xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2011 flu ob 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0004833
Ureksoy, Heather Elizabeth.
The role of program climate and socialization in the retention of engineering undergraduates
h [electronic resource] /
by Heather Elizabeth Ureksoy.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 108 pages.
(Ph.D.)--University of South Florida, 2011.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
ABSTRACT: Increasing women's participation in the fields of science, technology, engineering and mathematics (STEM) can promote a healthy economy by ensuring a diverse and well-qualified STEM workforce, not only in the quantity of females in the workforce, but diversity in thinking and creativity. It will also send a positive message to young women about the breadth of educational opportunities and career choices they have available to them. However, women continue to participate in engineering education in a far lower rate than men. Attracting and retaining female students has become a challenging problem for the academic engineering community. In this study, a classic model of student withdrawal is presented as a theoretical framework for examining the relationships between the environment and the people in undergraduate engineering departments, and how they can influence students' commitment to and persistence in their program. A sample of 1,369 engineering undergraduates enrolled in eight Florida universities participated in a survey assessing the climate of the engineering department, the socialization process, student commitment and withdrawal intentions. The results of a factor analysis reveal that faculty support, a sense of community, and encouraging and valuing diversity are all important elements of a climate for retention. In general, women perceived the academic climate as being less supportive than men did, reported lower levels of commitment, and greater withdrawal intentions. These climate factors, as well as socialization, also played a significant role in predicting the levels of student commitment to their program, and their intentions to withdraw or persist in their academic goals. However, there was little evidence for a moderating role of gender in these relationships. This research suggests the importance of having a supportive faculty and fostering a sense of community among students, both of which aid in the successful socialization of engineering students, and ultimately promote commitment and persistence.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Borman, Walter .
x Psychology Higher Education
t USF Electronic Theses and Dissertations.
The Role of Program Climate and Socialization in the Retention of Engineering Undergraduates by Heather Ureksoy A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Psychology College of Arts and Sciences University of South Florida Major Professor: Walter Borman, Ph.D. Tammy D. Allen, Ph.D. Michael Brannick, Ph.D. Joseph Vandello, Ph.D. Kathryn Borman, Ph.D. William Tyson, Ph.D. Date of Approval: April 6, 2011 Keywords: STEM, college women, persistence withdrawal Copyright 2011, Heather Ureksoy
Acknowledgements My journey through graduate school would not have been possible without the enduring support I received from my husband, Erol, whose inexhaustible patience and encouragement helped me stay the course; my mother, Diana, who always believed in me and who coo ked and cleaned for me so I could concentrate on writing; and my dear friend Cassie, without whom I would not have survived all the data collection trips we took for this project. I would also like to thank my advisor, Dr. Walter Borman, for all his suppor t and guidance, and his wife, Dr. Kathryn Borman, the principal investigator for the grant from which this dissertation was born. Finally, I would like to acknowledge the National Science Foundation for funding the grant that provided the data for this di ssertation.
i i Table of Contents List of Tables ................................ ................................ ................................ ..................... i v List of Figures ................................ ................................ ................................ ...................... v Abstract ................................ ................................ ................................ .............................. v i Chapter One : Introduction ................................ ................................ ................................ ... 1 Chapter Two : Review of the Literature ................................ ................................ ............... 7 Institutional experiences and organizational climate ................................ ............. 1 1 Integration and organizational socialization ................................ .......................... 1 6 Withdrawal, intent to withdraw and intent to turnover ................................ .......... 1 9 Commitment ................................ ................................ ................................ .......... 22 ................................ ..... 25 Chapter Three : Method ................................ ................................ ................................ ...... 30 Survey a d ministration ................................ ................................ ............................ 31 Participants ................................ ................................ ................................ ............. 32 Inclusion criteria ................................ ................................ ........................ 32 Sample characteristics ................................ ................................ ................ 32 Comparison of the sample characteristics to national enrollment ............. 32 Variables ................................ ................................ ................................ ................ 3 4 Involvement ................................ ................................ ............................... 35 Faculty s upport ................................ ................................ ........................... 35 Helpful ness ................................ ................................ ................................ 36 Diversity ................................ ................................ ................................ ..... 36 Integration ................................ ................................ ................................ .. 36 Socialization ................................ ................................ ............................... 36 Commitment ................................ ................................ .............................. 37 Intent to withdraw ................................ ................................ ...................... 37 Background information ................................ ................................ ............ 38 Chapter Four : Results ................................ ................................ ................................ ........ 39 Preliminary analysis ................................ ................................ ............................... 39 Data screening ................................ ................................ ............................ 39 Assumptions testing ................................ ................................ ................... 39 Factor a nalysis ................................ ................................ ................................ ....... 41 Multiple r egression ................................ ................................ ................................ 46 Gender differences ................................ ................................ ..................... 47 Hypothesis testing ................................ ................................ ...................... 49
iii Hypothesis 1 and 2 .........................................................................49 Hypothesis 3 and 4 .........................................................................51 Hypothesis 5 and 6 .........................................................................52 Hypothesis 7 and 8 .........................................................................55 Exploratory analysis...................................................................................55 Chapter Five: Discussion ...................................................................................................60 Dimensionality of a climate for retention ..............................................................60 Faculty support...........................................................................................61 Sense of community ...................................................................................64 Encouraging diversity ................................................................................65 Valuing diversity ........................................................................................66 Predicting indicators of persistence among engineering undergraduates ..............69 Climate and intentions to withdraw ...........................................................70 Socialization and intentions to withdraw ...................................................71 Climate and commitment ...........................................................................72 Socialization and commitment ...................................................................73 Theoretical and practical implications ...................................................................74 Limitations of the study .........................................................................................76 Directions for future research ................................................................................77 Conclusions ............................................................................................................79 List of References .............................................................................................................80 Appendix : Effects of College Degree Program Culture on Female and Minority Student STEM Participation: Student Survey ......................................93 About the Author ............................................................................................... END PAGE
iv List of Tables Table 1. Demographic characteristics of the sample ................................ ......................... 39 Table 2. Educational characteristics of the sample ................................ ............................ 40 Table 3. Pattern c oefficients for the c limate for r etention survey ................................ ...... 51 Table 4. Correlations among major study variables ................................ ........................ 52 Table 5. T tests of major study variables by g ender ................................ ........................ 54 Table 6. Regression of i ntent to w ith draw on c limate for r etention factors ...................... 57 Table 7. R egression of i ntent to w ithdraw on s ocialization ................................ ............... 59 Table 8. R egression of c ommitment on c limate for r etention factors ................................ 6 0 Table 9. Regression of c ommitment on s ocialization ................................ ........................ 62 Table 10. Climate rating by majors with high and low female participation .................... 63 Table 11. Ratings of major study variables by year in school ................................ ........... 64
v List of Figures Figure 1. ................................ ......................... 16 Figure 2. Participant enrollment by major ................................ ................................ ......... 41 Figure 3. Ratings of study variables by gender ................................ ................................ .. 55 Figure 4. Climate ratings by majors with high and low female participation .................. 63 Figure 5. Ratings of major study variables by year in school ................................ ............ 65
vi A bstract Increasin fields of science, technology, engineering and mathematics (STEM) can promote a healthy economy by ensuring a diverse and well qualified STEM workforce, not only in the quantity of females in the workforce, but diversity in t hinking and creativity. It will also send a positive message to young women about the breadth of educational opportunities and career choices they have available to them. However, women continue to participate in engineering education in a far lower rate t han men. Attracting and retaining female students has become a challenging problem for the academic engineering community. In this study, a classic model of student withdrawal is presented as a theoretical framework for examining the relationships between the environment and the people in undergraduate engineering their program. A sample of 1,369 engineering undergraduates enrolled in eight Florida universities participated in a survey assessing the climate of the engineering department, the socialization process, student commitment and withdrawal intentions. The results of a factor analysis reveal that faculty support, a sense of community, and encouraging and valuing divers ity are all important elements of a climate for retention. In general, women perceived the academic climate as being less supportive than men did, reported lower levels of commitment, and greater withdrawal intentions. These climate factors, as well as socialization, also played a significant role in predicting the levels of student commitment to their program, and their intentions to withdr aw or persist in their
vii academic goals. However, there was little evidence for a moderating role of gender in these relationships. This research suggests the importance of having a supportive faculty and fostering a sense of community among students, both o f which aid in the successful socialization of engineering students, and ultimately promote commitment and persistence.
1 Chapter One : Introduction Over the last thirty years, there has been dramatic participation, and with it an interest in examining what makes women a distinct group with unique values, attitudes and career oriented behaviors. This expanding body of career related choices. Even though women make up over half the work force (National Science Foundation [NSF], 2009), many occupations today continue to be sex stereotyped. Some are commonly characterized as historically male dominated, such as science, t echnology, engineering, and mathematics ( hereafter referred to as STEM), whereas others tend to be viewed as more female oriented, like education and healthcare. Ideas about gender typed occupations are pervasive, and evidence for them can be found in children as young as three years old (Stockard & McGee, 1990). Internalizing these beliefs about gender typed view of their own career potential and the options available to them. Although there is evidence that these stereotypes may be declining among college students (White, Kruczek, Brown, & White, 1989), out dated conceptualizations of gender typed jobs continue to play an important role in the development of many care er pathways. Women tend to be attracted to fields and occupations that provide an opportunity for social interaction and allow the individual to play a useful role in society. This can include jobs in fields such as healthcare or social services. In fact, when women in traditionally male dominated fields were asked about their career plans, many reported
2 they were more likely to consider changing to a career which would allow them to give back to the community in a way that their current field of study woul d not (Lightbody, Siann, Tait, & Walsh, 1997). A classic example of how women me e t with resistance participating in a male dominated field can be seen in the engineering discipline. Throughout the most of the 20th century, women studying or working in e ngineering were generally perceived as going against traditional gender norms. Historically, engineering has had masculine connotations. This can be attributed, in part, to the apprentice style training most working engineers participated in, as opposed to a formal educational program. The hands on nature of training engineers was considered unsuitable for women to participate in, because it often involved physically strenuous or dangerous conditions. Other social cues also reinforced the masculine image of engineering. Tool kits and model trains were interests were often encouraged towards more traditional disciplines (Bix 2002; Oldenziel 1999; Purcell 1979; Wajcman 1991). War II, when our country faced a serious manpower shortage. Women were trained on the job for their new responsibilities, yet the idea of women formally studying engineering in a university setting was still unheard of. Slowly, this attitude began to change, and women were allowed to apply to and participate in undergraduate engineering programs of study, although they were met with an incredible amount of resistance and subject to stereotyping to an alarming degree. There was a great deal of
3 concern about the consequences of having women study at a technical institution dominated by men (Bix, 200 0 ) To combat these negative attitudes concerning the sparse number of female engineering students, support groups started to emerge, and in 1952 the Society for Women Engineers (SWE) was created. SWE continues to advocate for women in engineering. SWE play s an important role in the encouragement of female engineering students by giving women a way to meet each other, and to build a sense of community with a shared purpose ( SWE, 2009 ) With support for female engineers starting to grow, a new generation of MIT coeds began to band together to consider remedies to emerging workplace challenges. They hosted meetings and conferences to address issues such as employment discrimination. Meanwhile, important changes were underway at the national level. The 1964 Fe deral Civil Rights Act had included language barring gender based employment discrimination. Major companies during the 1970s promoted the fact that they were "an equal opportunity employer". To encourage diversity, many companies deliberately marketed the ir recruitment efforts towards women (Woloch, 1999) Although we have come a long way from the need for female engineers as a quick fix for wartime labor shortages, women are still alarmingly underrepresented in the field. In 1979, women made up 12.1 per cent of undergraduates enrolled in engineering across the United States; currently, that number has risen less than ten percent. In 2006, women represented the majority of college students in four year institution (56%), yet only 17% of engineering underg raduates. This represents a smaller proportion of students than in the previous decade. Men, on the other hand, have exhibited a steady increase in
4 enrollment over the past eight years. Graduate school enrollment shows a similar trend for female engineers. Women enrolled in graduate engineering programs increased by 40% from the previous decade, but they still represent only a quarter of all engineering graduate students (NSF, 2009) Enrollment, however, is only half the picture. Over 68,000 engineering were awarded in 2007, yet less than one quarter of them went to women. fields, but women were vastly overrepresented in a small subset of these disciplines, such as biological science; this discipline often attracts female students as it serves as a n entry to the healthcare field. The doctoral degrees awarded mirror the trend seen in undergraduate degree attainment. While women earned half of all the doctoral degrees, men still graduated with a Ph.D. in engineering at a rate five times that of women. The doctoral degrees. At the state level, a similar distrib ution of women in engineering is found. Enrollment in four year, Florida State University System institutions has increased 40% between 1998 and 2007, significantly higher than the national figures (23% increase nationwide). Of these students, 56% are fema le, similar to the national rate. Yet in 2007, women made up 17% of engineering undergraduates enrolled and 22% of engineering graduate students enrolled (Florida Board of Governors, 2007). It is clear to see that uate institutions is not lacking, yet they consistently choose to enter and graduate from engineering programs at a fraction of the rate they enter many other STEM program, not to mention non STEM programs
5 nterest in STEM education and careers. In fact, the National Science Board (2007) identified the supply of scientists, engineers, and science teachers as one of the top 10 priorities of the early 21 st century. Increasing lds can promote a healthy economy by ensuring a diverse and well qualified STEM workforce, not only in the quantity of females in the workforce, but diversity in thinking and creativity. It will also send a positive message to young women about the breadth of educational opportunities and career choices they have available to them. However, women continue to participate in engineering education in a far smaller proportion than men. Attracting and retaining female students has become a challenging problem fo r the academic engineering community. Effective solutions for student retention must be designed in order to increase the retention of qualified and talented female engineers. The reasons for this underrepresentation are complex, but one factor continues t Sandler, 1982, 1984; Heller, Puff, & Mills, 1985; Constantinople, Cornelius, & Gray, 1988; Crawford & MacLeod, 1990; Pascarella, Whitt, Edison, Nora, Hagedorn, Yeager, & Terenzini, 1997; Wh itt, Nora, Edison, Terenzini, & Pascarella, 1999, Bix, 2004). After all, individuals do not exist in a social vacuum T he role of organizational conditions in the workplace is important for understanding the outcomes among individuals. Likewise, in underst anding the experiences of women in undergraduate STEM programs, we need to consider features of the program in which they study and attend class. These features are influential in student experiences across fields, but can be especially important in scientific fields where there is such gender disparity. Certain g roups of people, such as women compared with men, or underrepresented minority students compared to non
6 minorities, can have diff erent experiences in these organizational environments, with implications for differences in outcomes such as persistence educational experiences in and out of the classroom, a nd thus be subject to positive intervention, yet they do not receive the necessary attention in the literature. These both while in college and after graduation. Under standing that factors that influences number of students who actually do persis t in engineering education (Wyer, 2000). The most significant contribution this research is to apply established organizational theory and methodology to the study of undergraduate students in engineering and to delineate the components of an academic c limate w hich support the persistence of its undergraduate engineering students. The proposed research will apply organizational theory to the empirical study of the perceptions of academic climate among undergraduate engineering students, and the relations hip between these climate perceptions and undergraduate commitment and withdrawal intentions.
