The Effect of High School Socioeconomic Status on the Predictive Validity of SAT Scores and High School Grade‐Point Average

Research has often found that, when high school grades and SAT scores are used to predict first‐year college grade‐point average (FGPA) via regression analysis, African‐American and Latino students, are, on average, predicted to earn higher FGPAs than they actually do. Under various plausible models, this phenomenon can be explained in terms of the unreliability of predictor variables. Attributing overprediction to measurement error, however, is not fully satisfactory: Might the measurement errors in the predictor variables be systematic in part, and could they be reduced? The research hypothesis in the current study was that the overprediction of Latino and African‐American performance occurs, at least in part, because these students are more likely than White students to attend high schools with fewer resources. The study provided some support for this hypothesis and showed that the prediction of college grades can be improved using information about high school socioeconomic status. An interesting peripheral finding was that grades provided by students’ high schools were stronger predictors of FGPA than were students’ self‐reported high school grades. Correlations between the two types of high school grades (computed for each of 18 colleges) ranged from .59 to .85.     

Relationships of Measurement Error and Prediction Error in Observed‐Score Regression

The focus of this paper is assessing the impact of measurement errors on the prediction error of an observed‐score regression. Measures are presented and described for decomposing the linear regression's prediction error variance into parts attributable to the true score variance and the error variances of the dependent variable and the predictor variable(s). These measures are demonstrated for regression situations reflecting a range of true score correlations and reliabilities and using one and two predictors. Simulation results also are presented which show that the measures of prediction error variance and its parts are generally well estimated for the considered ranges of true score correlations and reliabilities and for homoscedastic and heteroscedastic data. The final discussion considers how the decomposition might be useful for addressing additional questions about regression functions’ prediction error variances.     

ENHANCING RECRUITMENT SUCCESS FOR TWO-YEAR COLLEGE STUDENTS OF COLOR

The increasing proportion of students of color enrolled in secondary institutions makes the issue of their recruitment into postsecondary institutions an increasingly important concern in higher education. Data from a national survey of chief student affairs officers (CSAOs) on recruitment barriers and strategies for students of color were merged with 1995 National Center for Education Statistics Integrated Postsecondary Education Data System Fall Enrollment Survey data. A regression analysis was conducted to identify significant predictors of the percentage of students of color at 562 two-year colleges. The percentage of students of color was used as a proxy measure of aninstitution's success inrecruiting students ofcolor.Demographic and institutional characteristics that emerged as predictors included (a) having a CSAO of color and(b) being an urban institution.Recruitment strategies that entered as predictors included (a) having recruitment materials in students' native languages, (b) working with minority high schools in the design of curricula, (c) having individuals of color as members of the board of trustees, and (d) participating in dual-enrollment programs with minority high schools. The percentages of faculty members and administrators of color and the amount of contact that CSAOs have with students of color emerged as the strongest predictors. The study found that two-year colleges reap the benefit of faculty-student and faculty-teacher interactions in the form of greater institutional success in increasing their percentages of students of color. Policy makers, administrators, and faculty members can use the results of this study to promote equity by designing and implementing more successful recruitment policies and practices for two-year college students of color.     

Environmental impact prediction using remote sensing images

Environmental impact prediction is an important step in many environmental studies. A wide variety of methods have been developed in this concern. During this study, remote sensing images were used for environmental impact prediction in Robatkarim area, Iran, during the years of 2005-2007. It was assumed that environmental impact could be predicted using time series satellite imageries. Natural vegetation cover was chosen as a main environmental element and a case study. Environmental impacts of the regional development on natural vegetation of the area were investigated considering the changes occurred on the extent of natural vegetation cover and the amount of biomass. Vegetation data, land use and land cover classes （as activity factors） within several years were prepared using satellite images. The amount ofbiomass was measured by Soil-adjusted Vegetation Index （SAVI） and Normalized Difference Vegetation Index （NDVI） based on satellite images. The resulted biomass estimates were tested by the paired samples t-test method. No significant difference was observed between the average biomass of estimated and control samples at the 5% significance level. Finally, regression models were used for the environmental impacts prediction. All obtained regression models for prediction of impacts on natural vegetation cover show values over 0.9 for both correlation coefficient and R-squared. According to the resulted methodology, the prediction models of projects and plans impacts can also be developed for other environmental elements which may be derived using time series remote sensing images.     

