CONCEPT AMONG COMMUNITY COLLEGE STUDENTS IN

Journal of Women and Minorities in Science and Engineering, vol. 11, pp. 209–229, 2005

INFLUENCE OF PRECOLLEGE EXPERIENCE ON SELFCONCEPT AMONG COMMUNITY COLLEGE STUDENTS IN SCIENCE, MATHEMATICS, AND ENGINEERING

Soko S. Starobin* and Frankie Santos Laanan Iowa State University Female and minority students have historically been underrepresented in the field of science, mathematics, and engineering at colleges and universities. Although a plethora of research has focused on students enrolled in 4-year colleges or universities, limited research addresses the factors that influence gender differences in community college students in science, mathematics, and engineering. Using a target population of 1,599 aspirants in science, mathematics, and engineering majors in public community colleges, this study investigates the determinants of self-concept by examining a hypothetical structural model. The findings suggest that background characteristics, high school academic performance, and attitude toward science have unique contributions to the development of self-concept among female community college students. The results add to the literature by providing new theoretical constructs and the variables that predict students’ self-concept.

INTRODUCTION Female and minority student populations have long been underrepresented in the field of science, mathematics, and engineering at America’s colleges and universities (National Science Foundation, 2000). Specifically, the shortage of female students in the fields of science, mathematics, and engineering at postsecondary education institutions has concerned educational researchers and policymakers despite the fact that overall postsecondary enrollment of women has been higher than that of their male counterparts (National Center for Education Statistics, 2002). The National Science Foundation (2000) reports that women received 47% of all science and engineering bachelor’s degrees, 39% of the master’s degrees, and just 33% of the doctoral degrees in 1996. It is distressing that the gender discrepancy in educational attainment increases as women’s educational experiences successfully progress at higher levels. Over the decades, America’s more than 1,200 private or public 2-year colleges have facilitated the increase of female representation in the field of science, mathematics, and engineering. The nation’s community colleges provide educational as well as vocational opportunities for students through vocational-technical education, academic transfer to 4-year colleges and universities, remedial courses, continuing education, and community service (Cohen & Brawer, 2003; Laanan, 2003). These community colleges enroll approximately 5.5 million credit students, which include more than 57% of female students in the student population pool (National Center for Education Statistics, 2002). Inasmuch as community colleges awarded more than 76,000 associate degrees in allied health fields, which serve the nation’s significant healthcare needs (Phillippe & Patton, 2000), limited

*Correspondence concerning this article should be addressed to Dr. Soko S. Starobin, Iowa State University, Educational Leadership and Policy Studies, N243 Lagomarcino Hall, Ames, IA 50011-3195; Tel.: 515-294-9121; Fax.: 515-294-4942; e-mail: [email protected]. ISSN 1072-8325/05$35.00 Copyright © 2005 by Begell House, Inc.

Electronic Data Center, http://edata-center.com Downloaded 2006-3-13 from IP 67.87.127.223 by Joanna Antosiuk

209

210

Soko S. Starobin and Frankie Santos Laanan

attention has been paid to unveil the contributions of community colleges to the nation’s science, mathematics, and engineering education. Concerning the nation’s greater need to strengthen its national defense and homeland security, policymakers have addressed the issue of educating competent scientists and engineers within the United States as their policy objective. Federal agencies, such as the National Science Foundation, have recently documented positive impacts of community college attendance on career and educational pathways of science and engineering undergraduates and graduates (Tsapogas, 2004a, 2004b). Furthermore, the Undergraduate Science, Mathematics, Engineering and Technology (SMET) Education Improvement Act - H.R. 3130, which became Public Law 107-368 in December 2002, specifically recognizes community colleges as contributors to the increasing number of undergraduate students in science, mathematics, and engineering education. Selected policy objectives of the Act that relate to community colleges are to increase overall workforce skills by • raising postsecondary enrollment rates of women, minorities, and persons with disabilities in science, mathematics, engineering, and technology disciplines; • increasing access to higher education in science, mathematics, engineering, and technology fields for students from low-income households; and • expanding science, mathematics, engineering, and technology training opportunities at institutions of higher education (P.L. 107-368, 2002). Since the introduction of the Undergraduate Science, Mathematics, Engineering and Technology (SMET) Education Improvement Act - H.R. 3130, the American Association of Community Colleges (AACC) and the National Science Foundation (NSF) have developed a collaborative partnership to address the role of community colleges in increasing the representation of women and minorities majoring in science, mathematics, engineering, and technology. Given the substantial percentage of female students enrolled in community colleges, AACC has successfully influenced policymakers to recognize that this pool of individuals serves as a potential group to be the future scientists through this partnership. A body of research that specifically addresses the disparities in mathematics and science education among male and female students, however, has focused its analysis on students at the secondary level or at 4-year institutions. Questions about the factors that influence gender differences in college students’ mathematical achievement and self-concept have been a popular inquiry. Evidence from previous research suggests that high selfconcept positively influenced academic performance (Astin, 1993; Bailey, 1971; Byrne, 1984; Hansford & Hattie, 1982; House, 1995). Currently, there is a dearth of research that investigates mathematical ability and the degree of encouragement to pursue study in science, mathematics, and engineering among students in community colleges. Moreover, little attention has been paid to uncover the effects of students’ background characteristics and pre-college experiences with respect to community college students’ self-concept. The objective of this study was to investigate the determinants of community college students’ self-concept. Specifically, students who indicated that their probable major was science, mathematics, or engineering were included in the target population. To examine the determinants of self-concept among aspirants of science, mathematics, and engineering majors in community colleges, the researchers hypothesized a latentvariable structural equation model by including variables identified in previous research

