Problem Of Sample Selection Bias Research Articles

Tobit 모형을 활용한 취업 만족도 결정요인 분석

In this paper, the factors affecting job satisfaction of college graduates were analyzed. Job satisfaction consists of various factors, but job satisfaction is limited to job match, job satisfaction, and wages in this paper. Job match, job satisfaction, and wages each have information on employment status and compound information on the magnitude of job satisfaction. However, many previous studies have used only information on the size of job satisfaction to analyze it, overlooking the problem of sample selection bias. In order to overcome this problem, the Tobit model was used in the analysis. The main results are summarized as follows. First, Tobit model was statistically significant. Second, it can be seen that as male compared to female, age, and average GPA increase, job satisfaction increases significantly. Third, it can be seen that natural science compared to arts is higher, and the higher the major satisfaction and parents' income, the higher the job satisfaction. Finally, as a result of homogeneous test of the regression coefficients for each category level of independent variables, it can be seen that there is a statistically significant difference in most of the independent variables.

An Assessment of the Relevance of the Mincerian Equation: The Case of Mauritius

This paper assesses the relevance of the Mincerian Equation in Mauritius. Using CMPHS data, the basic equation is augmented to cater for wage responsiveness to changes across different schooling levels. However, the use of the standard linear regression method does not take into consideration the endogeneity of the variable years of schooling and sample selection bias. The IV2SLS Method is adopted, using region as an instrument. The Heckit model is further adopted to address the problem of sample selection bias. This study also employs pseudo panel models as an alternative to IV2SLS as a weak instrument may significantly impact results.

Causal Feature Selection in the Presence of Sample Selection Bias

Almost all existing causal feature selection methods are proposed without considering the problem of sample selection bias. However, in practice, as data-gathering process cannot be fully controlled, sample selection bias often occurs, leading to spurious correlations between features and the class variable, which seriously deteriorates the performance of those existing methods. In this article, we study the problem of causal feature selection under sample selection bias and propose a novel Progressive Causal Feature Selection (PCFS) algorithm which has three phases. First, PCFS learns the sample weights to balance the treated group and control group distributions corresponding to each feature for removing spurious correlations. Second, based on the sample weights, PCFS uses a weighted cross-entropy model to estimate the causal effect of each feature and removes some irrelevant features from the confounder set. Third, PCFS progressively repeats the first two phases to remove more irrelevant features and finally obtains a causal feature set. Using synthetic and real-world datasets, the experiments have validated the effectiveness of PCFS, in comparison with several state-of-the-art classical and causal feature selection methods.

The Influence of Prince on Sleeping Time: An Empirical Study Based on Propensity Score Matching

With the rapid development of science and technology, it is of great significance to study the influence of prince on sleeping time to explore the knowledge value of sleeping beauty. In this work, we use the data of American Physical Society to identify sleeping beauties according to B-index, and use logistic model and propensity score matching model to analyze the data. Then, we explore the impact of prince on the sleeping time of sleeping beauty. The results show that the prince has a significant inhibitory effect on the sleeping time of sleeping beauty, and the sleeping time of sleeping beauty cited by prince is shorter. The citation behavior of prince has the feature of self-selection. Using propensity score matching method can effectively overcome the problem of sample selection bias and obtain more reliable estimation results.

Imports-ICT Linkage in Colombian Manufacturing

The objective of this paper is to explore the ICT impact on firm’s import decision and on firm’s performance in import markets in Colombian manufacturing. This is the first document exploring this linkage in Latin America. We analyse firms’ imports of raw materials (not exports as the tradition) because of the decisive role imports play in firms’ exports. To deal with the problem of sample selection bias we replicate the Tobit II-Heckman procedure, by using a dynamic random effects probit model with panel data for decision process and a dynamic Generalized Linear Model (GLM) for performance process including Mills ratio. We include the influence of exports on imports, the initial conditions problem and the depreciation pattern of import experience. Our main results reveal a strong and direct impact of ICT on imports, especially for SMEs. Other results also suggest sunk costs/persistence, self-selection, cross-effects, and a decreasing import experience.

Bivariate symmetric Heckman models and their characterization

A sample selection bias problem arises when a variable of interest or response is correlated with a latent variable. This problem is presented when the response variable has part of its observations censored. The Heckman sample selection model is based on the bivariate normality assumption and fits both response and latent variables. Recently, this assumption has been relaxed to more flexible models based on the Student-t distribution, which has appealing statistical properties. In this article, we introduce an extention of the Heckman sample selection model to the wide class of symmetric distributions. In the new class of sample selection models, covariates are used to describe its dispersion and correlation parameters explaining heteroscedasticity and sample selection bias, respectively. We derive mathematical and statistical properties of the introduced model, and estimate its parameters with the maximum likelihood method. The case of the bivariate Heckman-Student-t model, as a special member of the family of symmetric Heckman models, is analyzed. Monte Carlo simulations are performed to assess the statistical behavior of the estimation method. Two real data sets are analyzed to illustrate our results.

