# I. Introduction ince its independence back in 1971, there has been an upsurge in higher education in Bangladesh. Before the independence, only four public universities were available for tertiary education in this country. That figure is now 42 along with around hundred other private universities approved by the University Grant Commission (UGC) 1 . These numbers are in no mood to stop in near future. There are debates going on whether such large numbers of institutions for higher education, particularly the ever-increasing number of private universities are still needed in a developing country like Bangladesh. On the other hand, the quality in higher education is continuously moving in a declining trend. There are several issues responsible for such negative trend. But in this study, focus has been given to a specific issue regarding higher study, academic performance of the students who are admitted at undergraduate level in private universities. Several factors are associated to the academic performance of the undergraduate students such as enthusiasm and individual characteristics of the students, demographic and socio-economic characteristics, institutional characteristics and so on 2 . But only a little research has been done to identify the significant factors, especially in case of Bangladesh. As long as academic performance is concerned, Cumulative Grade Point Average (CGPA) is considered as the indicator by most of the researchers around the world 3 . The family background, educational environment, and financial states of the learners are crucial factors affecting academic performance 4,5 . Involvement of the students in part-time jobs to bear partial educational expenses also affect their academic performance 6,7,8 . Motivation and strong personality direct the students in the way towards the achievement of their academic goal 9 . Academic performance at pre-university level (S.S.C. and H.S.C. in case of Bangladesh) was found to have significant association with its counterpart at university level 10 . The objective of this study was to identify the factors that have significant impact on the academic performance of the undergraduate students of private universities in Bangladesh. The extent of association would also be tested through odds ratio applying logistic regression analysis. # II. Methodology Primary data for the study were collected from undergraduate students of Fall semester 2018 of International University of Business Agriculture and Technology (IUBAT). A pre-coded self-directed questionnaire was initially developed for the data collection purpose. It was then finalized after being pretested on a small sample of 30 students of BBA (Bachelor of Business Administration) program. There are approximately 10000 students in total under different colleges in IUBAT. An estimated sample size of 605 respondents was interviewed for the study considering 4% margin of error (3.86% to be exact) at 95% confidence interval 11 Conceptual framework coefficients represents odd ratios (OR) for the given variable. The analysis was performed using SPSS version 20.0. The dependent variable was dichotomous given value 1 if the delivery was attended by trained medical personnel and given value 0 otherwise. The model was fitted as: E(y) = ?? ???? 1+?? ???? , ? = [? 0 , ? 1 , ? 2 , ??., ? k ], X = [x 1 , x 2 , x 3 , ??., x k ] / here, y = dependent variable = academic performance, x i = corresponding independent variables, ? i = parameters of the model. # III. Results and Discussion The background characteristics of the respondents are shown in Table 1. Out of 605 students, more than half (52.2%) were from Engineering departments. Academic performance of almost twothird of them (63.5%) was at satisfactory level. As far as family educational background is concerned, less than half (43.8%) of the family had highly educated one. More than three-fourth of the students (78.5%) performed well in S.S.C. and H.S.C. examinations which are the two pubic exams that take place at national level before they sit for undergraduate admission at university level. The number of male students was 1.78 times more than its Volume XVIII Issue XI Version I Several in dependent variables were considered that can be classified as: socio-economic variables such as location of residence, religion, yearly family income, students' part-time job status, students' physical health status, and educational background of family; demographic variables such as age and gender; institutional variables such as department, teacherstudent relationship, class size and course load; individualistic variables such as SSC and HSC result, regularity, and punctuality, self-confidence and depression. Other than the department, all other variables were dichotomous. The dependent variable for the study was academic performance which is also dichotomous, where CGPA ? 3.00 was coded as 1 and CGPA < 3.00 was coded as 0.The association between academic performance and all the above-mentioned independent variables were checked by applying chisquare test, where p-value ? 0.05 indicated significant association. Finally, logistic regression model was fitted among academic performance and the independent variables that showed significant association with it. Logistic regression is the widely applied regression analysis when the dependent variable is dichotomous. No assumption regarding the distributions of the independent variables is needed. It delivers an estimated value for the strength of the association adjusting for other variables. The exponential of the female counterpart indicating the supremacy of male students at tertiary level education. A large proportion of the students (70.2%) belonged to the age group of above 20 years. Most of them (59.7%) came from urban areas which was expected since education in private universities is an expensive deal and is out of reach for most of the rural people. One of the positive aspects of the respondents was their class attendance. Around 93.7% of