7 Chapter Two : Review of the Literature I t is clear that women remain underrepresented in engineering and science undergraduate programs and careers. In this chapter the research on student persistence will be reviewed, with a focus on factors for retention at the undergraduate level. When possible, issues pertaining specifically to women and/or STEM programs will be highlighted. A classic model of stu dent withdrawal developed by Tinto (1975) is presented as a theoretical framework for the study, laying the groundwork for the relationships between the environment and the people in an academic program and how th commitment to their program and intentions to persist or withdraw. The m and discussed in the context of central industrial/organizational constructs. Much of the contemporary research on persistence in undergraduate science programs has generally focused on pre college attributes and influences such as socioeconomic status, parental influence, or high school course taking and achievement For instance, one study of pre college students used informati on about family, school and individual variables to predict success and achievement in college (Hansen, 2000 ). Another study examine d the effect of high school math and science achievement on subsequent college course taking (Wood & Brown, 1997). Yet research indicates that post matriculation experiences are equally important to pre matriculation student characteristics when explaining commitment and retention ( Terenzini & Pascarella, 1980 ).
8 The study of student experiences in high er education begins with a review of the research o n history leading up to college, but on what happens to a student after they enroll. They focus on inter individual origins of studen interaction with the environment around him and how this environment reinforces or alters their original goals and plans. The research on student change seeks to answer several critical questions about student experiences in college, such as : a) Can students have different experiences while in the same institutional environment, and b) Does the college experience create different outcomes for various subgroups of students? These models often incorporate factors reflecting academic climate, and can be used to help clarify the undergraduate experience in a STEM program for both men and women. This goal of the current study is to address both of these questions with respect to women in engineering. Research on stu E O Model and Theory of Involvement (1970) This was one of the original college impact models developed, and is based on the now familiar input -process (or in this case, environme nt) -output model. General Model for Assessing Change (1985) a causal model that specifies the role of Weidman (1989) created We i Model of Undergraduate Socialization (1989) is focused on predicting non cognitive changes, such as those involving career choices, attitudes and values, which contrasts odel, focused on learning and cognitive
9 (1989) model also specifies the role of socialization in much more specific terms Weidman (1989) believed that s college normative ideas, whi ch leads to outcomes such as the alteration of attitudes and (1989) model of student change places the role of student socialization at the forefront, it is limited in its ability to account for a variety of other institutional f actors which are also thought to influence student change, such as perceptions of the program climate. model of s tudent withdrawal (1975, 1993) provides a more comprehensive model of the factors associated with student change, with a specific focu s on the outcome of retention (see Figure 1) According to Tinto, a student enters college with a set of personal and academic characteristics and skills, which are then modified by the vast array of experiences the student has while in college. These expe riences, along program or institution. This classic model of student withdrawal has also been modified and applied by a number of subsequent researchers. For example, Bean (2005) applies a similar, model to the phenomenon of student withdrawal as Tinto does. Here, s tudents begin with pre matriculation attitudes, beliefs and goals. Upon entering college, students interact with the institutional environment, and these in teractions help to form the these attitudes, and a subsequent decision to withdraw can ultimately be traced back to The following sections will introduce, in detail, the main be re interpreted u sing common constructs industrial organizational p sychology.
11 I nstitutional experiences and organizational climate Tinto considers institutional experiences to be a combination of a number of components, including objective indicators such as grade point average and participation in extrac urricular activities, and subjective indicators like the quality of interaction with faculty, staff, and peers. Studies show that the institutional environment does influences program efficacy (Berger, 2001; Braxton, 2001; Noel, Levitz, & Saluri, 1985; Pas carella In this context, the concept of organizational climate can be used to illuminate perception of their environment and describes the atmosphere of the program. In other words, program climate refers to the experience of being a member of the program (Ostroff, Kinicki, & Tamkins, 2001, 2003). An academic program can be viewed as an organization with members including faculty, students and staff. Members of this program interact with one another on a daily basis as they would in other organizational Organizational climate focuses on how organizational participants perceive and make sense of t heir environments. This research has its roots in the work done by Kurt Lewin (1951) and has been applied to both organizational and educational settings. A distinction should be made between climate and culture; climate can be thought of as a description of what happens, and cultur e helps define why these things happen. Therefore, climate is a more proximal indicator of an organization than culture, and is more readily accessible upon entry into the organization. The physical appearance of the organization the
12 attitudes and values held by its employees and the treatment of newcomers all provide evidence for the climate of an organization. Much of the early research on climate explored the influence of the overall climate on organizational effectiveness and attitudinal outcomes (Schneider, Bowen, Ehrhart, & Holcombe, 2000). Recently, there has been a trend away from global definitions of climate, and a & Tamkins, 200 3). This strategic definition of climate is gaining popularity and acceptance in the literature, and has been successfully applied to issues such as safety (Zohar, 2000), service (Schneider, 1990), justice (Naumann & Bennett, 2000), and citizenship behavio r (Schneider, Gunnarson and Niles Jolly, 1994), to name a few. Individual level perceptions of climate have successfully been linked to affective and behavioral outcomes. For example, justice climate has been shown to be related to commitment and helping b ehaviors (Naumann & Bennett, 2000), and climate for sexual harassment has been linked to reports of harassment incidents (Hulin, Fitzgerald & Drasgow, 1996). The emerging research on strategic climates has been promising, but only a few examples of clim ate for have been given empirical attention. More attention to the climate for concept is needed to reinforce the nature of this construct. The current research will expand the climate for literature by introducing and measuring a climate for undergraduate retention. There has been some debate surrounding the precise nature of climate, and at what level it is most appropriately measured. To resolve the issue of level, a distinction was made between psychological climate, which is operationalized and measu red at the
13 individual level, and organizational climate, which is defined as an organizational variable (Ostroff, et al., 2003). At the individual level, perceptions of climate stem from of organizational climate should rely on the individual as the basic unit of measurement. If consensus about climate is reached among individuals in an institution, then these perceptions can be meaningfully aggregated into an organizational level construc t. Because similar people are attracted to the similar environments, exposed to similar features of an organization, and are socialized in a similar manner, it is likely that a consensus among climate perceptions of individuals will develop. This di stincti on between psychological climate as an individual level variable and organizational climate as a group level variable, when appropriately aggregated, is now widely accepted (Schneider et al, 2000). To further support the distinction between individual le vel psychological climate and group level organizational climate, researchers have given much attention to the variety of methods by which aggregated climate perceptions can be transformed into a meaningful group level indicator. The most common method is to use a mean score across individuals to represent a higher level climate. In order to justify meaningful agreement on organizational level climate, two criteria must be met. The first is to show that a sufficient amount of within group agreement exists, and the second is to demonstrate the degree of between unit variability. Although the exact definitions of what qualifies as sufficient within group agreement and across group variance is still under debate, if both of these conditions can be reasonably m et, a researcher is then able to justify the use of aggregated climate perceptions to reflect a cohesive organizational climate (Klein, Cohn, Smith, & Sorra, 2001).
14 There are two types of studies which address the impact of climate on the indi vidual: a) individual level studies, which examine the relationships between psychological climate perceptions and individual outcomes, and b) cross level studies using aggregated climate scores assigned to individuals, and relationships to individual outc omes are examined. The current research represents the former, in that individual level perceptions of climate will be used to predict withdrawal intentions and commitment of engineering students. Several studies have been conducted ( i.e., Hall & Sandler, 1982, 1984; Heller, Puff, & Mills, 1985; Constantinople, Cornelius, & Gray, 1988; Crawford & MacLeod, 1990; Pascarella, Whitt, Edison, Nora, Hagedorn, Yeager, & Terenzini, 1997; Whitt, Nora, Edison, Terenzini, & Pascarella, 1999, Etzkowitz, Kemelgor, & Uz zi, 2000 Bix, 2004, Herzig, 2004 ) that focus on how the climate of undergraduate STEM classrooms, programs, and departments may contribute to a higher level of program efficacy for women and minorities in STEM fields Yet little is still known about which individual elements of an academic climate best support program efficacy, especially for completion of STEM programs. enabling climates and cultures for female academics Specifically, what are the facets of the program environment that facilitate success in STEM programs and foster positive integration into the program on the part of the student? Researchers have found, for example, that departments or programs that focus on collaborati on rather than competition, are collegial rather than bureaucratic, and are student centered rather than institution centered, tend to be associated with increased success for all students, particularly female and minority students (Tinto, 1993; Smith, Ger bick, & Figueroa,
15 1997). However, this research may be just scratching the surface of how STEM program environments that disadvantage women. These include behaviors such as calling on male students more o ften than female students, paying more attention when men speak, and Sandler, 1984). Pri or research suggests that such behaviors and the environment they create often go unnoticed because they reflect socially accepted patterns of communication and the long held belief that men are more capable of working in the fields of hard science (Sandle r, Silverberg, & Hall, 1996; Brady & Eisler, 1999). led to increased self doubt in women, which results in their attrition from engineering fields. A study of male and female engineering faculty members addresses the role of gender in the perceptions of organizational climate (Fox, 2010). Climate was selected for this study because it can encompass aspects of the perceived atmosphere of an academic department along a variety of dimensions. Responses on perceptions of such dimensions (Reichers & Schneider, 1990, p. 22), and such characterizations can vary for male and female faculty. This study found marked differences in the perceptions of the
16 helpfulness and excitement) and significantly higher for negative aspects, including stressfulness. Furthermore, they reported lower levels of communication among colleagues, which lead to restricted sense of integration and membership in their work environment. As a result, female facu lty may feel excluded from the social networks in which they work (Fox, 2001 ). Other research has considered the influence of racial climate on student persistence as well as a climate focused on gender The underlying assumption of these studies is that under represented groups of students, whether that refers to their gender, race, or ethnicity, are sensitive to the climate of the institution and its predominantly white male population. One such study (Sidel, 1994) of campus racial climate found that un der represented groups, including racial or ethnic groups and women, were made to feel like outsiders within their institution. These feelings of alienation had a negative commu nity. I ntegration and organizational socialization institution. Tinto places a great deal of importance on the role of student integration into the academic and social systems of th e institution and its influence on student departure. values of peers and faculty in the institution and abides by the formal and informal structural requirements for mem (Pascarella & Terenzini, 2005, p. 54). I ntegration as a process by which students learn the
17 normative attitudes and values of the program and its members, and act in accordance with the requirements for pro gram membership. Tinto (1978) theorizes that positive encounters with the program environment lead to greater student integration. T here is much variability in the ways in which integration is operationalized. Many studies have traditionally relied on objective indicators of integration, such as grade point average, or the degree to which the student participates in socially and acade mically oriented groups or events. There is a lack of emphasis on the psychological conceptualization of community with the institution, which the proposed study seeks to ad dress in greater detail. One notable exception is research conducted by Hurtado and Carter (1997). In this le sought to measure the Feeling socially integrated and connected with other students is an im portant factor which can affect Col lege is a time for both intellectual and social growth the social lives of students in college and their exchanges with others inside and outside Not surprisingly, a person is more likely to meet difficult goals when they are surrounded by people who share a similar goal. It is the same for college students. A major part of the social acclimation process is adapting to the unfamiliar environment an d seeking out like minded peers. Having a social bond with peers in an academic environment can provide some much needed support as everyone moves towards a common goal: graduation. Having this bond
18 with peers in the same area of study is especially import ant, and can be critical in fields such as engineering, where the workload is often intense. These friendships are important because they allow students to participate in academic activities together, such as study groups and team projects. This, in turn, their academic pursuits. It is important for institutions to recognize the importance these institutional officials to recog nize that social connectedness is important for confidence, loyalty, and 228 229). In this c can be interpreted as organizati onal socialization. Organizational socialization refers to the process by which an individual learns and adopts the values, attitudes, and knowledge required of successful members of that organization (Bauer, Morrison & Callister, 1998). Socialization can occur whenever an individual begins a new role, but the greatest degree of socialization is experienced near the time of entry to the organization. R esearch on organizational socialization also provides support for the influence of student departure theory. Socialization is often linked to important outcomes for both the individual and the organization. Just as Tinto theorizes in his student departure model, organizational research has found that unsuccessful socialization often lead s to turnover (Campion & Mitchell, 1986). Conversely, positive socialization of an individual into an organization can have a lasting impact on their both the organi zation and to their shared goals (Bauer et al., 1998), which parallels
19 to their goals and intentions. Finally, the socialization process is a primary vehicle for commu nicating information about the organizational culture and climate. This is beneficial for both the student and the organization; student s are better able to interpret their environment and can predict the consequences of their actions, whereas the organiza tion is able to ensure the continuity of their values and goals (Louis, 1990). Withdrawal, intent to withdraw and intent to turnover Although model of student withdrawal, there are several more proximal variables in his model that program withdraw. This is influenced by institutional experiences and student integration, and is an influence on actual withdrawal. According to Tinto, withdrawal is a longitudinal process that results from interactions between a student and his/her environment. A departure decision is made by the student, which is influenced by the environmental experience, degree of integration within the program, and d egree of commitment felt by the student to the program. Negative experiences and integration lead to student departure, whereas posit ive experiences and integration encourage persistence (Tinto, 1975). Mobley, Griffeth, Hand and Meglino (1979) proposed a m odel of turnover in the workplace in which various features of the work environment influence the which in turn influences withdrawal cognitions and ultimately ends in em ployee turnover.
20 student departure, in that a re lationship between environment and withdrawal intentions is specified. odel is ( actual) withdrawal from the institution. In this context, withdrawal can be interpreted as turnover from an organization. In certain environments, turnover can be problematic to measure. There is often a low base rate of turnover occurrences, and is almost always requires a longitudinal design. Furthermore, it may be the case that an intention to do something is interesting in itself, and worthy of investigation. Because of this, researchers often rely instead on a surrogate variable, intention to turnover (Miller & Wheeler, 1992; Mone, 1994; Kirschenbaum & Weisberg, 1994; Rosin & Korabik, 1995; George & Jones, 1996; Sjoberg & Sverke, 2000; Freund, 2005) There are advantages to using a surrogate variable such as intent to turnover, or in this case, intent to withdraw from the program. For example, a dichotomous outcome of withdraw or persist has limited statistical variance, whereas intent to withdraw can be scaled to provide a wider range of data, as well as allow for varied experimental designs th at do not include a longitudinal element. Recommendations have been made about when it is appropriate to use surrogate variables, such as intent to turnover. First, the behavioral variable (i.e., turnover) must be unavailable for measurement, due to access or experimental design. Secondly, the surrogate variable is the only way to represent the variable of interest in the study. Finally, researchers suggest that when considering the use of a surrogate variable, its relationship with the actual behavior of i nterest (whether determined by empirical test or meta analysis) should exhibit a correlation of approximately 0.50 (Dalton, Johnson, & Daily, 1999). However, these recommendations only apply to the use of surrogate
21 variables in the place of the behavioral variable. If the surrogate is being modeled as a meditating influence, these recommendations do not apply. There have been several meta analytic reviews that provide estimates of the relationship between turnover and intent to turnover in a wide variety o f settings. Turnover intentions to actual turnover has been reported as low as 0.32 (Carsten & Spector, 1987), or 0.36 (Hom, Caranikis Walker, Prussian, & Griffeth, 1992), whereas higher correlations have been estimated at 0.50 (Steele & Ovalle, 1984) and 0.52 (Tett & Meyer, 1993). Based on the recommendations provided for use of the surrogate variable intent to turnover, and the adequate degree of correlation found in the literature between turnover and intention to turnover, it is believed that intentions to withdraw will provide a withdrawal. Based on this body of research, it is expected that institutional experiences and student integration will influence inte ntions to withdraw. H1. The extent to which students perceive a supportive climate for retention negatively predicts intentions to withdraw. The more supportive a climate for retention is perceived, the less students intend to leave the program. H2. The relationship between climate and intentions to withdraw will be moderated by gender. Supportive climate perceptions will be more predictive of withdrawal intentions for women than men. H3. The extent to which students are socialized in their program will negatively predict withdrawal intentions. The more positively students are socialized, the less they intend to leave their program.
