about this non‐identified correlation are updated from the data, and that generalized models of selectivity offer an improvement over the ‘textbook’ Gaussian model. Further, our results indicate that on average dropping out of high school has a large negative impact on senior‐year test scores. However, for those individuals who actually drop out of high school, the act of dropping out of high school does not have a significantly negative impact on test scores. This suggests that policies aimed at keeping students in school may not be as beneficial as first thought, since those individuals who must be induced to stay in school are not the ones who benefit significantly (in terms of test scores) from staying in school. Copyright © 2004 John Wiley & Sons, Ltd.     

Does credit quality matter for homeownership?

While micro-level household data on wealth and income are available for assessing income- and wealth-based constraints to homeownership, lack of data on household credit ratings has precluded evaluation of credit quality as a potential barrier to homeownership. The study, for the first time, measures the relative importance of credit-, income-, and wealth-based constraints and estimates how the effects of these constraints have evolved over the past decade. The results show that financing constraints continue to have an important impact on potential homebuyers. The wealth constraint has the largest impact, although its importance declined substantially during the 1990s. Credit quality based constraints have become more important barriers to homeownership during the 1990s, mostly reflecting an increase in the number of households with impaired credit quality. Thus, both wealth and credit constraints persist as barriers to the attainment of homeownership.     

Credit rationing when banks are funding constrained

Credit crunches, such as in the recent financial crisis, generally occur when banks are themselves funding constrained. We use this observation to repair the workhorse Stiglitz–Weiss model of credit rationing. Recent research has invalidated the distributional assumption on which that model is based. This paper shows that by adding the assumption that banks are capacity constrained, Stiglitz–Weiss rationing can occur again. It discusses how this finding can be related to the current policy debates on bank funding and credit provision.