Factor Analysis for Women in Engineering Education Program to Increase the Retention Rate of Female Engineering Students

Factor Analysis for Women in Engineering Education Program to Increase the Retention Rate of Female Engineering Students Jong Tae Youn Department of Graphic Arts Information Engineering, Pukyong National University, Busan, South Korea Song Ah Choi Department of Graphic Arts Information Engineering, Pukyong National University, Busan, South Korea Abstract The lack of women students in college of engineering is a serious problem as a lack of human resources in industrial fields of South Korea. The purpose of this study was to develop an effective measurement tool to measure the women students mind of dropout from engineering and to increase the retention rate of women students in college of engineering in South Korea. In order to develop an effective measurement tool, factor analysis for female engineering students at P University in South Korea was conducted. The analytical results showed that social and cultural effects are more significant ones than social effect alone, and the effects of gender discrimination in industrial fields or college to the women engineers are more significant than effects of regional or positional ones. Also results showed responses that the physical differences from man students, such as the ability to drive machines or to handle tools, their parents financial supports, the gender cognitive engineering education and the boosting for job recruiting are more significant effects than the scholastic ability. In this research, we developed the measurement tool to measure women students competencies in college of engineering and contributed to make the women in engineering education according to the results of factor analysis. These researches have been progressed since 2006 annually, and we had proposed to programs for the gender cognitive engineering education and the special programs for women engineering students. As the results, women engineering students dropout rate in college of engineering in P University as a model was about 5 % at 2006, it was much higher than 1 % of male students, however the dropout rate of women engineering students were decreased to 1% at 2015 due to women in engineering program and proposed programs based on the results of factor analysis of women engineering student s dropout from college of engineering. Key Words: Female, Women, Students, Retention, Korea Dropouts, Engineering Education Introduction In South Korea about 10% of women engineers are engaged in industries at 2014. The lack of national women power resources is a serious problem. Moreover women engineering students still have wanted to change their major because then are faced with a glass wall or glass ceiling of male dominant culture, circumstances that preferring males to females. Many of women engineering students had withdraw from engineering school temporarily, and about 5% of women engineering students changed their major from engineering to some other major or dropped out of completely in P University until 2006 (Youn 2011). So, one of the governments important campaigns of South Korea is the promoting and supporting women engineers in the industrial field. The most important issue of the campaign is the clear obstacles that hinder women s promotion to higher positions in industries and barriers of their continuing careers and supporting women in engineering and science. An increment of female engineers, human is an important national undertaking not only for the national economics but also for improving the quality of women s life and social status. In 2010, the OECD prospected that the rate of potential growth of the South Korean economy for 2010~2011 would be about 4 %, and they also prospected that the rate of growth would decrease as low as 2 % per year until 2025. Also, they predicted that the tendency of students avoiding engineering related jobs would bring about a lack of human resource power in production and in research and development, and eventually weaken the South Korean economy. To find a way out of these problems, the Korean government made a law of provided fostering and supporting women engineers and scientists (Park, 2008). These situations are almost same in some other nations. Female students are not interested in engineering related activities and consistently experienced much more difficulties than male students in engineering and science (Wolffram, 2009). Discrimination of women engineers in the industrial field is still real worldwide problem. Even though we have made improvements against, it is still devastating throughout the world. Women students can be discriminated and be experienced the prejudice not only at college, but at all sociological and cultural levels at home or in their society, either directly or indirectly. South Korea stands with circumstances of the worst gender pay gap and the worst glass ceiling index among the OECD nations. Therefore, it is necessary to understand how women engineering students thoughts differ from those of male 5657