22 H4. The relationship between socialization and intentions to withdraw will be moderated by gender. Socialization will be m ore predictive of withdrawal intentions for women than men. Commitment Another indicator of persistence (or lack thereof) withdrawal is student commitment. This is influenced by institutional experiences and student integration, and is an influence on actual withdrawal. Tinto theorizes that positive experiences while in college reinforce ve likelihood of persisting. Student commitment has been a popular topic in educational literature ( Bean 1980, Cabrera, Castaneda, Nora, & Hengstler, 1992) In this contex t, the construct of organizational commitment used in organizational research can be applied here to contribute to a better understanding of student commitment. Organizational ide Porter, 1979). In other words, organizational commitment involves a strong belief in the concept has received a great deal of attention in the literature, and has been evaluated as both an antecedent and a consequence for a variety of work related variables. Organizational commitment has been conceptualized and operationalized in several ways, yet most share an underlying theme: commitment constitutes a bond or a link to the
23 organization, and those individuals who have a strong bond are less likely to leave their organization then individuals who have a weak bond (Allen & Meyer, 1990). The mos t commonly studied conceptualization of commitment has been that of attitudinal commitment, the type of commitment measured in the current study. Attitudinal commitment has been shown to exhibit a relationship with several work related variables; personal characteristics, job characteristics, and organizational environment are thought to be antecedents to commitment, while behavioral intentions and turnover have been studied as consequences of commitment. A meta analysis by Mathieu and Zajac (1990) explores the relationships between such variables. For example, a review of commitment studies showed a small correlation between gender and commitment ( r = 0.145), indicating that women tended to be slightly more committed to their organizations than men. Grusky (1966) proposed that women feel more committed to their organizations because they had to overcome more obstacles and barriers in order to gain membership. Although this hypothesis may be a little out of date for working women in general, this theory coul d apply to environments where women are still the overwhelming minority, such as in academic engineering programs. Attitudinal commitment also showed moderate correlations with perceived competence (r = 0.630); it appears that individuals will become commi tted to an organization to the degree it provides an opportunity for growth and for individuals to meet their goals. When considering the consequences of organizational commitment, turnover and turnover intentions are the most commonly studied variables. Commitment positively correlates with attendance (r = 0.102), and negatively correlates with lateness (r = 0.116) and turnover (r = 0.277). However, correlations with turnover intentions are much
24 stronger than with actual turnover; intention to search of alternative jobs (r = 0.599), intention to turnover (r = 0.464). The primary antecedents of affective commitment are organizational experiences and individual experiences (Allen and Meyer, 1997). Organizational experiences include factors such as auton omy, inclusion in the decision making process, supportiveness and fairness. Individual experiences include personal fulfillment, having rewarding experiences, and the nature of organizational practices. The degree of organizational and individual experienc commitment. With respect to student commitment to their academic program, Withey (1990) found that the most critical antecedents to affective commitment of an academic program were investment mad e in education (both in terms of energy expended and finances allocated), and the social environment of the academic program. Most research on the antecedents of student commitment consistently points towards two factors: social integration and faculty stu dent interaction. These factors are prominently featured in (Bean, 1980; Pascarella & Terenzini, 1980, Tinto, 1997). Based on this body of research, it is expected that institutional experiences and student integration will influence commitment. H5. The extent to which students perceive a supportive climate for retention positively predicts affective commitment to the program. The more supportive a climate for retention is perceived, the more committed students are to their program.
25 H6. The rela tionship between climate and commitment will be moderated by gender. Supportive climate perceptions will be more predictive of commitment to the program for women than men. H7. The extent to which students are socialized in their program will positively p redict commitment. The more students are socialized, the more committed they feel to their program. H8. The relationship between socialization and commitment will be moderated by gender. Socialization will be more predictive of commitment for women than m en. Fox, 1986; Pascarella & Terenzini, 1983) when it was tested in traditional environments, with traditional undergrad uate students pursuing an education is a variety of disciplines. analysis to confirm five variables which play a role in the prediction of student retention: a ) peer gr oup interaction, b ) interactions with faculty, c ) faculty concern for student development, d ) academic and intellectual development, and e ) institutional and goal commitments. These results supported the existence of and relationships among the major varia bles proposed in Tinto model. These results also support their previous conclusion that post matriculation experiences are more important than pre matriculation student characteristics when accounting for student commitment ( Terenzini & Pascarella, 1978 ). However, the current landscape of the student population has shifted towards a greater percentage of non traditional students and environments, and the research needs to progress in order to remain reflective of the current environment of student learning.
26 of non traditional undergraduate students in a management program, with an emphasis on the roles of social and acade mic integration. The researcher s here sought to tes t this model with a sample of students, on average 40 years old. They made assumptions about implying that the more alike the students were, the richer the quality of social interaction would be. Academic integration was measured by the degree to which students wanted to integration variable, they found that social, but not academic in tegration was a significant predictor of retention for their sample of non traditional students. However, this assessment was based on professional homogeneity within classes, an imprecise measure of integration. ed with graduate students, as well as the doctoral students at a traditionally Hispanic institution. Like the previous study, Vaquera chose to focus largely on the rol e of academic and social integration as predictors of attrition. In addition, the departmental climate (here, with an emphasis on racial climate) was also included in the model as a predictor of doctoral student persistence. R esults of this study indicate that the more negative the departmental racial climate was perceived, the less likely that student s would persist in the degree program. Additionally, the role of academic and social integration was significant in predicting persistence, although academic integration was found to be more important than social integration
27 Research by Lovitts (200 1) and Golde (2005) also highlight the importance of academic and social integration in the persistence of doctoral candidates. Both studies found that students who completed their degrees perceived a greater degree of integration than their non completer counterparts. Golde (2005) found that frequent interaction with faculty and advisors predicted persistence in the sciences, which often feature an apprenticeship element to graduate study. This interaction with faculty was a less significant predictor of d egree completion among the humanities, which is less centered on the mentor apprentice relationship. Research conducted on the retention of students in STEM majors based on on site intervention programs can offer another glimpse into the undergraduate experiences of students For example, Seymour and Hewitt (1997) conducted a three year study that examined the factors affecting undergraduate persistence in STEM majors. Interviews with undergraduate students revealed a number of common themes, such as di ssatisfaction with the faculty and a competitive environment among peers. Beyond the factors that affect all students in STEM, they sought to uncover the specific obstacles faced by women. Women talked about issues such as feelings of alienation from the faculty, uncertainty when asking faculty for help, developing a social network and feeling like part of a community. Women who were successful in STEM programs named the positive socialization with other women in their program as important to their success This s ocialization included mentoring programs, relationships with advisors and female study groups. These relationships were found to be more important to the women in the sample than the men. Seymour and Hewitt (1997)
28 concluded that the existence of su pport programs specifically geared toward women in involvement and integration in their program In another study based on an intervention program, Hyde and Gess Newsome (1999) described Project Access, a university based progra m for women in STEM majors. The found that the relationships with female peers served to encourage persistence, because they had the ability to identify with other women sharing similar experiences. Finally, a longitudinal study at the University of Washin gton was conducted to increase the retention of women and identify factors that impacted the retention of women in STEM (Brainard & Carlin, 1998). They found that among upper division students (juniors and seniors), acceptance by their department, faculty and peers was predictive of persistence. Perceived barriers specific to women included lack of self confidence, feeling of intimidated and isolated, and having a poor advising relationship. The results of these studies directly point to the importance of climate and socialization factors in the retention of STEM undergraduates, especially women. This corresponds with the predictions made by Tinto (1975) in his model of student withdrawal. withdrawal provides the theoretical foun dation for the research being proposed. It specifies both the personal and environmental factors associated with voluntary student withdrawal, and highlights the role of institutional experiences, socialization, and commitment in student withdrawal. By co nducting the proposed research, I hope to identify some of the reasons behind the disproportionate rates of withdrawal by female engineering undergraduates. It is believed that because
29 they are in the minority, the social environment, or in this case, the program climate, will have a greater impact on women than men. By examining the program climate, interpersonal environment, socialization processes, and commitment, it is hoped that a better understanding of gender specific issues will be gained. This has important implications for both colleges and their students, as they can use this information to give well informed advice and guidance to struggling female students, or focus on encouraging potential female engineers to enter into the sciences. The more t hat is known about the climate of engineering programs, and the relationship between program climate and a desire to withdraw, the better we can prepare its students for success.
30 Chapter Three : Method T his study was developed from a National Science Foundation grant entitled (NSF STEM Talent Expansion Program STEP II Award #0525408 ) funded by NSF. The purpose of this grant was to investigate the program culture and other environmental conditions that encourage the successful completion of undergraduate degrees in science, technology, engineering and mathematics by undergraduates, with a special emp hasis on underrepresented populations, such as women and minority group mem bers. This research project was conducted in two and four year public institutions of higher education in the State of Florida An inter disciplinary, mixed methods approach to dat a collection was taken, and included classroom observations, intervi ews, focus groups and surveys The goals of the study were to contribute to the existing knowledge base concerning education and STEM careers, and to contribute more broadly to the literature on organizational culture and climate and the ir influence on important higher education outcomes. One goal of the research grant was to identify the specific factors which create a climate for retention, and to develop a measure that can quantif y this climate. In order to create a survey that represented a well rounded explication of retention climate, research on career decision making and outcomes conducted by anthropologists, sociologists and psychologists was considered. These perspectives in volve different theoretical foundations (person centered on the psychological side, structural/organizational from
31 the sociological perspective and cultural from the anthropological perspective) and different research methods. When integrated, they provide complementary sources of information on STEM persistence E xtensive reviews of the literature in the disciplines of higher education, psychology, sociology, and anthropology were conducted. This information, combined with the content of student and facult y interviews completed for this purpose provided the basis for a list of elements thought to reflect a climate for student retention in higher education in general, and specifically a climate for student retention within STEM programs. In total, nine eleme nts of student retention climate were specified: Involvement, Faculty Support, Institutional Support, Helpfulness, Diversity, Integration, Fit, Engagement, and Importance. Items for these elements were written and the climate for retention survey was creat ed and piloted. The result was a 53 item scale measuring nine elements of STEM program climate. Survey a dministration Administration of the climate for retention survey began in January 2007 and ended in May 2008. Data were collected from undergraduate students in Engineering and Chemistry programs enrolled in nine four year institutions (both public and private) and four community colleges throughout the state of Florida. Students volunteered to take the survey, which took approximately 3 0 minutes to complete, and were paid $10 for their participation. Surveys were administered as a paper and pencil instrument, and responses were recorded on a s cantron form. Students were required to complete the survey in person, and response forms were checked for accuracy before the student was awarded their compensation. Concurrently, qualitative interviews and classroom observations were also conducted with students, faculty and key members of the program
32 administration. However, the proposed study f ocuses on a subset of these data, specifically data representing Engineering students at four year institutions. Participants Inclusion criteria In order to qualify for inclusion in the study, participants must be at least 18 years old, enrolled full or part time at one of the eight targeted universities, and be a registered engineering or pre engineering major. Sample characteristics This sampl e contains 1,421 students enrolled as undergraduate engineering or pre engineering majors. Approximately 22% of the sample was female, and the average age was 22 years old ( SD =3.4 years). The sample of students was predominantly White (47%), and 24% identi fied themselves as Latino/Hispanic, 16% Black, and 7% Asian/Pacific. Most of the students in the sample were in their junior or senior year (73.4%) and over one quarter of the students majored in civil engineering (27.3% ; s ee Table 1 Table 2, Figure 2). Comparison of the sample characteristics to national enrollment. The distribution of sample demographics closely matches the national enrollment figures for Fall 2007 as reported by the Interactive University Database. The gender and racial/ethnic distribu tions of undergraduate students at four year institutions were as follows: 20% female; 55% White, 26% Latino/Hispanics, 18% Black, and 7% Asian. Not only was the sample representative of national enrollment figures, it was also representative of the enrol lment figures for the university as a whole. T he proportion of females enrolled in each of the institutions in the sample closely approximated the overall enrollment distribution of that school, with the most female engineers enrolled in Florida State Univ ersity/Florida Agricultural and Mechanical University (FSU/FAMU,
33 Table 1. Demographic characteristics of the sample Overall Female Variable N % % Gender Male 1067 77.9 -Female 302 22.1 -Age (years) a 22.1 3.4 18 115 8.4 20.0 19 114 8.3 19.3 20 151 11.0 26.5 21 274 20.0 26.6 22 269 19.6 21.6 23 133 9.7 24.1 24 80 5.8 13.8 25 48 3.5 27.1 26 30 97 7.1 19.6 31 35 30 2.2 10.0 36+ 15 1.1 20.0 Ethnicity Caucasian/White 625 45.7 19.0 Latino/Hispanic 315 23.0 25.7 African American/Black 211 15.4 23.2 Asian/Pacific Islander 93 6.8 25.8 American Indian 8 0.6 37.5 Other 78 5.7 25.6 a. Mean SD Tallahassee 21.1% female), followed by the University of Florida (UF, Gainesville; 20.7%), Florida International University (FIU, Miami; 18.7%), the University of South Florida (USF, Tampa; 16.7%), the University of Central Florida (UCF, Orlando; 14.9%), and Florida Atlantic University (FAU, Boca Raton; 14.4%). Enrollment figures for Florida Institute of Technology ( FIT, Melbourne) and Embry Riddle Aeronautical University (ERAU, Daytona Beach) were not available.