基于用户习惯偏好相似度的Slope One推荐算法

数据稀疏是协同过滤预测精度的一个重要影响因素。Slope One算法使用简单的线性回归模型解决该问题,但它只使用评分数据做计算,未考虑相似性。提出一种基于用户习惯偏好相似度的Slope One算法（UPS Slope One）。UPS Slope One首先基于用户习惯偏好聚类,得到三组不同偏好的用户,然后分别计算各组评分偏差,计算时将用户习惯偏好相似度融入其中,最后使用线性回归模型预测评分。在MovieLens数据集上的实验表明,该算法可得到更高的推荐质量、预测准确性和稳定性。     

PREDICTION OF GRADE POINT AVERAGES IN NURSING SCHOOLS USING SECOND-ORDER MULTIPLE REGRESSION MODELS

The present study was concerned with the prediction of first year grade point averages of associate degree nursing students. The primary objective of this investigation was to determine whether the inclusion of quadratic and/or interaction terms in a regression model would improve the prediction of student nurses' grade averages. The predictor battery included cognitive, biographical, and personality variables. Results indicated that interaction and quadratic terms improve both the predictability of grade point averages and the replicability of these predictions. The inclusion of higher order terms in prediction research is suggested as a means of improving predictive efficiency.     

VALIDITY OF THE SAT FOR BLACKS AND WHITES IN THIRTEEN INTEGRATED INSTITUTIONS

Differential prediction for black and white students was empirically investigated at 13 institutions by comparison of regression planes. Particular attention was given to the possibility that prediction procedures that are appropriate for white (majority) students would underpredict the performance of black (minority) students. The data tend to support, among others, the following generalizations: (a) a single regression plane cannot be used to predict freshman GPA for both blacks and whites in 10 of the 13 institutions studied; nevertheless, (b) if prediction of GPA from SAT scores is based upon prediction equations suitable for majority students, then black students, as a group, are predicted to do about as well as (or better than) they actually do; but (c) the multiple regression (SAT-V, M) prediction for blacks in 12 of the 13 institutions was lower in magnitude than for whites and was nonsignificant in 6 of the situations studied.     

SAT Validity for Linguistic Minorities at the University of California, Santa Barbara

The validity of the SAT as an admissions criterion for Latinos and Asian Americans who are not native English speakers was examined. The analyses, based on 1997 and 1998 UCSB freshmen, focused on the effectiveness of SAT scores and high school grade-point average (HSGPA) in predicting college freshman grade-point average (FGPA). When regression equations were estimated based on all students combined, some systematic prediction errors occurred. For language minorities, using only high school grades as a predictor led to predicted FGPAs that tended to exceed actual FGPAs, particularly for Latinos. Including SAT scores in the equation notably reduced prediction bias. Further analyses showed that, while HSGPA had the highest correlation with FGPA for most groups, SAT verbal score was the strongest predictor of FGPA for language minorities in 1998. An overriding conclusion is that combining data across language groups can obscure important test validity information.     

Using Wherry's Adjusted R 2 and Mallow's Cp for Model Selection From All Possible Regressions

Selecting a subset of predictors from a pool of potential predictors continues to be a common problem encountered by applied researchers in education. Because of several limitations associated with stepwise variable selection procedures, the examination of all possible regression solutions has been recommended. The authors evaluated the use of Mallow's C_p and Wherry's adjusted R ² statistics to select a final model from a pool of model solutions. Neither the C_p nor the adjusted R ² statistic correctly identified the underlying regression model any better and was generally worse than the stepwise selection method, which itself was poor. Using any of the model selection procedures studied here resulted in biased estimates of the authentic regression coefficients and underestimation of their standard errors. The use of theory and professional judgment is recommended for the selection of variables in a prediction equation.     