Classification of Fermi-LAT sources with deep learning using energy and time spectra

ABSTRACT Despite the growing number of gamma-ray sources detected by the Fermi-Large Area Telescope (LAT), about one-third of the sources in each survey remains of uncertain type. We present a new deep neural network approach for the classification of unidentified or unassociated gamma-ray sources in the last release of the Fermi-LAT catalogue (4FGL-DR2) obtained with 10 yr of data. In contrast to previous work, our method directly uses the measurements of the photon energy spectrum and time series as input for the classification, instead of specific, human-crafted features. Dense neural networks, and for the first time in the context of gamma-ray source classification recurrent neural networks, are studied in depth. We focus on the separation between extragalactic sources, i.e. active galactic nuclei, and Galactic pulsars, and on the further classification of pulsars into young and millisecond pulsars. Our neural network architectures provide powerful classifiers, with a performance that is comparable to previous analyses based on human-crafted features. Our benchmark neural network predicts that of the sources of uncertain type in the 4FGL-DR2 catalogue, 1050 are active galactic nuclei and 78 are Galactic pulsars, with both classes following the expected sky distribution and the clustering in the variability–curvature plane. We investigate the problem of sample selection bias by testing our architectures against a cross-match test data set using an older catalogue, and propose a feature selection algorithm using autoencoders. Our list of high-confidence candidate sources labelled by the neural networks provides a set of targets for further multiwavelength observations addressed to identify their nature. The deep neural network architectures we develop can be easily extended to include specific features, as well as multiwavelength data on the source photon energy and time spectra coming from different instruments.

Opportunistic strategy under cooperation

Must a transaction be based on reciprocal commitments? Has opportunistic behavior inevitably caused the termination of a transaction? To elucidate the positive effect of opportunistic behavior on long-term transactions, this study regards opportunistic behavior as a strategy used to respond to the perception of partners’ opportunistic behavior. Based on 113 transactions for contract farming in Taiwan’s organic agriculture, this study employed the Heckman model to deal with the sample selection bias problem of opportunistic behavior under cooperation. This study found that in Chinese society, to maintain a transactional relationship, the opportunistic strategy is put to use based on the perception of partners’ opportunistic behavior—our perception of the partners’ obvious unethical behavior increases our hidden and obvious unethical behavior; the perception of partners’ hidden unethical behaviors increases our hidden behavior, but does not affect our obvious unethical behavior.

Foreign Direct Investment and Export Decisions of Vietnamese Enterprises

This research quantifies the impact of foreign direct investment (FDI) on the export decisions of Vietnamese enterprises. To control the problem of sample selection bias, this study employs the Heckman model (1979) estimated for two equations on export participation and rate. Unlike previous researches, which mainly rely on one single proxy of FDI, this research adopts sensititivy analysis through the estimation of a model with two representative variables for FDI. It is indicated that FDI has a positive impact on the export decisions of Vietnamese enterprises. Nonetheless, FDI has an insignificant effect on the export rate of Vietnamese enterprises.

Can Purchase Restrictions Really Limit House Prices

This article uses the data of residential commercial housing prices in 70 large and medium-sized cities from 2007 to 2017, and on the basis of controlling the covariates that can affect housing prices on the demand side and the supply side. The method of double difference is used to study the two rounds of purchase restriction policies, that is, in 2010, the first round of purchase restriction policies promulgated by local governments from 20111 and the second round of purchase restrictions promulgated in 2016 have their respective effects on residential commercial housing prices and price growth rates. The regression results of the benchmark fixed-effects model show that the implementation of the first round of purchase restriction policies in cities can significantly reduce the price of residential commercial houses, and the long-term effects of the purchase restriction policies are relatively weak. Because there may be a certain reverse causality between the implementation of the purchase restriction policy and the city’s housing prices, this paper introduces the lag value of the explained variable as the instrumental variable of the purchase restriction policy, and uses the dynamic panel instrumental variable method to perform regression analysis. In addition, there may also be the problem of sample selection bias, that is, for the difference between the treatment group and the control group, this article first uses the propensity score matching method to determine the appropriate covariate, and then matches the treatment group to the appropriate control group, And finally perform double-difference regression analysis to reduce the bias of sample selectivity. In the analysis of heterogeneity, this paper divides the samples according to the different conditions of housing prices and housing price growth rates, and verifies whether there is a difference in the impact of the purchase restriction policy on the cities with excessively high housing prices, rapid housing prices, and high and rapid housing prices. At the same time, we will return to cities in the eastern, central and western regions to explore whether there are regional differences in the effect of the purchase restriction policy. Finally, through a series of robustness tests, the robustness of the conclusions of this article is demonstrated.