34 Table 2 Educational characteristics of the sample Variables As mentioned, the climate for retention survey contains nine elements of program climate, as well as a measure of socialization, commitment, and questions regarding student withdrawal. Only a subset of the items which best factors for student retention were included in the study, and these are discussed in more detail Variable Overall Female N % % University FAMU/FSU 232 16.9 22.4 USF 213 15.6 23.5 UF 207 15.1 29.5 UCF 188 13.7 21.8 FIU 180 13.1 22.8 FAU 138 10.1 17.4 FIT 122 8.9 17.2 ERAU 55 4.0 12.7 Grade Level 1 st year 142 10.4 18.3 2 nd year 127 9.3 19.7 3 rd year 299 21.8 27.4 4 th year 420 30.7 22.6 5 th year 289 21.1 17.6 Major Aerospace 77 5.6 10.4 Civil 380 27.8 24.2 Electrical 177 12.9 11.3 Mechanical 251 18.3 11.2 Chemical 88 6.4 45.5 Computer 128 9.3 16.4 Environmental 21 1.5 47.6 Industrial 73 5.3 45.2 Biomedical 38 2.8 52.6 Ocean 30 2.2 10.0 Other 63 4.6 33.3
35 Figure 2. Participant enrollment by major Involvement. academ ic life was developed using five items anchored on a 5 point Likert type scale (1 = strongly disagree to 5 = strongly agree; items 1 5). Students indicated their level of agreement with a set of items capturing faculty availability and help to students, re Faculty support Faculty support was meas ured with four items for which students were asked to indicate, on a 5 point Likert type scale (1 = strongly disagree to 5 = strongly agree; items 6 9), their level of agreement with items that captured the types of assistance provided by faculty and staff to help students master knowledge in their 0 50 100 150 200 250 300 350 400 Men Women
36 discipline and develop creative capacities. An example of an item in this scale was Helpfulness. Six items were used to as sess the extent to which students perceived members of the department were helpful by indicating their level of agreement on 5 point Likert type response scale (1 = strongly disagree to 5 = strongly agree; items 10, 15 16, Diversity. The extent to which students perceived members of their department embrace diversity was captured by 9 items. Students were required to indicate their level of agreement on 5 point Likert type response scale (1 = strongly disagree to 5 = strongly agree; items 11, 14, 21, 25 27, 29 30, 33) to statements regarding what happens in the department with respec t to cultural and gender diversity. Integration. Four items captured the extent to which students perceived they were integrated in their department by having them indicate, on a 5 point Likert type scale (1 = strongly disagree to 5 = strongly agree; it ems 13, 19, 22, 31), their level of Socialization. An adaptation of the new comer socialization scale by Chao, Kelly, Wolf, Klein & Gardner, 1994) was used to examine the extent to which items to better reflect the university environment point Likert type scale (1 = strongly
37 disagree to 5 = strongly agree; items 6 1 72). The measure includes six dimensions of socialization, and items from three of these factors were included in the climate for retention survey. The three factors chosen (People, 4 items; Organizational Goals and Values, 4 items; Performance Proficiency, 4 items) were selected beca use they best Commitment. An adaptation of the affective commitment scale (Allen & Meyer, 1990) was included to determine the extent to which students felt a sense of commitment to their program. Minor Students responded to four questions reflecting their degree of commitment on a 5 point Likert type response scale (1 = strongly disagree to 5 = strongly agree; items 54 57). This set of questions represents the proximal outcome in t withdrawal, which is student commitment. Intent to withdraw opportunity to enroll in the same degree program at a different but equally ranked ould (a) definitely maintain enrollment at their university, (b) probably maintain enrollment at their university, (c) no opinion, (d) probably enroll at the alternative university, or (e) definitely enroll at the alternative university. Respons es to this item were coded such that lower endorsed values represent a desire to stay at the current program, and higher endorsed values indicate a desire to withdraw from the current program. This question represents the distal outcome of the Tinto model, which is student withdrawal from the program.
38 Here, however, the student is being asked about hypothetical intentions to withdraw which allows a response independent of outside constraining factors, such as finances or location of the institution; this e ncourages the student to only consider factors related to the quality of their experiences within the department when reflecting on their intentions to withdraw. Background information. In addition to the climate for retention survey, students also comp leted a background information sheet which contained general demographic questions, as well as information regarding their choice of major, time in school, financial status and housing information. Gender was dummy coded for analyses as male = 0, female = 1.
39 Chapter Four : Results The goal of the study was first to examine the climate for retention scale and determine what dimensionality is represented by the scale content. After the dimensions of climate for retention were defined, hierarchical multiple regression was used to determine what relationships exist between the major variables of the study: climate for retention, socialization, commitment, and intent to withdraw. Preliminary a nalysis Data s creening All variables were examined for missing values and accuracy of data entry In the event of missing data, a determination was made whether to delete a participant from the analysis or replace the missing value(s). Participants who were missing more than 10% of their data poin ts (seven items) were removed from the data set (n=5 2 ) leaving a final sample of 1369 students. The remaining missing values were to that of the item with the missi ng data point. The use of scantron technology reduced the chance of data entry errors, and the range of values for each item was also examined to check for impossible values (i.e., no values of 6 on a five point response scale). No issues regarding accurac y of data were found. Data from the hand entered background information sheet was also reviewed to ensure no errors were made during data entry. Assumption t esting The data were evaluated to determine if they met the assumptions for regression and factor analysis Regression requires linearity of the relationship between dependent and independent variables, independence of errors,
40 normality of dependent variabl es, normality of residuals, and homoscedascticy of residuals In addition to the assumptio ns for regression, f actor analysis also requires a substantial sample size and adequate factorability among the data. The relationships specified in all hypotheses are anticipated to be linear in nature, and the literature does n ot suggest that potential polynomial or other non linear relationships should be expected. Independence of errors was also not expected to be problematic in this sample of data. Independence was verified by referring to the Durbin W atson coefficient for each regression equation F or a sample of this size the de sirable values for the Durbin Watson autocorrelation coefficient should fall between 1.9 and 2.1. In each case, the values fell close to or within this range, generally indicating independence of data. Normality of dependent variables was evaluated by visual inspection of histograms and boxplots, as well as calculating skewness and kurtosis values. The variable commitment appeared to be normally distributed, while the variable inten t to withdraw was positively skewed. This is believed to be due to the nature of the variable itself, given that intentions to withdraw from the current academic program are ins titution. Normality of residuals was evaluated by visual inspection of histograms of the standardized residuals The residuals from regressions predicting commitment were normally distributed whereas the residuals from regressions predicting intent to wit hdraw were very slightly negatively skewed Finally, homoscedasticit y of residuals was evaluated by plotting the standardized residuals against the standardized predicted values. The distribution of residuals appears to be random, providing evidence of hom o scedastic error. Although not all of the assumptions were completely met,
41 regression analyses tend to be robust to minor violations of normality ( Tabachnick and Fidell 1996) so the analysis proceeded using the original data Factor analysis generally requires a large set of data, and while there is no exact formula for determining what the appropriate sample size should be, several recommendations, or rules of thumbs, have been have been made with respect to the relationship b etween the number of items to be analyzed and the size of the sample. The number of participants in the proposed study well exceeds 1,000, and it is expected that no matter what rule of thumb is applied, the sample size will be more than sufficient for thi s type of analysis. Statistical analyses were conducted to check the factorability of the data The Kaiser Meyer 2 = 10651, p <.05) were calculated. A KMO value of well over 0.5 and large significant chi square value resulting from the Bartlett test indicate that items share common factors, providing support for factor analysis (Kim & Mueller, 1978). Factor a nalysis The Climate for Retention scale was written with the goal of addressing a number of dimensions theorized to be evidence of a climate for retention. To determine if the climate for retention scale reflects a multidimensional structure or one general factor, an exploratory factor analysis (EFA) was conducted. EFA can be used to identify the underlying structure of a relatively large set of variables. Because the climate scale was written to assess a number of dimensions, but this dimensionality was never tested, there are no justifiable priori assumptions abo ut specific items associated with any predicted factor to be made. The factor loadings found as a result of the EFA guided the factor
42 structure of the data. The goal of the factor analysis was to determine the dimensionality of the climate for retention sc ale, and to provide a foundation for creating composite scores for each of the factors. The EFA was conducted with a principal axis factoring (PAF) extraction, to maximize the variance extracted by the factors. PAF is frequently used to explain the const ructs accounting for the variance of a measure, and was selected instead of principle components analysis for this analys i s because it is most useful for identifying latent variables rather than simply reducing the number of items (Preacher & MacCallum, 20 02). Oblique rotation was chosen to allow for the factors to be correlated with each other, because it is reasonable to believe that the various elements represented in the climate for retention scale would be related to some degree. Tabachnick and Fidell (200 6 p. 646) suggest looking to the factor correlation matrix for a confirmation that an oblique solution is appropriate. If the factors show a correlation of 0.3 or higher, there is enough shared variance among the factors to support the decision to use an oblique rotation. If the correlations are smaller, however, it may be better to use an orthogonal rotation instead. The intercorrelation between the factors was moderate (ranging between r = .39 and r = .49), indicating that the oblique rotation was appropriate. The factor structure was evaluated using several different indicators, because it is best when the identification of the factors do not rely on any single indicator. Pattern coefficients eigenvalues and the scree plot were all examined to det ermine the final factor structure of the climate for retention scale. In the first step of the factor analysis, all 28 of the climate for retention items described above were included, and the Kaiser rule was followed, which recommends retaining all factor s with eigenvalues greater than one
43 (Kaiser, 1960). This resulted in five factors explaining 50.2% of the variance. This factor structure, however, was not interpretable, and had only one item loading on the fifth factor. In the next steps, a four factor m odel and a three factor model were forced on the data, and explained 46.2% and 41.3% of the variance, respectively. These factor structures, however, resulted in problematic cross loadings and somewhat uninterpretable item distributions among factors. In t he final step, five items were removed due to low communalities and low pattern coefficients and the factor analysis was re run with the 23 remaining items. A four factor model was requested from the data, which explained 52.2% of the variance, and result ed in a simple structure with interpretable factors. The final four factor model had i tem communalities that were generally strong; no item had a communality below .20, and no items loaded below .35 on their respective factor, or had a multiple factor load ing value greater than .30. The four factors derived from the previous analysis represent four distinct dimensions of a climate for retention. After considering the common themes among items within each factor, the resulting factors were named Faculty Sup port, Sense of Community, Encouraging Diversity, and Valuing Diversity. The first factor, Faculty Support, contains nine items (survey items 1, 3 9, 18), and accounted for 29.5% of the variance. It reflects ideas about the extent to which faculty seemed su pportive of students and encouraging of their educational and professional goals. The second factor, Sense of Community, contains five items (items 10, 13, 19, 22, 31), and accounted for 9.5% of the variance. It represents the degree to which students are perceived to work together and share strategies for success with each other. The final two factors represent behavioral and attitudinal elements of diversity. The factor Encouraging Diversity contains four
44 items (items 27, 29 30, 33), and accounts for 7.1% of the variance. It represents behaviors which are thought to be evidence of supporting diversity, such as the absence of discriminatory behavior, and the perception that all students are treated equally. The final factor, Valuing Diversity, contains five items (items 11, 14, 21, 25 26) and accounts for 5.9% of the variance. It represents values that are supportive of diversity, such as prioritizing the creation of a diverse environment and respecting cultural differences (see Table 3) After the factor s tructure of the scale was defined, factor scores were calculated for use in the subsequent regression analyses. A summative score was calculated to represent the scale score for each of the four climate factors. Faculty Support is represented by nine items so the range of possible scores on that factor ranges from a minimum of 9 to a maximum of 45. The mean rating was 3 1.61 ( SD = 5.62 ), and the reliability of this factor was the highest of all the climate factors, = 0.85. Sense of Community is represente d by five items, so the range of possible scores on that factor ranges from a minimum of 5 to a maximum of 25, with an average rating of 19.66 ( SD = 3.55). The reliability of this factor was the second highest of the climate factors, = 0.80. Encouraging Diversity is represented by four items, and a range of possible scores from a minimum of 4 to a maximum of 20. The average rating of Encouraging Diversity was 15.62 ( SD = 3.18) and a reliability of = 0.76. Valuing Diversity is represented by five items, and a range of possible scores from 5 to 25. Valuing Diversity had an average rating of 17.91 ( SD = 3.21) and a reliability of = 0.72.
45 Table 3. Pattern coefficients for the c limate for r etention survey Climate for Retention Survey Items Factor 1 2 3 4 8. Faculty and staff go out of their way to help students master the knowledge in their discipline .758 .030 .134 .012 5. Faculty are enthusiastic about teaching .717 .021 .087 .106 6. Faculty and staff provide students with strong academic and professional role models .694 .000 .026 .029 9. Faculty and staff are generally encouraging towards students .635 .049 .108 .052 7. Faculty and staff help students develop creative capacities .624 .031 .056 .078 3. Faculty and staff help students achieve professional goals .562 .019 .012 .064 4. Faculty and staff seem to take responsibility for students' success .550 .030 .061 .062 1. Faculty and staff are often available for students to see outside of regular office hours .496 .047 .013 .026 18. Faculty or staff will offer help to a student who appears lost or confused. .402 .082 .099 .071 19. Students share strategies for success .035 1.067 .044 .146 13. Students often work together on team projects (e.g., research projects) .029 .712 .022 .003 10. Current students try to make new students feel included .046 .659 .063 .032 22. Students often learn from each other .080 .357 .089 .293 31. There is an overall sense of community among the students .129 .354 .165 .233 30. Instructors treat all students the same, both men and women .009 .038 .915 .159 29. Instructors treat all students the same regardless of their race or ethnicity .013 .032 .833 .083 26. Special efforts are made to help racial and ethnic minority students feel like they "belong" .084 .054 .515 .022 33. I have never observed discriminatory behaviors (e.g., words or gestures) directed towards female students .012 .055 .474 .056 21. People value diversity .038 .029 .018 .782 11. It is a priority to create a diverse, multicultural environment .012 .016 .143 .616 27. The perspectives of men and women are equally valued .083 .018 .172 .571 25. There is a genuine concern for the needs of students of all races and ethnicities .102 .053 .109 .467 14. People show respect for cultural differences .039 .125 .266 .380
46 In addition to the four climate factors, two other variables were included in the subsequent analysis: socialization and commitment. Summative composite scores and reliabilities for the socialization and commitment scales were also calculated. As discussed previously, each of the climate factors were moderately correlated with one another. The la rgest correlation among climate factors was that of Sense of Community with Valuing Diversity ( r = 0.49). Commitment and socialization also exhibited moderate correlations with each of the climate factors, and intent to withdraw displayed an inverse relati onship with all other study variables, as predicted. That is, the greater the perceptions of a supportive climate, positive socialization and strong commitment, the less students were inclined to report intentions to withdraw from the program. The correlat ions among study variables as well as the psychometric properties of all scales were calculated (see Table 4 ) Table 4: Correlations among major study variables Variable M SD FS SC ED VD Social Commit FS 31.62 5.62 (.846) SC 19.66 3.55 .398 (.799) ED 15.62 3.18 .410 .390 (.759) VD 17.91 3.21 .447 .489 .438 (.715) Social 45.26 6.15 .408 .429 .319 .348 (.781) Commit 13.48 3.09 .453 .366 .292 .337 .570 (.720) Withdraw 2.33 1.207 .276 .172 .194 .164 .304 .316 Note: All correlations are significant at p <.01 Scale reliabilities appear in parentheses on the diagonal FS= f aculty s upport, SC= s ense of c ommunity, ED= e ncouraging d iversity, VD= v aluing d iversity Social=socialization, Commit=commitment, Withdraw= i ntent to w ithdraw. Multiple r egression One of the major goals of this study was to determine what differential effects of program climate and socialization had on men as compared to women. Therefore, before
47 the regression analyses were undertaken, some preliminary investigations regarding gender differences were done. Gender differences Independent t tests were used to test for gender differences in each of the four climate factors, socialization, commitmen t, and intentions to withdraw. Regarding climate for retention factors, significant gender differences were found on the climate dimension of f aculty s upport, t (1367) = 2.70 p <.05, with men reporting higher levels of faculty support ( M =31.82, SD =5.72) than women ( M =3 0.88 SD =5.22). Significant gender differences were also found on the climate dimension of s ense of c ommunity, t (1367) = 2.6 3 p <.05 (this alpha level was used throughout the entire analyses) with women perceiving a greater sense of communi ty ( M =20.13, SD =3.53) than men ( M =19.52, SD =3.55) and on the climate dimension of e n couraging d iversity, t (1367) = 3.99, p <.05, with men reporting higher levels of encouraging diversity ( M =15.8 1 SD =3.11) than women ( M =14.95, SD =3.32). However, no gender d ifferences were found on the climate dimension of v aluing d iversity, t (1367) = 1.15, ns. Gender differences on the degree to which students are socialized in their program were also evaluated using an independent t test. Significant differences between genders were found, t (1367) = 4. 45 p <.05, with men reporting higher degrees of socialization ( M =45.6 7 SD =6.00) than women ( M =43.82, SD =6.46). Gender differences in the degree to which students felt committed to their department were found to be signific ant, t (1367) = 6.12 p <.05, with men reporting higher levels of commitment ( M =13.75, SD =3.04) than women ( M = 12.52 SD program were evaluated using an independent t test. Women reported a greate r intention to withdraw from their program ( M = 2.98 SD =0. 96 ) than men ( M = 2.15 SD =1.21),
48 t (1367) = 12 .53, p <.05. Recall that endorsing higher values on this item represent a desire to withdraw from program, while endorsing lower values indicate a desire to stay at the current program (see Table 5, Figure 3). Table 5. T tests of m ajor study variables by g ender Variables M SD t d F aculty S upport 2.70** .172 Men 31.82 5.72 Women 30.88 5.22 S ense of C ommunity 2.63** .173 Men 19.52 3.55 Women 20.13 3.52 E ncouraging D iversity 3.99*** .267 Men 15.81 3.11 Women 14.95 3.32 V aluing D iversity 1.15 .075 Men 17.85 3.22 Women 18.09 3.15 Social ization 4.45*** .297 Men 45.67 6.00 Women 43.82 6.46 Commit ment 6.12*** .402 Men 13.75 3.04 Women 12.52 3.08 Intent to Withdraw 12.53*** .760 Men 2.15 1.21 Women 2.98 0.96 Note. p<.05; ** p<.01; *** p<.001 To test for the existence of multicollinearity among the predictor variables, the tolerance statistic was calculated (Morrow Howell, 1994) Small tolerance values are problematic, and those values below 0.1 are considered to be indicative of multicollinearity among the predictor variables. In each case, tolerance values were between 0.9 and 1.0.
