Search

UAE has made significant progress in the field of education, including Higher Education, by attracting students from all around the world to various colleges and universities. Student dropout or attrition is a major issue that is faced by Higher Education Institutions (HEI). Hence, we need effective techniques to identify student data that affects attrition. This research aims to explore the use of machine learning algorithms to predict student attrition at Rochester Institute of Technology- Dubai (RIT Dubai). In this research, the CRISP-DM (Cross-Industry Standard Process for Data Mining) methodology was used as it is widely used in projects based on data mining and machine learning. Using IBM SPSS Modeler and SPSS Statistics, various machine learning models including Random Trees, Logistic Regression, Linear Support Vector Machines (LSVM), and Neural Networks were explored to find the most suitable predictive model that could accurately find student attrition and support in providing early intervention to mitigate this problem at the HEI. The logistic regression model provided the best results with the highest accuracy, AUC, precision, and recall at 1.0. The research involves analyzing student data such as demographics, results of language tests, data on academic performance, acceptance and start dates, socioeconomic factors, and other contextual data from RIT Dubai. The results of the study provide a clear understanding of the most significant predictors of student attrition that were found through machine learning. This will prove to be an important instrument for RIT Dubai to improve student retention rates.

Student retention is essential for band programs at any level to succeed and grow. This study includes results of student surveys of current high-school students who took band during their middle school years, seeking essential factors affecting student perceptions about continuation into high school band programs. Research into band student enrollment and retention exist, yet limited resources focus on transitioning from middle to high school. Studies indicate that participation in large music ensembles, such as bands, provides social, emotional, and academic benefits. Increasing student retention across this transition could yield improved student experiences and achievement in music and beyond. Existing research combined with student survey results constitutes a qualitative thematic analyzing why students who have learned much about the basics of instrumental music in middle school band elect not to continue their studies in high school. Based on existing literature, peer-connectedness, parental support, and director rapport are likely indicators for continuation in the band. By improving band student retention rates, improvements in overall graduation rates may also be achieved.

The purpose of this study was to examine whether not wellness coaching, as implemented at St. Cloud State University, has a measurable impact on participant belonging, as operationalized by the SCSU Belonging Initiative's periodic Sense of Belonging Index (SBI) surveys, which are administered at the beginning of each semester (Davis et al., 2019). Contemporary research indicates that a sense of belonging, or a lack thereof, is a key component in whether or not students continue to enroll on higher education, and is a stronger prediction of retention than GPA (Davis et al., 2019; O'Keeffe, 2013). Moreover, interventions designed to increase belonging may have an impact on reducing achievement gaps between majority and minority groups (Silver Wolf et al., 2017; Walten & Cohen, 2007). Wellness coaching is an intervention that can impact known factors of belonging (ACHA, 2020; Cemalcilar, 2010; Chiu, 2016; Clark et al., 2016; Duran et al., 2020; Krumrei-Mancuso et al., 2013; Masika, 2016; Pearson, 2012; Samura, 2016; Sforzo et al., 2017; Slaten et al., 2016; Swarbrick et al., 2011; Tachinea et al., 2017; Vaccaro, 2015). This study was designed to explore the relationship between belonging and wellness coaching by examining SBI survey scores before and after wellness coaching and determine if the process of using SBI scores was feasible as a method of evaluation. This study concluded that, while not enough data exists to determine the efficacy of wellness coaching in increasing SBI scores, this method of program evaluation is feasible and can be used again in the future as more data becomes available.

The study of retention is important to the institution, to assist in ensuring financial stability, and to the student, to provide an environment supportive of student needs for success. The purpose of this correlational study was to examine the relationship between the subscales of the Priorities Survey of Online Learners (PSOL) and the enrollment decision of adult, online students at a Christian university in the southeast United States. This quantitative, non-experimental, predictive study used a correlational research design and a survey strategy of inquiry. The participants for the study were drawn from archival data consisting of 5,221 undergraduate and 3,799 graduate students enrolled in online classes in the fall of 2009 and the fall of 2014. A binary logistic regression was used to analyze the data to examine if a statistically significant and meaningful connection existed. The findings of this study indicate that there are select educational experiences that do have an impact on retention for online adult learners. For 2009 undergraduate students, the expectation gap of their perceptions of the institution itself had a statistically significant impact on their retention; however, for 2014 undergraduates and 2009 and 2014 graduate students there was no significant predictive relationship between the expectation gap on the survey scales and student retention. The data has shown that the needs of graduate students and undergraduate students is different, and the needs of one class of undergraduate students is not necessarily the same as those of a different class. It is suggested that future research should be focused on more specific subpopulations, and it should also look at the homogeneity of the specific predictor variables in the chosen survey within the context of the institution.