How are maths-anxious students identified and what are the key predictors of maths anxiety? Insights gained from PISA results for Korean adolescents

ABSTRACT

Maths anxiety has been of great concern for many educators and educational policymakers because of its adverse effects on students’ maths performance and career path. Various empirical studies have been conducted to explore the factors predicting maths anxiety, and they have typically been based on a limited set of pre-specified variables, such as maths performance and student self-concept. However, to fully grasp the nature of maths anxiety, an exploratory study based on more elaborate prediction models using a wider variety of variables can also benefit educators. To explore the important predictors of maths anxiety and examine the possibility of achieving an acceptable level of prediction accuracy, this study employed the random forest algorithm, logistic regression, and the hierarchical general linear model to build prediction models for maths anxiety based on 194 variables collected from PISA student questionnaires. Among the factors predicting maths anxiety, enjoying maths, self-concept, and attributions to failure were revealed as being the most significant predictors. Confidence in oneself, persistent behavioural characteristics, and pressures from parents or teachers were also selected as important predictors. Educational implications are drawn from the findings of this study, and the advantages and drawbacks of each prediction model are discussed.     

Factors Affecting Achievement in the First Course in Calculus

Three groups of students at Illinois State University (of respective sizes 235, 157, and 397) were used as subjects to determine which factors were significant predictors of success in the first course in calculus. The second and third groups were used to provide replications of the initial study. Academic independent variables considered were: ACT scores, high school rank, high school GPA, high school algebra grades, and the score from an algebra pretest. Biographical independent variables considered were: sex, birth order, family size, and high school size. The dependent variable was a function of the student’s course grade in the first semester of calculus. The use of stepwise and all-subsets regression procedures on the three groups revealed in each case that the best combination of predictors consisted of the algebra pretest and high school rank. From this result, the investigators concluded that the combination of algebraic skills, as represented by the score on the algebra pretest, and long-term perseverence and competitiveness, as measured by high school rank, play a significant role in the prediction of achievement in the first semester of calculus.     

High school science enrollment of black students

How can the high school science enrollment of black students be increased? School and home counseling and classroom procedures could benefit from variables identified as predictors of science enrollment. The problem in this study was to identify a set of variables which characterize science course enrollment by black secondary students. The population consisted of a subsample of 3963 black high school seniors from The High School and Beyond 1980 Base-Year Survey. Using multiple linear regression, backward regression, and correlation analyses, the US Census regions and grades mostly As and Bs in English were found to be significant predictors of the number of science courses scheduled by black seniors.     

Comparing dropout predictors for two state-level panels using Grade 6 and Grade 8 data

The purpose of this study was to examine the effectiveness of dropout predictors across time. Two state-level high school graduation panels were selected to begin with the seventh and ninth grades but end at the same time. The first panel (seventh grade) contained 29,554 students and used sixth grade predictors. The second panel (ninth grade) included 31,641 students and used eighth grade predictors. The predictors studied were age, poverty, attendance, gender, and standardized test scores. The data were analyzed using logistic regression. All variables were predictors of dropping out of high school. Age and poverty proved to be the most effective at discriminating between dropouts and graduates within each panel. Age became more effective with time. Attendance and test scores were stable indicators between panels. Gender predicted dropouts for only the ninth grade panel. Eighth graders that were female were approximately 22% less likely to drop out.     

Comparing Student Performance and Growth in 2- and 4-Year Institutions

This study examines the organizational characteristics of 51 higher education institutions in relationship to student performance and growth. The study first finds that organizational measures of mission, size, wealth, complexity, and selectivity are statistically represented by the 2-year versus 4-year college mission. Findings indicate that 2-year and 4-year campuses indeed do exert significantly different influences on undergraduate GPA and self-reported intellectual growth. Next, the study uses both OLS regression and HLM to examine these influences. High school percentile rank and college classroom experiences are better predictors of Cum GPA at 4-year institutions, while student effort is a better predictor of GPA at 2-year institutions. Whereas the most important predictors of Cum GPA include precollege measures such as high school percentile rank and SAT score, the most influential predictors of student intellectual growth are campus experiences including classroom vitality, peer support, student effort, commitment, and involvement. Controlling for all other variables, students at 2-year institutions receive higher grades, and students at 4-year campuses experience more growth.     

The intersection of culture and institutional support for Latino students’ academic success: Remediation or empowerment?