49 Figure 3. Ratings of study variables by gende r Hypothesis testing H ierarchical multiple regressions were conducted to evaluate the contribution of climate factors and socialization to student commitment and intentions to withdraw. The same process was followed to test each of the eight hypotheses. In the first step of the regression, the predictor variable(s) of interest were entered as one block. This step addresses hypotheses 1, 3, 5 and 7. Gender is added to the second step of the equation, to determine what influence it has on the prediction of the outcome of interest. The third step of the regression adds the interaction between gender and the predictor variable(s) in one block. The interaction terms were created by multiplying gender and the predictor variable. This step addresses hypotheses 2, 4, 6 and 8. Hypothesis 1 and 2 To test the first hypothesi s that a supportive climate for retention negatively predicts intentions to withdraw, intent to withdraw was regressed on the climate for retention factors All four climate factors were entered in the first block. 0 5 10 15 20 25 30 35 40 45 50 Faculty Support Sense of Community Encouraging Diversity Valuing Diversity Socialization Commitment Intent to Withdraw Men Women
50 The first step of the regression analyses revealed that the block of climate factors accounted for a sign ificant amount of variance (8.6 %) in intenti ons to withdraw, F (4, 1364 ) = 32.09 p <.05. A further inspection of the individual predictor s in the first block revealed that faculty s upport and sense of community significantly predicted intentions to withdraw, in the direction expected. Higher values on faculty s upport and sense of community predicted lower values on intent to withdraw. The corresponding b eta weights revealed that faculty su pport had more predictive influence on intentions to withdraw than did sense of community These results provide partial support for Hypothesis 1, indicating that some climate for retention factors negatively predicted intentions to withdraw Gender was added to the regression in a second block to understand how it influenced intentions to withdraw after the effects of climate were taken into account. There was a significant increase in the variance accounted for when gender was 2 =.073, p <.05 ) and the overall model was significant F (5, 1363)=51.73, p <.05. The positive beta weight associated with gender indicates that females displayed a greater intent to withdraw than males holding all other variables constant. In addition to gender, faculty support and sense of community remained significant predictor s of intentions to withdraw To determine if gender played a moderating role in the relationship between climate factors and intentions to withdraw, the third block of predictor variables were e ntered in the regression equation consisting of four gender by climate factor interaction terms. To test for significant moderator effects, the significance of the change in R 2 from the equation without the moderator s to the equation with the moderator terms was
51 calculated. Although t he third step of the regression model was signifi cant F ( 9, 1359 )= 29.53 p <.05 there was no significant increase in the amount of variance 2 =.004, ns ) Including gender as a moderator did not significantly increase the ability to predict intent to withdraw and no support was found for Hypothesis 2 (see T able 6 ). Table 6. Regression of i ntent to w ithdraw on c limate for r etention factors Variables b SE b R 2 R 2 f 2 Step 1 .086 .094 FS .047 .007 .220*** SC .017 .011 .050** ED .031 .012 .083 VD .002 .012 .005 Step 2 .159 .073*** .087 FS .042 .006 .195*** SC .029 .010 .086** ED .015 .011 .039 VD .010 .012 .027 Gender .830 .074 .276*** Step 3 .164 .004 .005 FS .035 .007 .164*** SC .032 .012 .095** ED .018 .013 .046 VD .018 .013 .047 Gender .711 .531 .244 FS x Gender .038 .016 .409 SC x Gender .013 .024 .093 ED x Gender .013 .026 .069 VD x Gender .045 .030 .282 Hypothesis 3 and 4 To test the third hypothesis that socializ ation will negatively predict withdrawal intentions i ntent to w ithdraw was regressed on s ocialization in the first block. The first step of the regression analyses revealed that s ocialization accounted for a significant amount of variance ( 9.2 %) in intent to withdraw, F ( 1 1 367 ) = 138 86 p <.05. There was a negative relationship between the two variables; higher levels of
52 socialization predicted lower ratings of intentions to withdraw. These results provide full support f or Hypothesis 3 Gender was added as the second block of the regression to understand how it influenced intent ions to withdraw after the effects of socialization were taken into account. There was a significant increase in the variance accounted for when gender was 2 =.063, p <.05) and the overall model was significant F (2, 1366)=125.54, p <.05. The po sitive beta weight associated with gender indicates that females displayed a greater intent to withdraw than males holding all other variables constant In addition to gender, socialization remained a significant predictor of intentions to withdraw. Beta weights revealed that socialization was slightly more influential in predicting intentions to withdraw than was gender. To test the moderating influenc e of gender on the relationship between socialization and intent to withdraw, a gender by socialization interaction term was entered in the third block of the regression. While t he overall regression equation was significant, F ( 3 136 5 ) = 83.64 p <.05, the re was no significant increase in the amount of variance explained ( 2 = .0 0 ns ), indicating that gender did not moderate the relationship between socialization and intentions to withdraw. Hypothesis 4 was not supported (see Table 7). Hypothesis 5 and 6 To test the fifth hypothesis that a supportive climate for retention positively predicts commitment, commitment was regressed on the four climate factors. All four climate factors were entered in the first block. The first step of the regres sion analyses revealed that climate factors accounted for a significant amount of variance ( 25.5 %) in commitment F ( 4 136 4 ) = 116.45 p <.05. All four of the climate
53 Table 7 Regression of i ntent to w ithdraw on s ocialization Variables b SE b R 2 R 2 f 2 Step 1 .092 .101 Social .060 .005 .304*** Step 2 .155 .063*** .075 Social .053 .005 .272*** Gender .053 .073 .253*** Step 3 .155 .000 .000 Social .053 .006 .270*** Gender .803 .511 .276 Social x Gender .002 .011 .023 Note. p<.05; ** p<.01; *** p<.001. factors positively predicted commitment which was in the direction expected. The corresponding Beta weights revealed that Faculty Support had the most predictive influence on commitment, followed by Sense of Community, with Valuing Diversity and Encouraging Diversity having the least amount of significant i nfluence on commitment. These results provide full support for Hypothesis 5, indicating that each of the climate for retention factors positively predicted commitment, which was in the expected direction. Gender was added as the second block of the regres sion to understand how it impacts commitment after the effects of climate were taken into account. There was a significant increase in the variance accounted for when gender was introduced, 2 =.024, p <.05) and the overall model was significant F (5, 1363) =105.03, p <.05. The negative beta weight associated with gender indicates that females reported lower levels of commitment than males holding all other variables constant With the addition of gender, all of the climate factors except encouraging diversit y remained significant predictors of commitment. To determine if gender had a moderating effect in the relationship between climate factors and commitment, a third block of variables were entered into the regression, consisting of four climate by gender interaction terms. The interaction terms
54 contributed a significant amount of additional variance ( 2 = .005 p<.05) and the overall model was significant, F ( 9, 1359 ) = 59.73 p <.05. Faculty support, sense of community, valuing diversity and gender remained significant predictors. Evidence for the moderating effect of gender on faculty support was also found. The nature of the moderating effect was such that faculty s upport was more predictive of commitment for women than men. However, there was no evidence of a significant moderating effect for the three remaining climate factors. These results provide partial support for Hypothesis 6, indicating that gender has a mod erating effect on one of the four climate factors ( see Table 8 ). Table 8 Regression of c ommitment on c limate for r etention factors Variables b SE b R 2 R 2 f 2 Step 1 .255 .342 FS .178 .015 .324*** SC .153 .024 .176*** ED .053 .027 .055* VD .079 .028 .082** Step 2 .278 .024*** .033 FS .170 .015 .310*** SC .171 .024 .196*** ED .029 .026 .030 VD .090 .028 .094** Gender 1.165 .175 .156*** Step 3 .284 .005* .007 FS .154 .016 .281*** SC .152 .027 .175*** ED .033 .030 .034 VD .110 .031 .114*** Gender 3.623 1.258 .486** FS x Gender .094 .039 .398* SC x Gender .088 .057 .242 ED x Gender .009 .062 .018 VD x Gender .116 .070 .287 Note. p<.05; ** p <.01; *** p<.001.
55 Hypothesis 7 and 8. To test the seventh hypothesis that socialization will positively predict commitment, c ommitment was regressed on s ocialization in the first block of the regression equation The first step of the regression analyses revealed that socialization accounted for a significant amount of variance (32. 5 %) in commitment, F ( 1, 1367 ) = 658.07 p <.05. These results provide full support for Hypothesis 7, indicating that socialization positively predicted commitment, which was in the expected direction. Gender was added as the second block of the regression to understand how it influences commitment after the effects of socialization were taken into account. There was a significant increase in the variance accounted for when gender was introduced, 2 =.009, p <.05) and the overall model was significant F ( 2, 1366 )= 342.26 p <.05. The negative beta weight associated with gender indicates that females reported lower levels of commitment than males holding all other variables constant With the addition of gender, socialization remained a significant predictor of commitment. To test for a significant moderating effect of gender on the relationship between socialization and commitment, a gender by socialization interactio n t erm was entered in the third block of the regression. Although t he overall regressi on equation was significant, F (3, 1365 ) = 228.79 p <.05 ) there was no significant increase in variance explained 2 =.001 ns ) This indicates that there was no sig nificant moderating effect of gender Hypothesis 8 was not supported (see table 9) Exploratory a nalyses In addition to testing the formally stated hypotheses presented in Chapter Two, additional analyses were conducted to further investigate the perceptio ns of climate and persistence indicators among student groups of theoretical and practical interest.
56 Table 9 Regression of c ommitment on s ocialization Variables b SE b R 2 R 2 f 2 Step 1 .325 .481 Social .286 .011 .570*** Step 2 .334 .009*** .014 Social .280 .011 .558 Gender .707 .166 .095 Step 3 .335 .001 .002 Social .272 .013 .542*** Gender 2.142 1.161 .287 Social x Gender .032 .026 .193 Note. p<.05; ** p<.01; *** p<.001. Independent t tests were conducted to determine if significant differences existed each of the four climate factors, socialization, commitment, and intentions to withdraw between students in majors with a large proportion of females versus a small proportion of females. Majors we re considered to have a high proportion of females if female students made up at least 25% of the students (see Table 2 for the percentage of females in each major) Of all the major study variables, significant differences were only found in two of the cl imate factors: Faculty Support, t (1173)=3.91, p <.05 and Sense of Community, t (1173)=2.13, p <.05 (s ee Table 10 Figure 4). A multivariate analysis of variance ( MANOVA ) was conducted to determine if there were differences in ratings of the climate variables, socialization, or persistence indicators by year in school. Results of this analysis indicate that there were significant differences found among the major study var iables, F (7, 1266)=14357, p <.05. Follow up univariate analyses revealed differences by year in school on the following study variables: Faculty Support, F (4, 1277)=4.98, p <.05; Encouraging Diversity, F (4, 1277)=2.80, p <.05; Valuing Diversity, F (4, 1277)=4. 31, p <.05; Socialization, F (4, 1277)=11.45, p <.05 and Commitment, F (4, 1277)=3.13, p <.05 (see Table 11, Figure 5).