students and why more female students than male students dropped out of college. Because of complex concepts such as sociological status, college educational systems, individual or personal problem, or individual scales corresponding to the female students dropout phenomena, the factor analysis for investigating variable relationships was used. To find the factors that contributed to the female students mind to dropout, we conducted both in-depth interviews and a questionnaire survey. The questionnaire was framed on the basis of the women students thoughts and the experiences in their campus life. In college of Engineering in South Korea, women engineering students dropout rate was 4.4 % and it was much higher than 1 % of male students. And according to the results of preliminary experiments, we found that over 50 % of the women students wanted to change their major from engineering. One of the most important growth engineering industries is manpower as well as womanpower (human resources). However South Korea had a problem that it has a small numbers of women engineering students, 18 % of undergraduate, 10 % of graduate and it only has 2 % of women faculties in college of Engineering. To understand the situation of college in South Korea, P University was selected for this research. Women students in college of engineering in P University constitute 23% of the student in 2014 and the worst dropout rate until 2006. Also P University is a leading university of women in engineering program. Similar researches have been progressed since 2006, and we have presented at the conference of quality in higher education (Youn, 2015). And we need to make an advanced research to the framework of factor analysis and to perform annually to conduct a follow up survey. Theoretical Framework Definition of Dropout Rates The factors of dropout vary according to how the concept is defined. When the definition of dropout and the manner in which it is defined are not consistent, comparisons are difficult to make, and when comparisons are made, analysis may be faulty. The areas contributing to definitional confusion include variation in grade levels or ages and variation in the length of time or period during which dropout in calculated. There have been many attempts to identify the dropout (Morrow 1986). It is classified dropouts by Morrow into five types, such as the push out of school after being judged as disqualified, the unaffiliated persons who refused to associate with others, the voluntarily dropout who leave the school as the college and the family demanded different scales from the students, the educational mortalities who are not able to complete the coursework, and the stop out students who will return to school. However in this research for the special purpose to increase the retention rate of women engineering students, we focused and limited the dropout refers to the willing of discontinuing coursework at a regular college of engineering studies. In other words, for a quantitative factor analysis we calculated the dropout as the women students who initially entered in the engineering departments but quit the school or transfer to some other major field instead of engineering. Factors of Women Engineering Students Dropout To understand better the factors behind students decisions from dropping out, we reviewed the research on the college student s dropouts. Although in any special research it is difficult to adopt directly to casual relationship between single factor and mind to dropout, a large number of reports with similar findings do indicate a connection. The factors that predict whether students dropout or change their major from engineering fall into four categories: (1) individual factors, (2) domestic factors (3) sociocultural factors (4) college educational factors. Table 1 shows a summary of reorganized factors form previous researcher on the phenomenon of college students dropping out. Indeed the tendency of students dropout is combined and connected effects of multiple factors (Cho, 2007). Boshier reported that the phenomena of dropping out increase when the internal harmony with one s self does not match the harmony of one s self with the environment (Boshier, 1973). More macroscopic analysis offered nine factors for dropping out such as the idealistic self, the trust in one s self, the adequacy of the process, the level of achievement, available support for individual students, control by the educational institution, mismatch of the ego with the super ego, the clarity of objectives and academic competence (Garrison, 1987). Table 1: A summary of the factors of dropout phenomenon Factors Examples of variables Individual Factor Self confidence, behaviors and attitudes Educational performance capability Academic ability Affiliation with Classmates and faculty members Marriage Age limit Domestic Factor Housework circumstances Sociocultural Regional characteristics, Regional disparity Factor Educational Major program Factor Matching between the contents of education and student s goal Difficulties of coursework Satisfaction of the educational environment Satisfaction of the curriculum Source: Reorganized the Bae s report (Bae Kyung Suk, 2004). Several aspects of above four factors have been widely identified in literature as predictors of dropping out, and mismatch with her major program and the circumstances are the important variables (Park, 2009). Also discontinuation of coursework in college life or in her major program is a significant variable (Oh, 2006).. 5658