Birnbaum–Saunders sample selection model

The sample selection bias problem occurs when the outcome of interest is only observed according to some selection rule, where there is a dependence structure between the outcome and the selection rule. In a pioneering work, J. Heckman proposed a sample selection model based on a bivariate normal distribution for dealing with this problem. Due to the non-robustness of the normal distribution, many alternatives have been introduced in the literature by assuming extensions of the normal distribution like the Student-t and skew-normal models. One common limitation of the existent sample selection models is that they require a transformation of the outcome of interest, which is common -valued, such as income and wage. With this, data are analyzed on a non-original scale which complicates the interpretation of the parameters. In this paper, we propose a sample selection model based on the bivariate Birnbaum–Saunders distribution, which has the same number of parameters that the classical Heckman model. Further, our associated outcome equation is -valued. We discuss estimation by maximum likelihood and present some Monte Carlo simulation studies. An empirical application to the ambulatory expenditures data from the 2001 Medical Expenditure Panel Survey is presented.

Semisupervised Two-Level Fusion-Based Autoencoded Approach for Low-Cost Domain Adaptation of Remotely Sensed Images

In this letter, a low-cost semisupervised domain adaptation technique has been proposed using a two-level fusion of artificial neural networks. The proposed technique has been aimed to solve the problem of sample selection bias using a one-time collection of a few patterns from the target domain under semisupervised framework. To assess the effectiveness, experiments are conducted on the two source–target data sets acquired over India. The results are found to be encouraging.

Adjacent Superpixel-Based Multiscale Spatial-Spectral Kernel for Hyperspectral Classification

The kernel-based spatial-spectral approach has been widely used for hyperspectral image (HSI) classification in recent years, where composite kernel (CK) and spatial-spectral kernel (SSK) are the most representative methods. Unlike CK, SSK measures the similarity of two clusters in kernel space to capture the hidden manifold in HSI, which has proven to be powerful to handle the well-known sample selection bias problem. However, the cluster feature in kernel space is always handled by a spatial filter which treats a cluster as a collection of individually independent pixels. This is neither efficient nor robust. Therefore, in this paper, we propose a method which treats each cluster as the combination of several subclusters. On the basis of the assumption that each manifold within a local region is a homeomorphic subspace in Euclidean space, we consider computing subcluster-based spatial-spectral features in the original input space and measure the similarities between subclusters to construct the proposed generalized SSK (GSSK). Moreover, due to the powerful ability to preserve the object boundaries, the adjacent superpixels (ASs) under an over-segmentation condition seem to be perfect candidates that serve as the subclusters. Therefore, we propose AS-based GSSK (ASGSSK) for HSI classification. Furthermore, the superpixel-based multiscale strategy is utilized in the framework of ASGSSK (termed ASMGSSK) to further explore the multiscale structures of HSI for improving the classification performance and sidestepping the selection of an optimal superpixel scale. Compared with the classic SSK method, the proposed ASMGSSK achieves 2.2% and 3.8% higher overall accuracy on average, with only 25% and 20% computational consumption for the Indian Pines and University of Pavia datasets, respectively. Experiments on two real HSI datasets demonstrate the superiority of the proposed method.

A Statistical Test of the Impact of Bank Lending on Corporate Productivity

Based on a large sample of industrial enterprises panel data, this paper statistical analyzes Chinese enterprises of different ownership type get bank loans, then use Heckman two-phase model to control problem of sample selection bias, empirically analyze the relationship between bank loan and the enterprise productivity, the study found that: whether state-owned enterprises or private enterprises, bank loans obtained significantly promoted the enterprise productivity. However, from the perspective of the contribution of financing to efficiency improvement, the efficiency improvement of state-owned enterprises mainly comes from bank loans, while private enterprises mainly come from their own internal funds. In addition, we also found that bank loans are typically mismatched in resources, and bank loans tend to flow to companies with poor internal funds in state-owned enterprises.

Demand for piped drinking water and a formal sewer system in Bhutan

In this study, we estimate demand for sewage connection and piped drinking water in Bhutan. To estimate household willingness to pay, we use data from a sample of 8968 households surveyed through the Bhutan Living Standard Survey of 2012. A hedonic model is estimated using Heckman two-step procedure to overcome the problem of sample selection bias. On average, households in Thimphu city are willing to pay Nu 452 (USD 7) and Nu 124 (USD 2) per month for piped drinking water connections and sewage, respectively. This translates to 8 and 2% of their monthly household expenditure for water and sewage connection, respectively. The households in Thimphu are willing to pay significantly more than the current joint charge of Nu 78 per month for water and sanitation. There is scope for municipal offices in Bhutan to increase their revenues from public services and to cover potential investment, maintenance and operational costs associated with water and sewage services.