57 Table 10. Climate r ating by m ajors with h igh and l ow f emale p articipation Variables M SD t d F aculty S upport 3.9 1 *** .230 32.35 5.47 31.07 5.67 S ense of C ommunity 2.13* .127 19.93 3.42 19.48 3.68 E ncouraging D iversity 0.2 2 .012 15.56 3.32 15.60 3.12 V aluing D iversity 0.7 4 .044 17.92 3.16 17.78 3.26 Social ization 1.6 5 .097 45.67 6.12 45.08 6.08 Commit ment 0.4 6 .029 13.57 3.18 13.48 2.99 Intent to Withdraw 0.20 .008 2.31 1.21 2.30 1.21 Note. p<.05; ** p<.01; *** p<.001. Figure 4. Climate r atings by m ajors with h igh and l ow f emale p articipation 0 5 10 15 20 25 30 35 Faculty Support Sense of Community Encouraging Diversity Valuing Diversity
58 Table 11. Ratings of m ajor s tudy v ariables by y ear in s cho ol Variable Year in School M SD Faculty Support 1 ST YR 33.38 5.54 2 ND YR 31.94 5.85 3 RD YR 31.36 5.66 4 TH YR 31.36 5.80 5 TH YR 30.93 5.32 Sense of Community 1 ST YR 19.72 3.16 2 ND YR 19.36 4.00 3 RD YR 19.63 3.65 4 TH YR 19.65 3.62 5 TH YR 19.87 3.23 Encouraging Diversity 1 ST YR 16.35 2.92 2 ND YR 15.40 3.32 3 RD YR 15.69 3.15 4 TH YR 15.35 3.32 5 TH YR 15.60 3.12 Valuing Diversity 1 ST YR 18.85 3.28 2 ND YR 17.92 3.42 3 RD YR 17.94 3.28 4 TH YR 17.70 3.11 5 TH YR 17.58 3.05 Socialization 1 ST YR 44.74 5.89 2 ND YR 42.81 6.26 3 RD YR 44.27 5.99 4 TH YR 45.62 6.38 5 TH YR 46.60 5.34 Commitment 1 ST YR 14.27 2.98 2 ND YR 13.43 2.99 3 RD YR 13.28 3.08 4 TH YR 13.34 3.17 5 TH YR 13.30 2.89 Intent to Withdraw 1 ST YR 2.11 1.09 2 ND YR 2.18 1.12 3 RD YR 2.32 1.14 4 TH YR 2.41 1.27 5 TH YR 2.38 1.26
59 Figure 5. Ratings of m ajor s tudy v ariables by y ear in s chool 0 5 10 15 20 25 30 35 40 45 50 Faculty Support Sense of Community Encouraging Diversity Valuing Diversity Socialization Commitment Intent to Withdraw 1st 2nd 3rd 4th 5th
60 Chapter Five : Discussion As discussed throughout this paper, it is clear that women have historically been deal of significant research has been conducted to determine what factors play a role in student retention in general in student retention in STEM fields, and specifically the experiences of underrepresented minorities in STEM. This research revealed myriad factors, ranging from role models and peer networks to quality of the instruction. This work has made an increasingly positive impact o n the population of students in question Although the number of women in STEM is growing, the distribution across majors is uneven and still low overall compared to men. This study focuses on how engineering students are influenced by the perceptions of their academic environment and the people in it, with a specific focus on how these factors impact female studen ts. By understanding a better understanding of how their gender affects their persistence and success within the science and engineering community can be gained Dimensionality of a climate for retention The fir st goal of this study was to describe an academic climate which encourages the retention of its students in STEM programs. Tinto theorizes that the role of student experiences is paramount in retention, and while many studies which followed Tinto included variables reflec ting student experience, they are varied in nature and inconsistently applied. Some research focuses on objective indicators such as grades and
61 test scores, others on participation in study groups and academically oriented clubs and organizations. Even tho se researchers who explicitly study academic climate operationalize it in a variety of ways. A n important first step of this study was to give some attention to the ways in which students described the climate of their STEM program, and to use that inform ation to identify the most important factors for retention Before any conclusions can be drawn about the influence of academic climate on its students, the nature of this academic climate must be better defined. T he factor structure of the Climate for R etention survey indicate s that there are four underlying factors associated with a climate for retention: f aculty s upport, s ense of c ommunity, e ncouraging d iversity and v aluing d iversity. Overall, these factors accounted for approximately half of the varia nce in climate scores. While this is not and their perception of the academic environment around them. Each of these factors is discussed in greater detail below. Faculty s upport Items that comprised the faculty support dimension of a climate for retention reflected ideas about faculty being involved with and supportive of their students. This factor includes items about program faculty being enthusiastic about teaching, going above and beyond to help students, acting as role models to help students achieve their pro fessional goals, an d being general ly encouraging towards students. Faculty support exhibited moderate positive relationships with all other climate dimensions, which indicates some overlap between this factor and other aspects of a climate for retention, while still remaining distinct from the other factors. With respect to the outcomes of interest, f aculty support exhibited a moderate positive relationship with
62 student commitment, and a small negative relationship with intent to withdraw. It is interestin g to note that faculty support was the climate factor with the strongest relationship to intent to turnover. Women perceived a lesser degree of faculty support than their male counterparts, indicating that they felt that the faculty was not as supportive, available and encouraging to them as they were to the male students. It is not surprising that the interactions between students and faculty are important to student success in college. Students can benefit even more by developing individual relationships with faculty members which go beyond the usual classroom or laboratory setting. These student faculty relationships need to be tended to by both parties, but it helps facilitate the deve lopment of the relationship if faculty members are approachable and available to their students (Kuh, Kinzie & Schuh 2010 p. 208 ). This openness on the part of faculty can take on many shapes, such as encouraging students to work on research projects wit h them, collaborating in a non academic activity like an academic club meeting, or having open and honest discussions about coursework. However, it is most important that faculty make themselves seem available and approachable to students. Interactions wi th faculty usually happen both inside and outside of the classroom, and neither should be overlooked. Interactions with faculty outside the classroom can encourage a dialogue that normally would not occur during class time, such as conversations about care er goals. They also allow for a more individual relationship with faculty by proving important face time with the instructor. Pascarella and Terenzini (2005) believe that time with faculty outside the classroom environment is consistently to promote student persistence, educational
63 for this: first, building relationships with faculty members help students to learn the norms and values of the institution, which is a critical component to effective institutions, and increase their commitment. Faculty members and other departmental figures maintain a unique perspect ive to the inner workings of their department, and have witnessed the strategies of successful students and the pitfalls of unsuccessful ones. This is vital information that students are not always privy to unless they begin to engage the faculty outside o f class and develop a relationship which would allow for the exchange of this and other helpful information. Although time with faculty outside the class room is beneficial the truth remains that for most students, the majority of exposure to faculty occu rs during class time. This environment and their evaluation of instructional quality ( Pascarella & Terenzini 2005) Additionally, faculty behavior during class can set the tone for encouraging or discouraging interactions outside the classroom. Tinto (1993) believes that the positive interactions with faculty outside of the class room originate with a supportive classroom environment. Pascarella and Terenzini (2005) believe that student perceptions of faculty support a nd approachability alone will encourage student persistence. By fostering the sense that faculty is interested in their student success, a supportive climate for students can be developed It is important for departments to both encourage faculty student i nteraction and to create a climate that fosters these relationships In addition to the obvious benefits
64 of students having a more interactive relationship with their faculty, the y will have access to knowledge about t he norms and values of the institution which promotes student socialization These interactions ultim ately improve the likelihood of student persistence (Bean, 2005). Sense of c ommunity Items that comprised the sense of community dimension of a climate for retention reflected ideas about a collaborative relationship among students who freely share knowledge with one another This factor included items about whether or not students learn from one another, share their strategies for success, make other students feel included and generally foster a sense of student community in the program. The sense of community factor and the faculty support factor both reflect the personal element of the academic environment; they are both focused on the ways in which other members of the department (whether it be faculty or other students) are welcoming, supportive and encouraging. Sense of community exhibited moderate positive relationships with all other climat e dimensions It also showed a moderate positive relationship to student commitment, and a small inverse relationship to intent to withdraw. The importanc e of a sense of community among the students is often mentioned as a critical component to student persistence and success. Tucker (1999) describes his any phenomena which made students feel a sense of belonging to the new ed ucational environment community and a network of peers for social support benefit in a number of ways. They are better able to face academic challenges (Tucker, 1999), to negotiate barriers to their success (Brain ard & Carlin, 1998), receive encouragement and develop a sense of deeper
65 social integration (Berger & Milem, 1999). D epartments or programs that are collaborati ve in nature rather than those which foster a sense of competition among students tend to be as sociated with increased success for all students, particularly female s and other under represented groups ( Smith, Gerbick, & Figueroa, 1997). Interestingly, women felt there was a greater degree of community among the students than did their male counter parts. This could be due to the fact that women, as members of an underrepresented group, joined together in classrooms, study groups and in academic clubs for support and encouragement. There are also a number of intervention programs targeted specificall y for women in engineering programs, which may also foster a sense of community that is more significant to women than it is to men. Men in the program may not perceive the sense of community among students to be as high because it is not a feature that is salient to them, or because they do not feel as much of a need to participate in classroom activities or study groups. In fact, one group of female engineering upperclassmen said the most important piece of advice they could give to women just entering th e program would be to join a study group (Vogt, Hocevar, & Hagedorn, 2007) Encouraging d iversity Items that comprised the encouraging diversity dimension of a climate for retention reflec ted behaviors that can be interpreted as supportive of a divers e learning environment. This factor included items reflecting the and treatment of underrepresented minor ities, specifying treatment towards both women and racial or ethnic groups. Encouraging diversity had the strongest pos itive relationship with the other diversity oriented factor, valuing diversity, and smaller positive relationships with the remaining climate factors, as well as
66 commitment. Encouraging diversity also had a small inverse relationship with intentions to wit hdraw. Sim ilar to faculty support, women perceived their program to be less en couraging of diversity than male students did. Valuing d iversity Items that comprised the valuing diversity dimension of a climate for retention reflected attitudes of people in the department towards underrepresented groups. These items focus not on overt discriminatory behavior (or lack thereof) but on the underlying values held by department members. This includes items about the degree to which people prioritize the creation of a diverse environment, value group membership. The encouraging diversity and valuing diversity factors both speak to the degree that a diverse environment is supported and prioritized, but each express that in a different way. It is important to examine both the attitudes and behaviors regarding diversity beliefs align to promote a supportive environment for students of all backgrounds. The valuing diversity factor exhibited similar relationships in strength and direction to the other climate factors, commitment and intent to withdraw as did encouraging diversity. One interesting discrepancy to note is that there were no differences between men and women on their perceptions of the extent to which their programs valued diversity. Recall that women rated their programs to be significantly lower on encouraging diversity, the dimension that reflects behaviors towards underrepresented groups, yet they believe that ion. Perhaps there is a disconnect between the services that are provided for women by their department and individual behaviors of students and other department members. For
67 example, a female student could value the funding and support provided by her dep classroom by her fellow male students, or perceive preferential treatment is being given to male students by the instructor. Another explanation could do with the ratings of enc ouragement given by male students. Because males are not the underrepresented group in engineering, they may be less sensitive to the behaviors of other department members towards women, and be more apt to believe that there is no differential treatment oc curring. Women, on the other hand, would be more likely to notice if they felt slighted by a fellow student or instructor, in turn lowering their ratings of equal treatment by department members. Gender differences in the perceptions of departmental attit udes towards diversity, and the existence of discriminatory behavior towards underrepresented groups are supported by a multitude of studies, especially research focused on women in STEM programs. Vogt, Hocevar and Hagedorn (2007) predicted that women woul d report more instances in which they felt discriminated against, either overtly or subtly, than men did in engineering programs. In fact, women felt that their male peers did not have as much respect for them as they did other male students, and that male students had an advantage over females in the program. Although explicit discriminatory behaviors and actions taken towards women may not be common there is evidence that women do experienc e more subtle methods of discrimination, such as being left out of group discussions or receiving less encouragement from their peers or instructors than their male counterparts Seymour and Hewitt (1997) echo this by warning that male attitudes can often be covertly derogatory towards women These attitudes can create a range of negative
68 situations for female students from feeling vaguely uncomfortable around classmates to obvious attempts to exclude them Peers can have a damaging effect on female students if they feel as though their contri butions are being marginalized, and such experiences only and self confidence One example of such subtle discrimination can be found in the creation of study groups. W hen students were allowed to choose their own groups for class assignments, they often gravitated towards groups members that were similar to themselves in race and gender. Because men are the majority of students in engineering and other STEM programs, this tendency often excludes women or other underrepresente d groups (Rosser, 1998). Furthermore, women tended to withdraw from group activities if they were the only female group member, and in engineering (and other male dominated fields) this trend is even more prominent (Vogt, et al., 2007). However, it is not just peers that can discourage female participation in STEM programs. Faculty and other departmental staff can also be influential in the female student experience. Instructors are a main component in establishing the classroom environment This can exert a great deal of choice to pursue an engineering degree, and subsequent interactions in and out of the classroom can impact their willingness to persist in their field, especially if it is in STEM (Zeldin & Pajares, 2000). Unfortunate ly, f aculty may or may not be aware of the crucial role they play in pursue and persist in engineering Many faculty members reject the notion that they are discriminatory towards their female students, but subtle biases are perceived by female students, often with discouraging consequences.
69 For example, one study found that 84% of women who switched out of STEM cited a lack of adequate advising by faculty and staff as a major factor in their decision to switch majors. C onversely, when faculty members were perceived to be interested and invested in the goals of their students, this had a far greater positive impact on female students than it did male students (Seymour & Hewitt, 1997) Similarly, Vogt, et al. (2007) noted that female engineering students recalled more interactions with faculty who were discouraging than did male engineering students. T here is both statistical and theoretical support for the four dimensions of a climate for retention described above : havin g a supporting faculty (faculty support) feeling a sense of community among students (sense of community) witnessing equal treatment of all students (encouraging diversity) and perceiving that the academic department values its diverse student populatio n (valuing diversity) Of course, there are m any other elements responsible for fostering a supportive climate for student retention which were not featured in the current study but the information presented here certainl y sheds light on some major areas in which an engineering program can help support its students through the ir journey towards graduation Predicting indicators of persistence among engineering undergraduates The first goal of the study was to describe the dimensions of a climate for rete perspectives on these dimensions as they apply to their undergraduate engineering department. The second goal of this study was to take this four factor model of climate that was described above, and use it within the framework of withdrawal to examine the role these factors, among others, play in the prediction of
70 student commitment and intent to withdraw. Tinto theorizes in his model of student withdrawal that institutional experiences (which are operationalized here as the dimensions of climate) and student integration (which is operationalized here as socializati on) ultimately predict student withdrawal. However, in his model, there are also proximal indicators of student persistence (of lack thereof) that are important consider, inc and their intentions to withdraw. The current stud y considered the predictive influence of a climate for retention and socialization on commit ment and intentions to withdraw, as well as the role of gender as a both a pre dictor and as a moderator of climate and socialization. Climate and intentions t o withdraw It was hypothesized that each of the climate for retention factors would negatively predict withdrawal intentions, based on institutional experiences lead to student persistence ( H ypothesis 1) If positive perceptions of a climate for retention predict student persistence, which is a positive outcome, then it follows that the same climate perceptions would have an inverse relationship with withdrawal intentions, which is a negative outcome. This is generally the pattern of results that was found Higher ratings o n faculty support and sense of community significantly predicted lower ratings of withdrawal intentions among engineering students with faculty support having a greater influence on withdrawal intent ions than sense of community However, the factors encouraging and valuing diversity were not significant predictors of withdrawal intentions. The role of gender was also of central interest in the prediction of withdrawal intentions. Adding gender to the model after the effects of climate were established approximately doubled the amount of explained variance in withdrawal intentions
71 Although not formally hypothesized, it was expected that gender would be a significant predictor of withdrawal intentions based on the retention literature and on the pattern of t test results discussed previously, and that female students would have higher reports of withdrawal intentions than male students would. This was the pattern of results found Also, wi th the addit ion of gender, the original two climate factors remained significant. Finally, the role of gender as a moderator of the climate withdrawal intentions relationship was examined ( H ypothesis 2). There was no evidence of a moderating effect of gender on any of the climate factors, although the original predictors from the second step of the model remained significant. This indicates that while climate factors such as faculty support and sense of community predict withdrawal intentions, t he se relationship s do no t differ for male and female students. Socialization and intentions to withdraw In addition to the climate for retention prediction of withdrawal intentions was also examined. Socialization plays an important role in the integration of a studen t to the academic department, because socialization represents the degree to which that student has gained knowledge about the people, the work and performance expectations, and the norms and values represented in that department (Chao, et al., 1994). It w as hypothesized that socialization would negatively predict withdrawal intentions, student integration would lead to student persistence (Hypothesis 3 ). If positive socialization predicts student persistence, which is a positiv e outcome, then it follows that the same degree of socialization would have an inverse relationship with withdrawal intentions, which is a negative outcome. This is precisely what the results indicated. Socialization was a
72 significant negative predictor o f intentions to withdraw, and accounted for a small percentage of the variance. When gender was entered into the model, even more variance in withdrawal intentions was accounted for. However, there was no evidence of an interaction between gender and socia lization in the prediction of withdrawal intentions (Hypothesis 4). Although socialization did significantly predict withdrawal intentions, th e strength or direction of this relationship did not differ between male and female engineering students. Climate and commitment It was predicted that each of the climate for retention positive institutional experiences lead to persistence (Hypothesis 5). This relationship w as expected to be in the positive direction, unlike the proposed relationships with withdrawal intentions, because as perceptions of climate and socialization increase T he data supported this hypothesis. Each of the four climate for retention factors were significant predictors of commitment, in the direction expected. Furthermore, these climate factors accounted for a far greater proportion of variance in commitment (approximate ly 25%) than they did in withdrawal intentions. When compared to each other, faculty support had the most influence on commitment, just as it did in the model for withdrawal intentions. Sense of community also had a comparatively large impact on commitment scores. Adding gender to the model as a predictor of commitment slightly increased the amount of explained variance, with female students reporting lower levels of commitment than male stud ents. It also rendered encouraging diversity an insignificant
73 predictor of commitment, but the other three climate factors (faculty support, sense of community, valuing diversity) remained significant. Finally, the moderating role of gender in the climate commitment relationsh ip was tested, and one significant interaction was found (Hypothesis 6). Although it added little to the amount of explained variance in commitment, there was evidence for the moderating effect of gender on the relationship between fa culty support and commitment. This relationship was such that faculty support was more predictive of commitment for female engineering students than it was for male students. Socialization and commitment It was hypothesized that socialization would posi lead to student persistence (Hypothesis 7 ). Similar to the relationships between climate factors and commitment, this relationship was predicted to be in the positive directi on, because greater levels of integration should lead to greater levels of commitment. The results of this analysis support th i s prediction. Socialization was a significant predictor of commitment in the direction expected and accounted for a lmost a third of the variance in commitment When gender was entered into the model, additional variance in withdrawal intentions was accounted for. However, there was no evidence of an interaction between gender and socialization in the prediction of commitment (Hypot hesis 8 ). Although socialization did significantly predict commitment the strength or direction of this relationship did not differ between male and female engineering students. There are several points that can be made about the pattern of resul ts found in this study. First, there are multiple dimensions of an academic climate whic h encourage retention. These climate dimensions represent interactions with faculty and peers both
74 generally and specifically in regard to encouraging and valuing diversity among the student body. The climate factors, along with soci alization play ed a significant role in predicting the indicators of persistence among engineering undergraduates, specifically, commitment and withdrawal intentions. These relationships mirror the model of student withdrawal described by Tinto. Second, some dimensions of climate are more important in predicting indicators of persistence than others. Perceiving the faculty to be supportive of their students was the most influential climat e factor in predicting both commitment and withdrawal intentions. Feeling as though there is a sense of community among students was also influential in predicting persistence indicators, while factors representing beliefs about diversity played the smalle st role in student outcomes. Third, t here was a general lack of evidence for the moderating role of gender. Although female students perceive the faculty in their department to be less supportive, and feel socialized to a lesser extent than their male coun terparts, this does not seem to account for gender differences in the indicators of persistence. One notable exception to this is the interaction of gender and faculty support in the prediction of student commitment. Finally, both the climate factors and s ocialization played a larger role in the prediction of commitment than they did on intentions to withdraw. Theoretical and p ractical i mplications One of the goals of this study was to describe which aspects of an undergraduate engineering environment foste r a climate for retention. R esults of this study provide evidence for the importance of faculty support, a sense of community among students, and the perception that a diverse student body is valued and encouraged by all members of the department This research brings a greater degree of specificity regarding the
75 individual components of an academic climate in a way that previous research has not. of situations which may discourage female participation in the sciences. Since then, the idea of a chilly climate has been conceptualized in a variety of ways, such as ussions and activities, making inappropriate comments about female students, or suggesting that female students are less committed, less prepared or less able to complete the coursework (Morris & Daniel, 2008). However, there has been no systematic examina tion of what factors interact to create this chilly climate Once a clear definition of what a climate for retention is (or it can be used to evaluate and compare a broader range of STEM and non STEM departments and programs. An obvious advantage of this is the ability to compare programs with different proportions of women or other underrepresented minorities, and programs with various levels of success in attracting and retaining wo men and minorities to see how they differ on the dimensions of climate. Defining the dimensionality of a climate for retention would also focus the development of intervention programs targeted at retention efforts. For example, from this research it is c lear that faculty support plays a critical role in both student commitment and withdrawal intentions. An intervention designed to generate more opportunity for relaxed, extra classroom interactions between faculty and students could foster the development of informal mentoring or advising relationships, provide socialization), and generally support student persistence.