Non-traditional college students enrolled in online academic studies experience higher attrition rates than traditional aged students in a campus setting. In recent years, the body of knowledge regarding academic success of not only online students has been widely examined, as has the success of students enrolled in undergraduate nursing programs. However, research that utilizes available survey instruments to determine characteristics of successful students in online RN to BSN programs is minimal to this point. Increased retention of students in online RN to BSN academic programs is critical to ensure student satisfaction, education institution revenue generation, and to address the projected future nursing shortage.Using Bean and Metzner’s non-traditional student attrition model as a framework, existing instruments were modified and utilized to capture student characteristics at the onset of their online baccalaureate nursing study in an effort to determine whether student academic success in an online RN to BSN program could be determined by similar characteristics as online students in general. Quantitative statistical analysis was completed of Likert-scale data aligning with Bean and Metzner’s non-traditional student attrition model. Additional student characteristics examined were age, gender, marital status, computer availability, educational background, hours worked weekly, current semester hours of enrollment, previous online college education enrollment, length of time since previous college enrollment, and current educational funding source. It was concluded that characteristics affecting attrition and success of online undergraduate students in general are dissimilar to characteristics affecting RN to BSN student success. However, it was determined that grade point average level of the initial baccalaureate course in question was found to be positively correlated to future enrollment in subsequent terms. Other factors specific to this academic program warrant further research, including the utilization of a compressed, five-week course format, and the fact that all students have demonstrated academic success in their prior associate degree nursing programs.

In this study, I identified institutional actions that reduce student attrition using a mixed-methods research design. The research question asked, “What multitheoretical retention levers can be identified to reduce student attrition at CVU?” A theoretical framework influenced by Braxton’s (2000) research on the complex problem of student departure guided the study. Quantitative data from a student satisfaction survey were used to develop and implement a qualitative protocol. A thematic analysis of phone interviews resulted in the generation of a number of institutional actions that are known to improve student retention. These included providing students with clear lines of communication about campus goals, values, policies, and procedures; communicating expectations related to academic policies and course requirements; and optimizing the advisor–student dynamic. The research findings demonstrated that multitheoretical retention levers can be identified using a mixed-methods design.

The purpose of this exploratory study was to determine if student background variables (age, gender, race, academic preparation); student status (part/full-time student), and student circumstances (commuting distance, independent/dependent financial student status) could predict fall-to-spring attrition (2006-2007) in a sample of 542 first-time, first-semester students at a community college in Southern Indiana. This study also included the results from an exit survey administered in the spring semester by the college. The sample consisted of 403 students who persisted to the next semester, and 121 students who did not persist to the next semester. Demographic variables, COMP ASS math and reading scores, independent/dependent student status, fulltime/ part-time student status, and driving distances were collected from the sample population. Descriptive statistics were used to identify differences between persisters and non-persisters. Hierarchical logistic regression was used to determine which study variables best predicted second semester student attrition. The exit survey analysis was used to provide anecdotal snapshots of student attrition. The students who persisted to the second semester were slightly older and had a higher percentage of males. The students who did not persist to the second semester had higher percentages of females and higher percentages of minority students and students with entrance scores requiring remedial classes before taking program-specific classes. Results of the hierarchical logistic regression indicated the COMPASS reading score variable and the full-and part-time student status variable were statistically significant predictors of attrition.

Under the George W. Bush U.S. presidential administration, the federal government pushed for greater accountability among institutions of higher education for educational outcomes. Graduation rate is a key performance indicator of institutional accountability. Previous researchers of student attrition focused primarily on the effects of student level factors on student persistence/withdrawal behavior. Recently, researchers put more focus on the effects of institutional characteristics on graduation rates, but most of these studies were exploratory and based on multiple regression models. No institutional model has existed to synthesize their results within a theoretical framework. Such an institutional model is needed to explain the process of student persistence at the institutional level. The purpose of this study was to develop a model of institutional performance in graduation rate for four-year, public and private not-for-profit, Title IV institutions in the United States. This study validated the institutional model based on the IPEDS dataset using the structural equation modeling (SEM) technique. Further group comparison analyses are conducted by fitting the same SEM model to several subgroup datasets based on grouping variables such as control, geographical region and state. Benchmarking analyses were conducted to demonstrate how administrators and policy-makers can use the institutional model to compare the performance of an institution with its peers and what policy changes can they pursue to improve graduation rates.