Unsupervised Domain Adaptation Using Exemplar-SVMs with Adaptation Regularization

Domain adaptation has recently attracted attention for visual recognition. It assumes that source and target domain data are drawn from the same feature space but different margin distributions and its motivation is to utilize the source domain instances to assist in training a robust classifier for target domain tasks. Previous studies always focus on reducing the distribution mismatch across domains. However, in many real-world applications, there also exist problems of sample selection bias among instances in a domain; this would reduce the generalization performance of learners. To address this issue, we propose a novel model named Domain Adaptation Exemplar Support Vector Machines (DAESVMs) based on exemplar support vector machines (exemplar-SVMs). Our approach aims to address the problems of sample selection bias and domain adaptation simultaneously. Comparing with usual domain adaptation problems, we go a step further in slacking the assumption of i.i.d. First, we formulate the DAESVMs training classifiers with reducing Maximum Mean Discrepancy (MMD) among domains by mapping data into a latent space and preserving properties of original data, and then, we integrate classifiers to make a prediction for target domain instances. Our experiments were conducted on Office and Caltech10 datasets and verify the effectiveness of the model we proposed.

Demand for Piped Drinking Water and a Formal Sewer System in Bhutan

In this study, we estimate demand for sewage connection and piped drinking water in Bhutan. To estimate household willingness to pay, we use data from a sample of 8968 households surveyed through the Bhutan Living Standard Survey of 2012. A hedonic model is estimated using heckman two step procedure to overcome the problem of sample selection bias. On average, households in Thimphu city are willing to pay Nu 452 (USD 7) and Nu 124 (USD 2) per month for piped drinking water connections and sewage respectively. This translate to 8 percent and 2 percent of their monthly household expenditure for water and sewage connection respectively. The households in Thimphu are willing to pay significantly more than the current joint charge of Nu 78 per month for water and sanitation. There is scope for municipal offices in Bhutan to increase their revenues from public services and to cover potential investment, maintenance and operational costs associated with water and sewage services.

Does microfinance reduce poverty? Some international evidence

AbstractIn this paper, the hypothesis that microfinance is an effective tool for reducing poverty at the macro level is tested using a unique cross-country panel data set from 106 countries for the period 1998–2013. Taking into account the potential problem of sample selection bias and endogeneity, this paper shows that microfinance has a negative effect on poverty. The results are robust to the choice of microfinance measures and poverty indicators. They suggest that in developing and emerging countries, the establishment of more MFIs should be encouraged, and more funds should be directed from development agencies and governments into MFIs, to reduce poverty.

Is the Gender Pay Gap in the US Just the Result of Gender Segregation at Work?

This study examines the gender wage gap between male and female workers in the US using a cross-section from the Current Population Survey (CPS) It shows that the extent of gender segregation by both industry and occupation is significantly greater than previously supposed. For the wage gap this creates problems of sample selection bias, of non-comparability between male and female employment. To address these problems the study uses a matching approach, which we also extend to a more recent methodological version with a yet stronger statistical foundation – Inverse Probability Weighted Regression Adjustment (IPWRA) – not previously used in related studies. Despite this, doubts remain about even these well founded and appropriate techniques in the presence of such strong gender segregation . To secure even greater precision we repeat the matching analysis for a small number of industries and occupations, each carefully selected for employing similar numbers of men and women. This is an approach that has not previously been explored in the relevant literature. The findings for the full sample are replicated at the level of industry and occupation, where comparability is more reliable. The study supports the view of the existing literature that the gender wage gap varies by factors such as age and parenthood. But it also finds that, even when these and other important “control” variables such as part-time working, industry and occupation are taken into account, a statistically significant gender wage gap remains.

On the fundamental diagram for freeway traffic: A novel calibration approach for single-regime models

The speed-density or flow-density relationship has been considered as the foundation of traffic flow theory. Existing single-regime models calibrated by the least square method (LSM) could not fit the empirical data consistently well both in light-traffic/free-flow conditions and congested/jam conditions. In this paper, first, we point out that the inaccuracy of single-regime models is not caused solely by their functional forms, but also by the sample selection bias. Second, we apply a weighted least square method (WLSM) that addresses the sample selection bias problem. The calibration results for six well-known single-regime models using the WLSM fit the empirical data reasonably well both in light-traffic/free-flow conditions and congested/jam conditions. Third, we conduct a theoretical investigation that reveals the deficiency associated with the LSM is because the expected value of speed (or a function of it) is nonlinear with regard to the density (or a function of it).