76 Limitations of the s tudy The c ross sectional design of th is study does not allow for a causal interpretation of the relationships between climate, socialization, commitment and withdrawal intentions In Tinto he specifies that there is a longitudinal component. The assimilation of institutional experiences and integration into an academic department are phenomena that occur over time, not all at once upon entry into the program. Therefore, a longitudinal design would allow for an exam ination of this development as it occurs, and would also provide a way to measure changes in student perceptions over time. Nevertheless, even a cross sectional design such as this is an important first step towards an understanding of student experiences in STEM, because it first identifies what the important variables are, and can focus the efforts of subsequent and often more expensive longitudinal research. In addition, t he way in which student withdrawal was approximated and operationalized did no t reflect the ultimate outcome of interest, which is actual student withdrawal from the program This study rel ied on intentions to withdraw as an approximation of actual withdrawal, and the survey item reflecting this was designed in such a way that stude nts were asked to respond to a hypothetical situation of withdrawal. The intention of designing the question this way was to isolate the influence of program experiences in the decision to intend to withdraw, so that other confounding factors could be rule d out, such as geographic or financial restrictions. Therefore, the conclusions Although the climate for retention survey was grounded in actual student interviews as well as reflecting current lit erature on retention, t here may be a variety of
77 other variables that interact to create a climate for retention that were not reflected in the survey A broader conceptualization of a climate for retention could have accounted for mo re of the variance in persistence indicators than what was found in the current study Finally, the sample of students included in the study does not represent the entire population of undergraduates in STEM programs. Only engineering undergraduates i n four year Florida universities were included in the sample, so the results of this study cannot be interpreted and applied to other STEM majors. In addition, s tudent s for this study were recruited in classrooms in academic club meetings, and around camp us in locations where engineering students can commonly be found Although efforts were made to target females and other underrepresented minorities so a more demographically representative sample could be assembled, these recruitment strategies may have m issed other populations of students, such as those who work off campus, commute, or for whatever other reason do not spend much time on campus outside of attending class. Recruiting in academic clubs may have also disproportionately encouraged the particip ation of students who are highly involved in the social aspects of the program, and represent a different perspective on the climate of the program, or be more committed because of their engagement in extracurricular academic activities. Directions for f uture r esearch over the last few decades, continued research is needed to identify the factors that may gender, what barriers to persistence exist for students, and what intervention programs and
78 ent ire educational pathway, from elementary school to post baccalaureate study. backward to pre K 12 experiences and influences but also forward to undergraduate and graduate e Research building on the current study should seek to address the longitudinal component to student persistence, and examine the unfolding process es by which students become acclimated to their progra m, learn the norms and values of the department, develop relationships with faculty and peers, and then track their progress through graduation to capture those students who graduate and those who withdraw. The population of interest should also be expan ded to include universities in other geographic regions, other STEM majors in addition to engineering and other underrepresented populations besides gender, such as racial or ethnic minorities. In addition, these results could be compared to a sample of n on STEM majors, to disentangle the reasons behind student withdrawal from STEM and reasons behind withdrawal from any undergraduate program. Finally, this line of research should examine retention factors at the program and departmental level, as well as at the individual level. By aggregating individual level climate data, the degree of consensus among students about the academic climate of their department can be assessed. Program level information about the factors that encourage persistence can help gu ide policy and direct the efforts of intervention programs for successful retention of STEM students.
79 Conclusion At the beginning of this paper, two critical questions were posed about student experiences in college: a) Can students have different e xperiences while in the same institutional environment, and b) Does the college experience create different outcomes for various subgroups of students? This goal of this current study was to address both of odel of student withdrawal while using key concepts from industrial/organizational psychology applied to the undergraduate engineering environment. Overall, t he answer to both of these questions is yes. Male and female engineering students do have differe nt perceptions of their academic environment and these experiences can lead to different outcomes for men and women. Promoting gender diversity in STEM fields remains a key concern for our educational system, and research in this area p rovides an opportunity to further tease apart the reasons for these gender differences. The research presented here is a response to efforts towards increasing proportional distribution of women in STEM promotes women in general, and also has broader implications for the global competitiveness of the American workforce By persistence, we can better ensure the success of our future scie ntists and engineers.
80 References Aiken, L., & West, S. (1991). Multiple regression: Testing and interpreting interactions Newbury Park, CA: Sage Publications. Allen, N., Meyer, J. (1990). The measurement and antecedents of affective, continuance and normative commitment to the organization. Journal of Occupational Psychology, 63, 1-18. Ashar, H., & Skenes, R. (1993). Can Tinto's student departure model be applied to nontraditional students? Adult Education Quarterly, 43 (2), 90-100. Astin, A. (1970). The methodology of research on college impact, part one. Sociology of Education, 43, 223254. Baron, R., & Kenny, D. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51 1173-1182. Bauer, T. N., Morrison, E. W., & Callister, R. R. (1998). Organizational socialization: A review and directions for future research. Research in personnel and human resources management, 16, 149214. Bean, J. (2005). Nine themes of college student retention. In A. Seidman (Ed.), College Student Retention: Formula for Student Success (pp. 215-243). Westport, CT: Praeger Publishers.
81 Berger, J. B. (2001). Understanding the organizational nature of student persistence: Empirically based recommendations for practice. Journal of College Student Retention: Research, Theory and Practice, 3 (1), 3 21. Berger, J. & Milem, J. (1999). The role of student involvement and perceptions of integration in a causal model of student persistence. Research in Higher Education, 40 641 664. education. IEEE Technology and Society Magazine, 19 20 26. Bix, A. (2002). Equipped for life: Gendered technical training and consumerism in home economics, 1920 1980. Technology and Culture, 43 7 28 54. Bix, A. (2004) From "Engineeresses" to "Girl Engineers" to "Good Engineers": A history of women's U.S. engineering education. NWSA Journal, 16 27 49. Brady, K. L., & Eisler, R. M. (1999). Sex and gender in the college classroom: A quantitative anal ysis of faculty student interactions and perceptions. Journal of Educational Psychology, 91 127 145. Brainard, S.G. & Carlin, L. (1998). A longitudinal study of undergraduate women in engineering and science, Journal of Engineering Education 87(4). Bra xton, J. M., & McClendon, S. A. (2001). The fostering of social integration and retention through institutional practice. Journal of College Student Retention: Research, Theory and Practice, 3 (1), 57 71. Campion, M. A., & Mitchell, M. M. (1986). Manageme nt turnover: Experiential differences between former and current managers. Personnel Psychology, 39 (1), 57 69.
82 Carsten, J., & Spector, P. (1987). Unemployment, job satisfaction, and employee turnover: A meta analytic test of the Muchinsky model. Journal of Applied Psychology, 72 374 381. Chao, G. T., O'Leary Kelly, A. M., Wolf, S., Klein, H. J., & Gardner, P. D. (1994). Organizational socialization: Its content and consequences. Journal of Applied Psychology, 79 (5), 730 743. Clewell, B. & Campbell, P. Journal of Women and Minorities in Science and Engineering, 8 255 284. Cohen, J., & Cohen, P. (1983). Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. Constantinople, A., Cornelius, R., & Gray, J. (1988). The chilly climate: Fact or artifact? The Journal of Higher Education, 59 (5), 527 550. Crawford, M., & MacLeod, M. (1990). Gender in the college cl assroom: An assessment of the chilly for women. Sex Roles, 23 101 122. Dalton, D., Johnson, J., & Daily, C. (1999). On the use of "Intent to..." variables in organizational research: An empirical and cautionary assessment. SAGE Social Science C ollections, 52 ,(10), 1337 1350. Etzkowitz, H., Kemelgor, C., & Uzzi, B. (2000). Athena unbound: The advancement of women in science and technology New York: Cambridge University Press. Florida Board of Governors Interactive University Data. Tallahassee, FL: Retrieved from http://www.flbog.org/resources/iud
83 Fox, M. (2001). Women, science and academia: Graduate education and careers. Gender & Society, 15 654 666. Fox, M. (2010). Women and men faculty in academic science and engineering: Social organizational indicators and implications. American Behavioral Scientist, 53 997 1012. Fox, R. (1986). Application of a conceptual model of college withdrawal to disadvantaged students. American Educational Research Journal, 23 415 424. Freund, A. (200 5). Commitment and job satisfaction as predictors of turnover intentions among welfare workers. Administration in Social Work, 29 (2) 5 21. George, J., & Jones, G. (1996). The experience of work and turnover intentions: Interactive effects of value attainme nt, job satisfaction, and positive mood. Journal of Applied Psychology, 81 318 325. Golde, C. (2005). The role of the department and discipline in doctoral student attrition: Lessons from four departments. The Journal of Higher Education, 76 (6), 669 700. Grusky, D. (1966). Career mobility and organizational commitment. Administrative Science Quarterly, 10 488 503. Hall, R. M., & Sandler, B. R. (1982). The Classroom Climate: A Chilly One for Women? Washington, D.C.: Association of American Colleges. Hall, R. M., & Sandler, B. R. (1984). Out of the Classroom: A Chilly Campus Climate for Women? Washington, D.C.: Association of American Colleges.
84 Hansen, S. (2000). Gender, families, and science: Influences on early science training and career choices. Journal of Women and Minorities in Science and Engineering, 6 169 187. Hays, W. (1994). Statistics (5th ed.). Fort Worth: Harcourt College Publishers. Heller, J. F., Puff, C. R., & Mills, C. J. (1985). Assessment of the chilly college climate for women. The Jo urnal of Higher Education, 56 (4), 446 461. Herzig, A. H. (2004). Becoming mathematicians: Women and students of color choosing and leaving doctoral mathematics. Review of Educational Research, 74 (2), 171. Hom, P., Caranikas, W., Prussian, G., Griffith, R ., (1992) A meta analytical structural equations analysis of a model of employee turnover. Journal of Applied Psychology, 77 890 910. Hulin, C., Fitzgerald, L., & Drasgow, F. (1996). Organizational influences on sexual harassment. In. M.S. Stockdale (ed.) Sexual Harassment in the Workplace Thousand Oaks, CA: Sage. Hurtado, S., & Carter, D. (1997). Effects of college transition and perceptions of the campus racial climate on Latino college students' sense of belonging. Sociology of Education, 70 324 345. Hyde, Michelle S.& Gess Newsome, Julie (1999). Adjusting educational practices to increase female persistence in the sciences. Journal of College Student Retention, 14. James, L. & Brett, J. (1984). Mediators, moderators, and tests for mediation. Journal of Applied Psychology, 69 307 321.
85 Jordan, C., & Bilimoria, D. (2007). Creating a productive and inclusive work environment. In A. J. Stewart, J. E. Malley, & D. LaVaque Manty (Eds.), Transforming science and engineering: Advancing academic women (pp. 22 5 242). Ann Arbor: The University of Michigan Press. Kaiser, H. F. (1960). The application of electronic computers to factor analysis. Educational and Psychological Measurement 20 141 151. Kim, J. & Mueller, C. (1978). Factor analysis: Statistical methods and practical issues Beverly Hills, CA: Sage Publications. Kirschenbaum, A., & Weisberg, J. (1994). Job search, intentions, and turnover: The mismatched trilogy. Journal of Vocational behavior, 44 17 31. Klein, K.J., Cohn, A. B., Smith, D.B., & Sorra, J.S., (2001). Is everyone in agreement? An exploration of within group agreement in employee perceptions of the work environment. Journal of Applied Psychology, 86 3 16. Kuh, G., Kinzie, J., & Schuh, G. (2010). Student Success in College: Creating Conditions that Matter San Fransisco, CA: Jossey Bass. Lau, L., (2003). Institutional Factors affecting student retention. Education 124, 126 136. Lewin, K. (1951). Field theory in social science New York: Harper and Row. Lightbody, P., Siann, G., Tait L., & Walsh, D. (1997). A fulfilling career? Factors which influence women's choice of profession. Educational Studies, 23 (1), 25 37. Louis, M. R. (1990). Acculturation in the workplace: Newcomers as lay ethnographers. In Organizational climate and cu lture (pp. 85 129). San Francisco: Jossey Bass.