Concerns about student retention and attrition rates in higher education have been steadily increasing over the years. Efforts to identify and treat potential withdrawers have grown considerably; now, more than ever, universities are getting more competitive and financially driven, as the issue of student attrition is threatening to affect the way in which universities continue to run (Harvey, Drew & Smith, 2006). More universities are now developing strategic retention plans that aim at keeping students enrolled in their courses, as there is no doubt that as attrition rates increase, universities’ funding could be at risk (Schwartz, 2007). Literature on student attrition and retention in higher education shows that there has been a substantial amount of research conducted into the issues of attrition and retention in general. Varied conceptual models have been developed to explain and acknowledge student attrition as problematic and these models have focused on determining the factors contributing to student withdrawal, as well as the elements that contribute to student success. However, until now, the literature available shows that there seems to be a shortage of research into the area of foreign language student attrition and retention. As the number of students learning a foreign language in Australian high schools is at its lowest ever, this lack of research is worrying (Group of Eight, 2007). This ultimately illustrates the importance and significance of research studies conducted to determine the reasons why students are withdrawing from foreign language classes at tertiary institutions. This study investigates student attrition and develops a prognostic approach to identify students at risk of withdrawing from their courses. The study was divided into three stages. The first stage examined class assessment reports to provide a picture of existing attrition rates in the course being examined, Elementary Spanish. The second stage explored factors contributing to withdrawal, using survey and interview data collected from 12 students who withdrew from Elementary Spanish, identifying probable factors triggering withdrawal, and comparing these factors to those identified in the literature. These data were contrasted with those collected from 24 students who did not withdraw. The third stage saw the development and piloting of a prognostic instrument and approach aimed at identifying students potentially at risk of withdrawal, using theoretical insights gained from the previous two stages. Previous studies have taken a diagnostic approach to attrition—that is, looking at the problem after it has taken place. The prognostic instrument was piloted with a group of 18 first year Elementary Spanish students in 2007 to identify those at risk of withdrawal, and to evaluate the performance of the instrument. As a result of the pilot in 2007 parts of the approach were improved and in 2008 it was implemented on a group of 31 students. The results were positive in determining which students could be at risk. While the study examined Spanish language students, and thus the results may not be applicable to other languages or other disciplines, this study was the first of its kind in taking a prognostic rather than diagnostic approach to an issue of great importance to both education providers and students.

Student retention continues to be a concern for two-year and four-year colleges and universities; there are a variety of factors contributing to the institutional attrition rates. The purpose of this study was to identify and examine factors associated with how freshman students enrolled in the First Year College (FYC), at North Carolina State University during the 1999 - 2000 academic year perceive their academic and social integration. The data for this research study was obtained using the College Student Experience Questionnaire (CSEQ). The dependent variable in this research study was grade point average (GPA) of the respondents. The independent variables for this study were the academic and social factors associated with the perceptions the respondents have regarding their academic and social integration. These academic and social factors are based on the 13 activity scales on the CSEQ that measure the quality of effort known to be important to student development. They are: clubs and organizations; library; computer and information; technology; course learning; writing experiences; art, music, theater; science and quantitative experiences; experiences with faculty; campus facilities; personal experiences; student acquaintances; topic of conversation; and information in conversations. The results revealed that there is no significant difference between male and female student involvement in organizations. Both male and female students in the FYC participated in campus clubs and organizations equally. A correlation analysis revealed a positive relationship, indicating that as FYC student's GPA increases the amount of academic integration increases. The quality of effort score for Course Learning, Writing Experiences, and Scientific and Quantitative Experiences had the highest correlations significant with GPA. In reference to social integration, a correlation analysis revealed that as the student's GPA increases the amount of social integration increases. The findings suggest that the quality of effort score for Experiences with Faculty and Topics of Conversation had the highest correlations significant to GPA. A logistic regression analysis was used to predict academic and social integration from GPA. There were five independent predictor variables and one dependent variable used in the model for predicting academic success. Of the five possible independent variables used in the logistic regression model, Course Learning, from the academic integration category was the only independent variable selected in the model.

The problems with which this dissertation is concerned are undergraduate student attrition and ways to reduce it. This study describes the problem of undergraduate student attrition by presenting an overview of the extent of such attrition. As a result of this study, the following conclusions, among others, are advanced: 1) Approximately 50 per cent of those who undertake a baccalaureate degree program complete their programs within four years. 2) Attrition data provide a measure of efficiency of higher education. 3) Undergraduates who do not complete course or degree requirements in either attendance or academic performance contribute to and constitute a measure of waste in the economic sense of (a) the labor of teachers, (b) the use of classrooms, laboratories, and equipment, (c) the failure of colleges and universities to achieve educational objectives, and (d) the competent individuals who never enter college. 4) Attrition is a distinct, interactive process. 5) Institutions of higher learning are responsible for changing their characteristics and environments when such variables adversely affect congruence of otherwise qualified and competent students. 6) Texas' senior college presidents who were surveyed believe that an extremely important variable in reducing attrition is the establishment of a relationship between the student and the institution through a competent, concerned faculty or staff member or peer.

The problem with which this study is concerned is the methodology that is used to assess the relationship between student perceptions of the college environment and student attrition. The population of the study was 329 students from a metropolitan community college who took the Student Opinion Survey, a publication of the American College Testing Service. Data on course withdrawals and non-return in a subsequent long semester were collected for the student population. The data results were analyzed statistically using analyses of variance, Pearson product moment correlation, multiple regression analysis using step-wise procedures, and factor analysis. Data were considered statistically significant at the .05 level in relation to seven hypotheses on combinations of variables that include areas of student satisfaction with the college environment, student background data, course withdrawal, and non-return in a subsequent long semester.