Journal of Women and Minorities in Science and Engineering


Self-Concept Among Community College Students

211

findings as being significantly related to the development of self-concept of students in science, mathematics, and engineering.

REVIEW OF THE LITERATURE The research studies on students in science, mathematics, and engineering at the secondary (Carlone, 2003; Drake, Clewell, & Sevo, 2002; Spears, Dyer, Franks, & Montelone, 2004) and postsecondary levels (Huang & Brainard, 2001; Hughes, 2000; Wyer, 2003) are extensive in theoretical and conceptual approaches. For instance, the application of the differential coursework hypothesis (Ethington & Wolfle 1984, 1986; Nora & Horvath, 1990; Nora & Rendón, 1990; Pallas & Alexander, 1983), the spatial visualization (Fennema & Sherman, 1977; Sherman, 1980, 1982, 1983), the expectancy-value theory (Lips, 1995), and the college impact on retention (Pascarella, Smart, Ethington, & Nettles, 1987) were used in the past studies. These empirical studies used these theoretical and conceptual frameworks to investigate the factors that influence gender disparities in science and mathematics education. Furthermore, the findings of these studies are consistent regarding the importance of participation in science and mathematics courses, as well as psychosocial and cultural issues, in shaping the confidence and educational success of female students in science and mathematics. Studies on Participation in Science and Mathematics Courses To measure the postsecondary academic achievement of students in science and mathematics, secondary students’ levels of participation in science and mathematics courses were identified as the attributing factor (Ethington & Wolfle, 1984, 1986; Nora & Horvath, 1990; Nora & Rendón, 1990; Pallas & Alexander, 1983). Known as the differential coursework hypothesis, this assumption is grounded in the premise that there is a plausible relationship among gender differences in academic performance and differences in students’ programs of study during high school. To test this hypothesis, Pallas and Alexander (1983) studied a cohort of 6,119 students in 24 public high schools from 1961 to 1968, corresponding to grades 5 to 12, by examining the effects of differences in the pattern of quantitative coursework in high school on the gender difference in quantitative SAT performance. The study revealed that as girls outperformed boys in quantitative courses, their SAT mathematics scores increased— thus closing the overall gender performance gap. On the basis of their findings, Pallas and Alexander (1983) concluded that this study supports the differential coursework hypothesis between males and females. Additionally, interventions for encouraging female students to enroll in quantitative high school courses were identified as a critical factor in their future success in quantitative studies. Using the data drawn from High School and Beyond, a nationwide longitudinal study of high school sophomores and seniors in 1980 that included 16,555 students, Ethington and Wolfle (1986) found a similar influence of prior exposure to mathematics on students’ academic performance. The structural equation model designed for this study revealed that students’ sophomore-year exposure to mathematics, mathematics and verbal abilities, and attitudes toward mathematics were predictors of mathematics achievement of students in their senior year. Nora and Horvath (1990) modified the structural equation model used in the Ethington and Wolfle study to examine the effects of sociodemographic factors,