86 Lovitts, B.E. (2001). Leaving the Ivory Tower: The Causes and Consequences of Departure from Doctoral Study Maryland: Rowman and Littlefield Publishers, Inc. Mathieu, J. and Zajac, D. (1990). A review and meta analysis of the antecedents, correlates, and consequences of organizational commitment. Psychological Bulletin, 108 171 194. Miller, J., Wheeler, K., (1992) Unraveling the mysteries of gender differences in intentions to leave the organization. Jour nal of Organizational Behavior, 13 (5) 465 478. Mobley, W., Griffeth, R., Hand, H., & Megliano, B. (1979). Review and conceptual analysis of the employee turnover process. Psychological Bulletin, 86 493 522. Mone M., (1994). Relationships between self conc epts, aspirations, emotional responses, and intent to leave a downsizing organization. Human Resource Management, 33 281 299. Morris, L., and Daniel, L. (2008). Perceptions of a chilly climate: Differences in traditional and non traditional majors for wom en. Research in Higher Education 49, 256 273. Morrow Howell, N. (1994). The M word: Multicollinearity in multiple regression. Social Work Research, 18 247 251. Mowday, R., Steers, R., &Porter, L. (1979). The measurement of organizational commitment. Jo urnal of Vocational Behavior, 14 224 247.
87 National Science Board. (2007). National Action Plan for Addressing the Critical Needs of the U.S. Science, Technology, Engineering, and Mathematics Education System. Arlington, VA: National Science Foundation. Retrieved from www.nsf.gov/nsb/documents/2007/stem_action.pdf National Science Foundation, Division of Science Resources Statistics. (2009). Women, Minorities, and Persons with Disabilities in Science and Engineering: 2009 Arlington, VA: National Scienc e Foundation. Retrieved from http://www.nsf.gov/statistics/wmpd Naumann, S. E., & Bennett, N. (2000). A case for procedural justice climate: Development and test of a multilevel model. The Academy of Management Journal, 43 (5), 881 889. Noel, L., Levitz, R., Saluri, D., & Marchese, T. J. (1985). Increasing student retention San Francisco: Jossey Bass. Oldenziel, R. (1999). Making Technology Masculine: Men, Women, and Modern Machines, 1870 1945 Amsterdam: Amsterdam University Press. Ostroff, C., Kinick i, A. J., & Tamkins, M. M. (2001). Organizational culture and climate. In Comprehensive handbook of psychology: Industrial and organizational psychology (Vol. 12, pp. 565 593). Hoboken, NJ: John Wiley & Sons, Inc. Ostroff, C., Kinicki, A., and Tamkins, M. (2003). Organizational Culture and Climate. In Borman, W., Il gen, D., Kilmoski, R., & I. Weiner (Eds.). Handbook of Psychology, Volume 12: Industrial Organizational Psychology Hoboken, NJ: Wiley.
88 Pascarella, E. T. (1985). College environmental influences on learning and cognitive development: A critical review and synthesis. In Higher education: Handbook of theory and research (Vol. 1, pp. 1 61). New York: Agathon Press. Pascarella, E., & Chapma n, D. (1983). Validation of a theoretical model of college withdrawal: Interaction effects in a multi institutional sample. Research in Higher Education, 19 25 48. Pascarella, E., Terenzini, P., (1983). Predicting voluntary freshman year persistence/withd rawal behavior in a residential university: A path analytic validation of Tinto's model, Journal of Educational Psychology 75 (2), 215 226. Pascarella, E. T., & Terenzini, P. T. (2005). How college affects students [vol. 2]: A third decade of research. San Francisco: Jossey Bass. Pascarella, E. T., Whitt, E. J., Edison, M. I., Nora, A., Hagedorn, L. S., Yeager, P. M., & cognitive outcomes during the first year of college. Journa l of College Student Development, 38 (2), 109 110. Journal of College Student Retention, 1 163 175. Purcell, C., (1979) Toys, Technology, and Sex Roles in America, 1920 1940. In M. Moore (Ed.). Dynamos and Virgins Revisited: Women and technological Change in History Metuchen, NJ: Trescott Reichers, A., & Schneider, B. (1990). Climate and culture: An evolution of the constructs. In B. Schneider (Ed.), Organizational climate and culture (pp. 5 39). San Francisco: Jossey Bass.
89 Rosin, H., Korabik, K., (1995). Organizational experiences and propensity to leave: A multivariate investigation of men and women managers. Journal of Vocational Behavior, 46 1 16. Rosser, S. (1998). Group work in science, engineering, and mathematics. College Teaching, 46 82 88. Sandler, B. R., Silverberg, L. A., & Hall, R. M. (1996). The chilly classroom climate: A guide to improve the education of women (pp. 1 127). Washington, D.C.: National Association for W omen in Education. Schneider, B. (1990). The climate for service: An application of the climate construct. In Organizational climate and culture (pp. 383 412). San Francisco: Jossey Bass. Schneider, B., Bowen, D. E., Ehrhart, M. G., & Holcombe, K. M. (2 000). The climate for service: Evolution of a construct. In Handbook of organizational culture and climate (pp. 21 36). Thousand Oaks, CA: Sage Publications, Inc. Sch n eider, B., Gunnarson,S., & Niles Jolly, K. (1994). Creating the climate and culture of success. Organizational Dynamics, 23 17 29. Seymour, E., & Hewitt, N. M. (1997). Talking about leaving: Why undergraduates leave the sciences (pp. 1 138). Boulder, CO: Westview Press. Sidel, R., (1994). Battling bias: The struggle for identity and community on college campuses New York: Viking. Sjoberg, A., &Sverke, M. (2000). The interactive effect of job involvement and organizational commitment on job turnover revisited: A note on the mediating role of turnover intention, Scandinavian Journal of Psychology, 41 247 252.
90 Smith, D. G., Gerbick, G. L., Figueroa, M. A., Watkins, G. H., Levitan, T., & Moore, L. C. (1997). Diversity Works: The Emerging Picture of How Students Benefit (pp. 1 159). Washington, D.C.: Association of American Colleges and Universities. Society for Women Engineers. (2009). The SWE Story: Timeline of Achievement Retrieved f rom: h ttp://societyofwomenengineers.swe.org/index.php?option =com_content&task=view&id=62&Itemid=66 Steele, R., & Ovalle, N. (1984) A review and meta analysis of research on the relationship between behavioral intentions and employee turnover. Journal of Applied Psychology, 69 673 686. Stockard, J., & McGee, J. (1990). Children's occupational preferences: The influence of sex and perceptions of oc cupational characteristics. Journal of Vocational Behavior, 36 (3), 287 303. Tabachnick, B & Fidell, L. (2006). Using Multivariate Statistics (5th ed.). Boston: Allyn and Bacon. Tett, R., & Meyer, J. (1993). Job satisfaction, organizational commitment, t urnover intention, and turnover: Path analysis based on meta analytical findings. Personnel Psychology, 46 259 293. Tinto, V. (1975). Dropout from higher education: A theoretical synthesis of recent research. Review of educational research, 45 (1), 89 125. Tinto, V. (1993). Leaving college: Rethinking the causes and cures of student attrition Chicago: University of Chicago Press. Vaquera, G., (2007). Testing theories of doctoral student persistence at a Hispanic serving institution. Journal of College Student Retention, 9 (3), 283 305.
91 Vogt, C., Hocevar, D., & Hagedorn, L. (2007). A social cognitive construct validation: The Journal of Higher Education, 78 337 364. Wajcman, J. (1991). Feminism Confronts Technology University Park: Pennsylvania State University Press. Weidman, J. (1989). Undergraduate socialization: A conceptual approach. In Higher education: Handbook of theory and research (Vol. 5, pp. 289 322). Norwell, MA: Kluwer Aca demic Publishers. White, M. J., Kruczek, T. A., Brown, M. T., & White, G. B. (1989). Occupational sex stereotypes among college students. Journal of Vocational Behavior, 34 (3), 289 298. Whitt, E. J., Nora, A., Edison, M. I., Terenzini, P. T., & Pascare lla, E. T. (1999). Additional evidence. Journal of College Student Development, 40 (2), 163 177. Woloch, N., (1999) Women and the American Experience. New York: McGraw Hill. Wood, B., & Brown, L. (1997). Participation in an all female algebra class: Effects of high school math and science course selection. Journal of Women and Minorities in Science and Engineering, 3 265 277. Wyer, M. (2000). Intending to stay: Positive ima ges, attitudes, and classroom experiences Dissertation Abstracts International 62(2 B), 1063A. (UMI No. AAI3005371). Retrieved August 8, 2010.
92 Zeldin, A. & Pajares, F. (2000) Against the odds: Self efficacy beliefs of women in mathematical, scientific, and technological careers. American Educational Research Journal, 37 215 246. Zohar, D. (2000). A group level model of safety climate: Testing the effect of group climate on micro accidents in manufacturing jobs. Journal of Applied Psychology, 85 (4), 587 596.
93 Appendix : Effects of College Degree Program Culture o n Female and Minority Student STEM Participation : Student Survey
Appendix 94 Effects of College Degree Program Culture on Female and Minority Student Science, Technology, Engineering and Mathematics (STEM) Participation: Student Survey The purpose of this survey is to learn about the academic department, college, and university th at you currently attend. There are no right or wrong answers. We want to know your experiences, opinions, attitudes, and impressions. Participation is voluntary. Your answers will be kept completely confidential, and combined with responses from others to provide an overall description of your department. This survey will take approximately 30 minutes to complete. As you answer these questions, please keep the following definitions in mind: Department refers to the division with in your college, through whi ch you are working to Department of Civil Engineering, Department of Physics, etc.). College is a division within your university that contains multiple departments, including your department (e.g., College of Arts and Sciences, College of Engineering). Faculty refers to any tenured or non tenured professors, lecturers or adjunct instructors teac hing courses or conducting research in your department. Staff in your department includes the any staff and administrative personnel, such as advisors and office personnel. Filling Out the Answer Sheet In order to complete this survey, you ha ve been supplied with a green answer sheet. Please turn to side 1 of the answer sheet and then complete the following steps: 1. 2. 3. For the Grade or EDUC section, pl than four years of h igher education (for example you are a fifth year senior or are a graduate student). 4. located on your Student Background Information Sheet. Fill in the first number under column 5. For each question on the survey, match the survey question number in this booklet with the corresponding set of response bubbles on the answer sheet. For example, to respond to question #1, use the bubbles for response #1 on the answer sheet. 6. Remember to use ONLY a #2 Pencil and fill in your bubbles completely. Please, DO NOT WRITE ON THIS SURVEY PACKET. 7. Questions 61 73 will use side two of the green answer sheet. Please turn the page to begin the survey.
Appendix (continued) 95 Using the scale below and the provided green answer sheet, please indicate your level of agreement with each of the following sentences: Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree A B C D E 1. Faculty and staff are often available for students to see OUTSIDE OF regular office hours. 2. Faculty and staff provide opportunities for students to work on research projects. 3. Faculty and staff help students achieve professional goals. 4. Faculty and staff seem to take responsibility for students' success. 5. Faculty are enthusiastic about teaching. 6. Faculty and staff provide students with strong academic and professional role models. 7. Faculty and staff help students develop creative capacities. 8. Faculty and staff go out of their way to help students master the knowledge in their discipline. 9. Faculty and staff are generally encouraging towards students. 10. Current students try to make new students feel included. 11. It is a priority to create a diverse, multicultural environm ent. 12. Students are encouraged to develop critical, evaluative, and analytical qualities. 13. Students often work together on team projects (e.g., research projects). 14. People show respect for cultural differences. 15. Faculty and staff genuinely care about students' well being. 16. The interpersonal atmosphere among students is cold. 17. I feel like I fit in well.
Appendix (continued) 96 18. Faculty or staff will offer help to a student who appears lost or confused. 19. Students share strategies for success with each other. 20. Students have to study very ha rd to succeed. 21. People value diversity. 22. Students often learn from each other. 23. The courses provide state of the art information about our field. 24. I sometimes feel out of place. 25. There is a genuine concern for the needs of students of all races and ethnicitie s. 26. Special efforts are made to help racial and ethnic minority students feel like they "belong". 27. The perspectives of men and women are equally valued. 28. Students are highly engaged in coursework. 29. Instructors treat all students the same regardless of their race or ethnicity. 30. Instructors treat all students the same, both men and women. 31. There is an overall sense of community among the students. 32. Students are well prepared to obtain very good jobs when they graduate. 33. I have never observed discriminatory behavi ors (e.g., words or gestures) directed towards female students. 34. There is an emphasis on developing vocational and occupational competence. 35. Faculty and staff make students feel inferior. 36. Students are often too concerned with their own success to help each o ther. 37. The research conducted by our professors and students is some of the best in our field. 38. The degree I am working on is in an exciting field. 39. Individuals getting a degree in my major are respected by most people. 40. My future occupation makes an important contribution to society.
Appendix (continued) 97 Below is a list of supports and services your institution may provide to help you succeed in school. If you participated in any of the following, please indicate how helpful you found this service. If you did not participate in t he service, please indicate so. Very Unhelpful Unhelpful Neither Helpful nor Unhelpful Helpful Very Helpful in this service but it is available My institution does not offer this service. A B C D E F G 41. Pre college outreach or training (i.e., summer programs for high school students). 42. College transition support (i.e., bridge programs for entering freshman or programs to assist transfer students). 43. 1st or 2nd year introductory courses offered by your department. 44. Advising support. 45. Tutoring support. 46. Real life training or applied classroom project. 47. Off campus training opportunities (i.e., internships). 48. Career guidance or job search support. Using the scale below and the provided green answer sheet, please indicate your level of agreem ent with each of the following sentences: Never Occasionally Sometimes Often Very Often A B C D E Since coming to the department, how often have you done the following: 49. Worked in small, ethnically diverse groups with other students in the department. 50. Participated in an ethnic or racial oriented student organization. 51. Socialized with someone of another race or ethnic group. 52. Studied with other students in your major 53. Spent time with students in your major doing non academic activities
Appendix (continued) 98 Using the scale below and the provided green answer sheet, please indicate your level of agreement with each of the following sentences: Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree A B C D E 54. I enjoy discussing my department with the people outside it. 55. I do not feel like "part of the family" in my department. 56. I do not feel "emotionally attached" to my department. 57. My department has a great deal of personal meaning to me. 58. My department has a great deal of personal meaning to me. 59. I am satisfied with my department. 60. I do not like my department. 61. I do not consider any of my fellow students as my friends. 62. I would be a good representative of my department. 63. 64. Within my department, I would easil 65. I understand the goals of my department 66. I have mastered the knowledge required to function well as a student in this department. 67. I am usually excluded in informal networks or gatherings of people within my department. 68. 69. I have not fully developed the appropriate skills and abilities to perform successfully as a student. 70. I believe most of my fellow students like me. 71. I support the goal s that are set by my department. 72. I understand what I need to do to be a successful student.
Appendix (continued) 99 73. Given the opportunity to enroll in the same degree program at a different but equally ranked University, I would: a. definitely maintain enrollment at this University. b. probably maintain enrollment at this University. c. I don't know no opinion. d. probably enroll at the alternative University. e. definitely enroll at the alternative University.
About the Author Heather Ureksoy received a B achelor of A rts in C riminology and a B achelor of S cience in P sychology from the University of Florida in 2004. She received a M I ndustrial/ O rganizational P sychology in 2008 and a Ph.D. in in I ndustrial/ O rganizat ional P sychology in 2011, both from the University of South Florida. Heather is employed as a project manager in the research and development department at Psychological Assessment Resources, Inc., a prominent test developer in Lutz, Florida. She primarily studies the educational and career pathways of individuals in science, technology, persistence. Heather and her husband currently reside in Valrico, Florida with their cats and d og.