Research Method Research Procedures The main objective of this study is to identify the reason why women students dropout and leave the college of engineering. Factor analysis of women engineering students dropouts from an engineering college was attempted to identify the selected factors or variables that explain the pattern of correlations within a set of observed variables. Factor analysis is often used in data reduction to identify a small number of factors that explain the variance observed in a much larger number of manifest variables. By using factor analysis we could generate hypotheses regarding casual tendency of dropout or screen variables for subsequent analysis. Source data were obtained by conducting both indepth interviews and a questionnaire survey. A quantitative measurement tool was developed on the basis of the factors that reflected the characteristic features of women engineering students dropping out. These were extracted from the interviews. The validity and reliability of the measurement tool were confirmed by applying SPSS 21 software. In descriptive statistical analysis, standard deviations were below 0.1% whose Skewedness and Kurtosis were above ±2.0. For the reliability test, we used Cronbach s alpha coefficient. Cronbach s alpha value of 0.5 0.6 is considered as relatively good and alpha value above 0.7 is good enough to be accepted (Nunnally, 1976; Lee, 1991). In this research, we used Varimax of the right angle rotation method for the rotation method. The number of factors was determined by using the Scree plot, eigenvalue, gross dispersion rate of more than 60%, and interpretation potentials, and the validity was confirmed through the eigenvalue, the common dispersion versus gross dispersion rate, and the interpretations of the content. Data collection In-depth Interview To collect the data, we interviewed random 20 women engineering students who changed their major from engineering and another 20 students attended in college of engineering at P University. The depth-interview is a qualitative method of analysis, which proceeds as a confidential and secure conversation between interviewer and a respondent. The interview ensures that the conversation encompasses the topics that are crucial to ask for the sake of the purpose and the issue of the survey (Boyce & Neale, 2006). As P University has the biggest engineering students in South Korea and has the largest number of women engineering students in terms of gender ratio, it was selected as the best model of engineering college to research the women dropouts (Youn, 2014). An in-depth interview took place in a private home or her room in dormitory, where the respondents are in their natural surroundings. Therefore the respondents were relaxed and often willing to reply to the exhaustive questions. An in-depth interview conducted between October and December, 2015 and varies between 1.5 and 2 hours. Interviews were recorded with permission of respondents for the sake of the following analysis and writing of the report. For the sake of the respondents, these recordings were deleted half a year after at the latest. The transcribed information obtained from interviews were summarized, and while analyzing over the classified responses, the key word and thematic responses corresponding to factors were extracted. The preliminary testing factors and framework of questionnaire were derived from obtained students information. By means of the data obtained from depth-interviews questionnaire was subsequently formulated on the basis of that information. Questionnaire survey The survey questionnaire was given to 300 women engineering students including 20 women students who changed their major from college of engineering. Total of 280 cases were collected through a random sampling method by means of questionnaires and questions were put to the respondents to measure the levels of satisfaction. Survey questionnaire also collect the coed information about students opportunities to learn both in and out of the classroom as well as students experiences. The framework for questionnaires, developed in 2006 and advanced for this research in 2015, could guide the collection and responding of women engineering students information. Survey questions undergo a multi step response based development process before being used. This process includes preliminary testing questions with small samples of students as well as piloting draft questions to large samples. Prior to pretesting, pilot testing, and operational administration, the survey questions were submitted. Students were given 15 minutes to complete the questions, which are located at the end of the assessment. Students were encouraged to answer as many questions as they feel comfortable with, and they could skip any question they do not want to answer. All responses were kept confidential. Student names were never reported with their responses or with the other information collected by ours. The regular engineering female students and dropped out students questionnaires were administrated separately. The measurement tool was extracted by the information from the interviews of women students, and its validity was confirmed by three professors of engineering, one doctor of education, and two graduate students majoring in education. Table 2 is a framework of survey questionnaire. The questionnaire consisted of 56 questions, including 14 items of individual factors, 5 items of familial factors, 14 items of university factors, 18 items of sociocultural factors and 5 items of basic demographic questions. To prevent the tendency of respondents to check the middle point, a 5 point scale was designed (very much, somewhat, so and so, not very, and absolutely no.) 5659