The study is framed by critical race theory to explore the intersection of cultural and institutional factors that influence Latino students’ completion of high school. The purpose of this study is to determine the extent to which factors related to students’ background, culture, socioeconomic status, and institutional-support such as participation in mentoring and/or dropout-prevention programs, can predict Latino students’ successful completion of high school. The overarching research question is: To what extent do family background, students’ educational aspirations, and institutional support programs predict whether Latino students’ complete high school? Using data from the Education Longitudinal Study of 2002 (ELS: 2002), from the National Center for Education Statistics (NCES) with 2,217 Hispanic participants, the study used a logistic linear regression model for the analysis. The findings identified students’ gender, socioeconomic status, first language, educational aspirations as well as the aspirations of their parents, school poverty concentration, and school support programs to be significant predictors of high school completion. The logistic regression model correctly classified between 78%, 85%, and 81% of the cases included in the group for timely completion according to first-, second-, and third-generation respectively. A similar classification was found for high school completion-within-two-years. The discussion highlights marked differences between the effect of dropout-prevention programs and that of mentoring programs on Latinos’ high school completion. In addition, that the factors represented by individual and institutional variables might not operate in isolation but instead might intersect with socioeconomic and cultural factors that ultimately create barriers for this minority group.     

世界杯足球比赛多分类预测模型研究

分类问题一直是数据挖掘、模式识别等领域的重要研究内容,应用大数据技术处理与分析海量数据可实现预测分类。数据科学研究一般过于依赖LGBM和XGBoost,但在某些情况下,线性回归的效果比GBM树更好。采用机器学习中的logistics回归算法对足球比赛历史数据进行分析处理,从而挖掘数据之间的关联。通过对训练集的后视检验得到每种结果的概率,对足球比赛结果进行预测。对决策树和集成算法Adaboost建模,提高了预测准确率。该方法对预测世界杯足球比赛结果具有指导作用。     

谈外语课堂中合作性学习活动的语言问题

正面的合作学习观认为合作学习能够增加学生语言互动的机会,培养学生的合作自主学习能力和策略,但合作学习要取得实效,需要正确对待学生语言输出中的错误,促进语言向准确、流利和具有一定的复杂程度的综合体发展.错误的合作学习现担心学生在合作学习中不具备互相改错的能力而导致学习的低效率.     

Spelling,Reading Abilities and Speech Perception in Deaf Children with a Cochlear Implant

ABSTRACT

This study aims to compare word spelling outcomes for French-speaking deaf children with a cochlear implant (CI) with hearing children who matched for age, level of education and gender. A picture written naming task controlling for word frequency, word length, and phoneme-to-grapheme predictability was designed to analyze spelling productions. A generalized linear mixed model on the percentage of correct spelling revealed an effect of participant’s reading abilities, but no effect of hearing status. Word frequency and word length, but not phoneme-to-grapheme predictability, contributed to explaining the spelling variance. Deaf children with a CI made significantly less phonologically plausible errors and more phonologically unacceptable errors when compared to their hearing peers. Age at implantation and speech perception scores were related to deaf children’s errors. A good word spelling level can be achieved by deaf children with a CI, who nonetheless use less efficiently the phoneme-to-grapheme strategy than do hearing children.     

Does High School Matter? An Analysis of Three Methods of Predicting First-Year Grades

This research evaluated the usefulness of 3 approaches for predicting college grades: (a) traditional regression models, (b) high-school-effects models, and (c) hierarchical linear models. Results of an analysis of the records of 8,764 freshmen at a major research university revealed that both the high-school-effects model and the hierarchical linear model were more accurate predictors of freshman GPA than was the traditional model, particularly for lower ability students. Counter to expectations, the hierarchical linear model was not more accurate than the high school effects model.     

The determination of factors related to academic achievement in the university: Implications for the selection and counselling of students

An initial diagnosis of some educational and psychological capacities of students on arrival at university were studied. This enabled us to find out what factors had a greater influence on academic achievement at the end of the first year. Using the techniques of multiple regression we established the optimal achievement performances expected from each of the students. Secondary school marks, the academic achievement tests and the intermediate examinations at university were the best set of predictors of academic performance. Differential aptitudes of intelligence increase considerably the accuracy of the prediction. Values of R of between 0.71 to 0.88 were reached depending on the criteria used. The usefulness of the prediction equations as a tool for increasing personalized attention to students is pointed out and a case made for the establishment of objective mechanisms for admission to higher education.