Volume 11, Issue 3


212


precollege factors, and attitudinal factors on the differences in academic preparation among students in Hispanic-serving 2-year institutions. Using Self-Concept as a Measure of Gender Differences In addition to the students’ levels of participation in science and mathematics courses, confidence levels and self-concept of students have been examined as —contributing factors in determining academic achievement of students in science and mathematics. Research studies on self-concept pertaining to science and mathematics education are diverse. This concept has been referred to as academic self-confidence (Huang & Brainard, 2001), mathematics attitudes (Ethington & Wolfle, 1986; Nora & Horvath, 1990; Nora & Rendón, 1990), confidence in learning mathematics (Fennema & Sherman, 1977; Sherman, 1982, 1983), course confidence (Hughes, 2000), self-confidence (Lips, 1995), and mathematical self-concept (Sax, 1994a, 1994b). Sax (1994a) investigated factors predicting mathematics self-concept and their gender differences by using a national sample of approximately 15,000 students drawn from the Cooperative Institutional Research Program (CIRP) 1985 Freshman Survey and 1989 Follow-up Survey. It was concluded that female students were less confident in their mathematical abilities than their male counterparts, and the gender disparity increased as they progressed through college. Additionally, Sax (1994a) reported that for female students in particular, strong academic performance as well as mathematics and science preparations in high school had far stronger influence on mathematics self-concept than the impact of college. The results of this study offer two significant interpretations. The finding of the relationship between mathematics and science preparations in high school supported the applicability of the differential coursework hypothesis to the measurement of self-concept, which was defined as the outcome variable to test the gender disparities. Another notable finding from this study was that universities’ competitive learning environments and low enrollments of women negatively affected the development of mathematics selfconcept of female students. This finding especially invites impetus for an investigation of self-concept of female students in community colleges, where the majority of students are female and academically less competitive. Thus, the learning environment in community colleges appears to be more desirable for female students to pursue their studies in science and mathematics. Research on Community College Students in Science and Mathematics Despite the plethora of research studies that addressed gender differences in academic performance, attrition, and retention in students in science and mathematics education, the focus has been limited to students at secondary and 4-year university levels. It is important, however, to address the gender differences of students in science, mathematics, and engineering and the extent to which characteristics of students at community colleges are different from the characteristics of students at secondary or 4-year institutions. Nora and Rendón (1990) examined the gender and racial differences in mathematics and science preparation and participation among students in Hispanic-serving 2-year institutions using sociodemographic, academic preparation, and attitudinal variables. Their results suggested that Hispanic female students were likely to be in the traditional 18- to 24-year-old student




213

pool and received adequate academic preparation, whereas White female students comprised an older and poorly prepared student population. Clearly, these women chose community colleges as their initial point of entry into higher education. Further, the authors argued that “even the brightest Hispanics may be choosing to begin their college careers in mathematics and science education in community colleges, and that 2-year institutions may be the primary collegiate vehicle Hispanics use to attain career, economic and social mobility” (Nora & Rendón, 1990, p. 37). Community colleges, specifically for traditionally underrepresented student populations pursuing postsecondary education in mathematics and science, have continuously played a significant role. Analyzing the data from the National Science Foundation’s 2001 National Survey of Recent College Graduates, Tsapogas (2004a) found that 44% of science and engineering (S&E) bachelor’s and master’s graduates attended community colleges “at some point in their educational path” (p. 6). Considering the background characteristics of the S&E graduates, Hispanics, females, and nontraditional age are more likely than their counterparts (other ethnic groups, males, traditional-age students) to attend community colleges (Tsapogas, 2004a).

CONCEPTUAL FRAMEWORK This study uses two conceptual frameworks: (1) the input-environment-outcome model (I-E-O) and (2) the differential coursework hypothesis. Developed by Astin (1993), the I-E-O model accesses the impact of various environmental experiences by determining whether students grow or change differently under varying environmental conditions. Although the I-E-O model is designed to examine the college experiences of 4-year college students, the model can be applied to community college freshman students with model modifications. In the present study, the inputs refer to the background characteristics of the student; environment refers to the high school experiences of the student; and outcomes refer to the student’s characteristics (cognitive and affective) after exposure to their high school experiences. To measure the high school experience of students, the differential coursework hypothesis is applied to examine the influence of students’ levels of participation in science and mathematics courses on the gender differences in students’ cognitive and affective characteristics. Although the research on the differential coursework hypothesis has been well documented (Ethington & Wolfle, 1984, 1986; Pallas & Alexander, 1983), only a few studies investigated community college students. Nora and Horvath (1990) and Nora and Rendón (1990) added a new dimension to the hypothesis by examining the effects of sociodemographic factors, precollege factors, and attitudinal factors on the differences in academic preparation among community college students. The authors concluded that differences in the effects were explained by students’ gender and ethnicity. The conceptual framework provides a useful guide in developing a hypothetical structural model to examine the determinants of self-concept of community college students. First, several key variables used in previous studies of the differential coursework hypothesis were identified for the selection of latent variables in this study. To measure the latent variables, 11 observed variables were selected based on Sax’s (1994a) study in which she predicted gender differences in math self-concept among 4-year college students. The next step was to determine the relationship among the latent variables. The modified I-E-O