Table 2: Formulated framework of questionnaire Questions Individual Factors Item 1 Attention to transfer to other major or dropout of college as of now Item 2 Intention to change the major due to the influence of students around you Item 3 Your friends who transferred to other major or who dropped out Item 4 Major agreement Item 5 Consistency between the degree of major and your goal Item 6 Job recruiting Item 7 Difficulties of coursework in the majoring field Item 8 Motivation of selecting your major Item 9 Information on the engineering field at the time of selecting the program Item 10 Satisfaction of the curriculum of the major Item 11 Availability of consultants (mentors) to discuss your individuals Item 12 Possibility of continuing a job in the field you are majoring in Item 13 Inner conflict about your major Item 14 The main agent that decides your career and your satisfaction Questions Domestic Factors Item 15 Parents perception of women engineering circumstances Item 16 Financial assistance of your family Item 17 Part time work and payment of your school expenses Item 18 Albeit Item 19 Parents occupations Questions Sociocultural factors Item 1 Society s prejudice against women engineers Item 2 Gender differences in the job hunting competition Item 3 Business owners perception Item 4 Gender differences experienced during one s career as a engineer Item 5 Preference for men or women in Business Item 6 Comparison of engineering with other majors in terms of job landing and success Item 7 Cultural gender differences Item 8 Role-models in the engineering fields Item 9 Job landing in terms of culture Item 10 Obstacles for continuing the carrier (marriage, child birth, etc.) Item 11 Male-dominant culture Item 12 Gender differences for job hunting Item 13 Agreements between female culture and business culture Item 14 Glass ceiling barriers Item 15 Environments of career jobs Item 16 Discrimination due to culture differences Questions College Educational Factor Item 1 Commuting distance Item 2 Item 3 Item 4 Item 5 Item 6 Item 7 Item 8 Item 9 Item 10 Item 11 Item 12 Gender cognitive teaching methods Gender cognition in the curriculum Ratio of female faculties in college Circumstances of engineering college for gender priority Difficulties in Lab classes Leaders sex in the team projects or classroom work Scale of gender cognitive lectures Educational supplementary training for women students Courses directly related to employment of women Women in engineering (WIE) program Special attention for women engineering students in college Results and discussion Extraction of Factors through In-depth Interviews The results of this study have verified the contracted four factors of the women engineering students dropout measuring tool which consisted of five preliminary test categories such as individual, domestic, social, cultural and college educational factors. The reason was that students responses for most of the social factors were involved to the more wide range of sociocultural factors. This study was looking at the educational factors of the negative gender cognition learning environments, responses showed that similarity as women students experimented in college, they perceived the male dominant cultures in business and industrial field. Also from the results shows that college should be provided with the opportunity to develop gender cognitive subjects and cubiculum for female students so as to avoid a more difficulties then male students in engineering education. However, in many engineering classes and lectures included lab experiments which were not aware of how to measure women engineers individual, social, cultural, environmental, and technical competencies. Therefore, this measurement tool developed in this study can be used to measure women students competencies in college of engineering and to make a program for women in engineering (WIE) education (Enyegue, 2004). In addition, it is necessary to provide an additional social and academic support program in order for boosting women students capability and for improving the retention rates in engineering. We introduced the physical competencies of handling the tools or machines to one of the items in questionnaire format as an educational factor (Grant, 2006) and disparity of age (Bauman, 1967). Driving machines, handling tools, assembling constructions and technical competencies in engineering lab etc. are factors that have negative effects on women students academic achievement in coeducational environments, and women students often needed to help her male classmate. Many of Korean women students have a tendency to rely on their parents while American woman students have an independent mind, most of expenses of tuition and living expenses are assisted or 5660