them attended their respective classes regularly. But the scenario was entirely the opposite one in case of library attendance. Only 19% of the respondents had regular attendance in library. One possible reason might be the increasing access of internet via smart phones where study materials are readily available in various formats in many cases, discouraging the students from spending some quality time in the library to exercise their brain. The teacherstudent relationship was found to be friendly in most of the cases (87.1%). The self-confidence level of 70.2% students was found to be high. There was nearly identical number of respondents as far as the depression status of the students is concerned. Most of the students were Muslim (89.9%) which was expected since a huge portion of the population of the country is Muslim. Relatively much higher proportion of students (85.3%) were involved with part-time job. The logistic regression results are shown in Table 3. Other than attendance, teacher-student relationship and depression, the remaining variables were still showing significant association with academic performance. # Table 3: Logistic Regression Results # C.I. = Confidence Interval, OR = Odds Ratio, RC = Reference Category One notable part of the findings of logistic regression model was that academic performance of female students were 1.49 times better than their male counterparts. Students suffering from depression were less like to perform better than those without depression. The relationship pattern with the teachers also seemed to impact the academic performance of the students. As expected, irregular students were found to be less likely (OR = 0.60) to perform better than the regular ones. Students undertaking not more than nine credits were 1.77 times more likely to perform better than those undertaking above nine credits. This was also usual that the less the amount of credits, the less the burden of study materials for the students; consequently, more time is available to focus on less amount of study materials. Self-confidence level of the students seemed to have highly significant association (p-value = 0.000) with the academic performance of the students. Achievements of the students with low level of confidence were 0.44 times less likely to perform better. The findings of this study will have crucial strategic implications to the managements of private universities in Bangladesh. Not only the students but also their families will realize which factors would have significant impact on students' academic success. The study will also help to undertake proper actions by Institutional Quality Assurance Cell(IQAC) in the universities as well. The number of private universities will continue its ever-increasing trend in the upcoming days. Therefore, it is high time to take necessary steps to ensure the quality of education at this level. It is the huge amount of tuition feesbeing paid by the students in the private universities that ultimately runs the life of the institutions. Hence, their academic performance level must be taken care of seriously and sincerely. # IV. Conclusion 1FrequencyPercentage 2 2Year 2018Academic Performance Satisfactory Not SatisfactoryTotalSignificanceBusiness12157178Engineering209106315?? 2 = 14.51DepartmentOthers5458112p-value = 0.001Total384221605Education of FamilyHighly Educated Moderately Educated Total171 213 38494 127 221265 340 605?? 2 = 0.22 p-value = 0.634SSC & HSC ResultGood Not Good Total316 68 384159 62 221475 130 605?? 2 = 8.90 p-value = 0.003( G ) Global Journal of Human Social Science -Family Income Gender Age Residence Class Attendance Library Attendance Teacher-Student RelationshipGood Not Good Total Male Female Total At most 20 Above 20 Total Urban Rural Total Regular Irregular Total Regular Irregular Total Friendly Unfriendly Total162 222 384 236 148 384 115 269 384 228 156 384 368 16 384 77 307 384 343 41 38489 132 221 151 70 221 65 156 221 133 88 221 199 22 221 38 183 221 184 37 221251 354 605 387 218 605 180 425 605 361 244 605 567 38 605 115 490 605 527 78 605?? 2 = 0.21 p-value = 0.645 ?? 2 = 2.87 p-value = 0.05 ?? 2 = 0.20 p-value = 0.483 ?? 2 = 0.04 p-value = 0.846 ?? 2 = 7.983 p-value = 0.005 ?? 2 = 0.74 p-value = 0.388 ?? 2 = 4.59 p-value = 0.032Self-confidenceHigh Low Total298 86 384127 94 221425 180 605?? 2 = 27.22 p-value = 0.000Physical HealthGood Not Good Total311 73 384172 49 221483 122 605?? 2 = 0.995 p-value = 0.318DepressionNo Yes Total209 175 384102 119 221311 294 605?? 2 = 3.843 p-value = 0.050© 2018 Global Journals * Official Website Of Wikipedia October 2018 6 * The Determinants of Undergraduate Grade Point Average: The Relative Importance of Family Background, High School Resources, and Peer Group Effects Betts The Journal of Human Resources 34 1999 * Factors affecting business students' performance in Arab open university: Case of Kuwait AAl-Mutairi International Journal of Business and Management 6 5 2010 * The impact of students' socioeconomic background on academic performance inuniversities, a case of students in Kisii University College CKOkioga American International Journal of Social Science 2 2 2013 * Determinants of Degree Performance in UK Universities: A Statistical Analysis of the 1993 Student Cohort Smith Oxford Bulletin of Economics & Statistics 63 1 2001 * Preventing students from dropping out AlexanderWAstin 1975 Jossey-Bass San Francisco 1st ed) * The impact of paid work on the academic performance of students: a case study from the University of Canberra Applegate Australian Journal of Education 50 2 2006 * University students, Experience Work and academic performance in Spain Benito 2014 * Personal, family, and academic factors affecting low achievement in secondary school LADiaz Electronic Journal of Research in Educational Psychology 1 1 2003 * The Influence of Type of High School Attendedon University Performance Birch Australian Economic Papers 46 2007 * References Références Referencias Volume XVIII Issue XI Version I