provided by their parents. Therefore we added to the factors of household situations, and we found thier parents assistance helps significantly motivates women engineering students study. Analysis of Dropout Factors Descriptive Statistical Analysis The minimum and maximum values of the individual, domestic, sociocultural and college education factors were 2.22~3.62, 2.51~3.26, 2.13~2.94, and 2.73~3.33, respectively. These results approved the variables as normally distribution of data based on the degrees of skewedness and kurtosis because two degrees were less than the absolute value of one. The rule of thumb could be applied to test the even normal distribution as the samples of responses were large. The standard deviations of the four factors were 0.88~1.36, 0.93~1.22, 0.86~1.15, and 0.99~1.42, respectively. The results were well distributed. The Skewedness were distributed between 1.32 and 0.39 and Kurtosis showed between 0.69 and 1.79 on average, which means that the responses obtained by the questionnaires were not much biased but normally distributed. Reliability and validity Items analysis was performed to verify the reliability and validity of the developed measurement tool in this research. The test score on the reliability of the items of female engineering students were between.618 and.869, respectively. These results are comparatively high score for the obtained results in a field of social science. (See Table 3). Table 3: Reliability scores Factors Cronbach`s alpha Individual Factor.869 Domestic Factor.618 Sociocultural Factor.832 Educational Factor.642 Factor Analysis Based on the results of the factor analysis, faculties in college of engineering or government persons in South Korea can use the measurement tool for a better understanding of the mind of dropping out of women engineering students by measuring above four competencies. Factor analysis was performed to measure the variable factors according to the main items analysis results and by using the Varimax which is the right angle rotation method. Table 4 shows the results of individual factors and classified new items related to individuals. Factor 1 on Table 4 refers to the satisfaction of her major because the score of Item 9 (the information on the engineering field at the time of choosing her major) and the score of Item 8 (the time to determine her major) shows the highest scores, respectively. Factor 2 refers to the influence of the circumstances because the value of Item 3 (students friends who changed their major from engineering and classmates or friends who took a leave of absence) was.807. Factor 3 refers to the accommodation and networking with classmates and faculties because it is a factor concerning the counselors on her anxiety of her major program or getting a job after graduation. Factor 4 refers to recruiting and continuing her engineering career. The Eigenvalue of variance that we calculated are a measurement of how much of the variance of the observed variables a factor explains. Any factor with an eigenvalue 1 means more variance than a single observed variable. Same as individual process, as shown on Table 5, the domestic factor was consisted of five items. Factor 1 refers to the financial conditions and Factor 2 refers to the parents assistance. Table 4: The results of factor analysis of individual factors Questions Factor 1 Factor 2 Factor 3 Factor 4 Item 9.865.123 -.052.111 Item 8.834.244 -.049 -.049 Item 10.701.173.280.275 Item 13.482.200.297.472 Item 7.421.379.057 -.170 Item 3.091.807.105.383 Item 1.151.772.320.161 Item 2.163.800 -.109.091 Item 5.465.629.350 -.104 Item 11 -.148.109.790.113 Item 14.493.005.575.237 Item 4.518.299.551 -.212 Item 6 -.101.143 -.069.809 Item 12.291.084.398.689 3.217 22.978 2.570 18.354 1.842 13.201 1.851 13.012 Table 5: The results of factor analysis of domestic factors Questions Factor 1 Factor 2 Item 18.879 -.138 Item 17.851 -.031 Item 16.781.212 Item 19 -.155.891 Item 15.481.765 2.512 46.882 1.812 23.124 Table 6: The results of factor analysis of educational factors Questions Factor 1 Factor 2 Factor 3 Item 2.712 -.051.056 Item 4.682.078.071 Item 5.651.051.112 Item 6.579.397.041 Item 3.551 -.091.451 Item 12 -.179.669.312 Item 9 -.001.632 -.041 5661

Item 10.391.617 -.371 Item 8.171.581.028 Item 1.248 -.171.591 Item 7.171.173.541 Item 11 -.312.397.527 3.572 18.498 2.031 17.217 1.821 12.531 Table 7: The results of factor analysis of sociocultural factors Questions Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Item 2.771.151.068.041 -.012 Item 3.742.072.115.246 -.167 Item 1.701.098 -.126.149 -.268 Item 5.691.121.115.096.201 Item 11.681.178.161.109.298 Item 14.129.789.047 -.029.237 Item 15.251.721.125.119.081 Item 16.052.691.076.335 -.081 Item 13.101.601.498 -.137 -.199 Item 9.091.098.805.197 -.052 Item 4.272 -.159.557 -.004.152 Item 6.312.051.041.827.057 Item 7.132.301.327.467.141 Item 12.151.298.397.421 -.081 Item 8 -.061.051 -.107.121.831 Item 10.497.247.191 -.179.517 3.818 21.987 2.452 14.327 1.723 11.716 1.370 8.565 1.587 8.914 The college educational factor consists of three categories (see Table 6). Factor 1 on the Table 6 refers to the circumstances of engineering campus and the relationships between female students and male faculties because factor 1 involved the items 2 through 6 related to gender cognitive teaching and learning method. Factor 2 on the Table 6 is composed of items for existence of special program for women engineering students such as women in engineering and science (WISE) or women in engineering program (WIE) etc. P University is a leading university of WIE program since 2006 in South Korea and one of the universities have a WISET which is women in science, engineering and technology program. Factor 2 refers to gender cognitive engineering education. Factor 3 on the Table 6 refers to gender cognitive curriculum and workforces for women engineering students. The score of factor 3 was.591 which was the highest value. Table 7 shows the results of factor analysis of the sociocultural factors. We composed five items on the sociocultural factor. Factor 1 on the Table 7 refers to the perceptions in society or industries to the women engineers because the score of Item 2 (gender difference in competition) and Item 3 (business owners perception of women) played an important role. Factor 2 refers to the discrimination against female engineers and the score of item 14 concerning the culture of engineering fields and the glass ceiling index item was.789. Factor 3 on the Table 7 refers to the job recruiting and Factor 4 refers to priority of man for employment or promotion etc. and Factor 5 on the Table 7 refers to her role model in engineering. These researches have been progressed since 2006, and we had proposed to programs for the gender cognitive engineering education and the special programs for women engineering students. As shown in Figure 1, women engineering students dropout rate in college of engineering in P University was about 5 % at 2006, it was much higher than 1 % of male students, however the dropout rate of women engineering students were decreased to 1% at 2015 due to women in engineering program and proposed programs based on the results of factor analysis of women engineering student s dropout from college of engineering.. Figure 1: The rate of dropout students of engineering at P University for 9 years. Conclusion In order to increase the retention rate and decrease the dropout rate of women engineering students, factor analysis for female engineering students in P University in South Korea was conducted. The analytical results showed that social and cultural effects are more significant ones than social effect alone, and the effects of gender discrimination in industrial field and college to the women engineers were more significant than effects of regional or positional ones. Also results showed responses that the physical difference from man students, the ability for driving machines or handling tools, her parents financial support, the gender cognitive engineering education and the boosting for job recruiting are more significant effects than the scholastic ability or the ageism. In this research, we developed the measurement tool to measure women students competencies in college of engineering and to make programs for women in engineering (WIE) education according to the results of factor analysis. As the results, factor analysis and women in engineering education programs have progressed successfully. Acknowledgments This work was supported by a Research Grant of Pukyong National University (2015 Year) 5662

References [1] Youn, J. T. and Han, H. W., 2011, Effectiveness of women in engineering course, Journal of Engineering Education Research, 14(3), pp.3-8. [2] Kim, J. H., Oh, M. S. and Jeong, Y. J., 2007, Study on women's engineering workforce for improved adaptability site. KOTEF. [3] Park, J. E., Kim, J. H., Jeong, Y. K. and Oh, M. S., 2009, Analyses of female engineering education programs abroad. Journal of Engineering Education Research. 12(3), pp.79-95. [4] Park. H. Y., 2008, Women in science technology and their supporting policy. Women s Study Research. 18, pp. 141-181. [5] Wolffram, A., Derboven, W. and Winker. G., 2009, Women withdrawers in engineering studies: Identity formation and learning culture as gendered barriers for persistence, Equal opportunities international, 28(1), pp. 36-49. [6] Youn, J. T. and Choi, S. A., 2015, Factors of women engineering students dropouts in South Korea, The Online Journal of Quality in Higher Education, 2(4), pp. 24-34. [7] Oh, Y. J., 2006, An analytic study on the university life experiences of freshman, Education Research. 25, pp. 141-164. [8] Lee, Y. J., 1991, Multivariate analysis using the + SPSS / PC. Seokjeong Press, Korea. [9] NRF, 2010, Women into engineering (WIE) fourth consulting report. NRF (National Research Foundation) of Korea. [10] Boshier, R. W., 1973, Educational participation and dropout: A theoretical context: an analysis of Boshier ' s congruence model. Adult Education. [11] Boyce, C. and Neale, P., 2006, Conducting in-depth interviews: A guide for designing and conducting in-depth interviews for evaluation input. Pathfinder. [12] Youn, J. T., Han, H. W. and Choi, S. A., 2014, A study on the factors of women student s dropout from engineering college, Journal of Engineering Education Research, 17(6), pp. 46-52. [13] Garrison, D. R., 1983, Psychosocial correlates of dropout and achievement in an adult high school completion program. Alberta Journal of Educational Research. 29(2), pp. 131-139. [14] Garrison, D. R., 1987, Dropout prediction within a broad psychosocial model. Adult education. 23(4), pp. 255-282. [15] Morrow, G. P., 1986, Standardizing practice in the analysis of dropouts. Teacher College Record. 87. [16] Nunnally, J. C., 1976, Psychometric theory. McGraw-Hill. [17] OECD, 2009, OECD Korea depth analysis of innovation policy. Recommendations for the elimination of all Forms of discrimination against women in science and technology in South Korea. Quarterly. 37(4). pp. 212-222. [18] Grant, M. and Hallman, K., 2006, Pregnancy-related school dropout and prior school performance in south Africa, Population Council, 212, pp. 1-30. [19] Bauman, K. E., 1967, The Relationship between age at first marriage, school dropout, and marital instability: An analysis of the Glick effect, Journal of Marriage and Family, 29(4), pp. 672-680. [20] Enyegue, E. and Papfait, M., 2004, Pregnancyrelated dropouts and gender inequality in education: A life-table approach and application to Cameroon, Demography, 41(3), pp. 509-528. 5663