Volume 11, Issue 3


214


model was applied to determine the relationship as well as directions of the latent variables in the hypothetical model. It was hypothesized, based on previous research (Astin, 1993; Pascarella et al., 1987), that background characteristics (i.e., socioeconomic status and parental education) and high school experiences (i.e., high school GPA and high school coursework) are associated with scientific orientation (i.e., goals), as well as students’ selfconcept (i.e., academic and mathematical ability). Although academic performance and scientific course enrollment have been selected as the outcome measurements of examining the differential coursework hypotheses in previous studies (Ethington & Wolfle, 1984, 1986; Nora & Horvath, 1990; Nora & Rendón , 1990; Sherman, 1980, 1982, 1983), this study specifically focuses on examining the effects of three constructs: (1) background characteristics; (2) high school academic performance; and (3) attitude toward science on the self-concept of science, mathematics, and engineering aspirants in community colleges. Objectives The objective of this study is to understand the influence of students’ background characteristics, high school academic performance, and attitude toward science on their self-concept. Specifically, this study addresses gender differences and the extent to which each construct influenced students’ self-concept. The following research questions guided this study: • What are the characteristics of students who aspired to baccalaureate degrees in science, mathematics, and engineering? • To what extent is the self-concept of students predicted by background characteristics, high school academic performance, and attitude toward science? • To what extent is the self-concept of female students predicted by background characteristics, high school academic performance, and attitude toward science?

DATA SOURCE AND METHODS The data for this study were drawn from the CIRP 1996 Freshman Survey, which was sponsored by the University of California at Los Angeles (UCLA) Higher Education Research Institute (HERI). The cohort of students who were first-time freshmen enrolled in public and private 2-year institutions in fall 1996 responded to the Student Information Form (SIF). The SIF was administered to students during the beginning of the fall 1996 semester to elicit a wide range of student information, which includes biographic and demographic information as well as attitudinal and affective measures of this student cohort (Sax, Astin, Korn, & Mahoney, 1996). The target population for the study consisted of 1,599 first-time, full-time students enrolled in public community colleges in fall 1996 who indicated their probable major as biological science, physical science (includes mathematics), or engineering on the survey. Variables and Measures This study was delimited to use a target population of students enrolled in the public community colleges that participated in the 1996 CIRP American Freshman Study. A




215

complete description of the research methodology, as well as other technical issues related to the 1996 CIRP American Freshman Study, is provided in The American Freshman: National Norms for Fall 1996 (Sax et al., 1996). The data for the present study is cross sectional and relies on reflective self-report data to examine the transitional influences on students’ self-concept. The variables used in preparing descriptive statistics in the study include (1) gender; (2) age; (3) race/ethnic background; (4) students’ probable major; (5) parents’ annual income; (6) father’s educational attainment; (7) mother’s educational attainment; and (8) high school academic performance, which consists of years of study in mathematics, physical science, biological science, and Grade Point Average (GPA). For the structural equation model (see Fig. 1), the 11 observed variables were selected from the SIF. The coding and scaling of these variables are presented in the Appendix. To test the structural model, two exogenous variables were measured by seven observed vari-

Parents' Income Academic Self-concept Father's Education

Background characteristics

Self-concept Social Self-concept

Mother's Education

High School Mathematics Scientific Orientation

High School Physical Science HS Academic Performance

Attitude tow ard Science

High School Biological Science

Materialism and Status

High School GPA

Figure 1. Structural model for self-concept: influence of background characteristics, high school academic performance, and attitude toward science on self-concept.

Volume 11, Issue 3


216


ables. The first independent latent variable, ξ1 = background characteristics, was measured by three variables: x1 = parents’ income, x2 = father’s education, and x3 = mother’s education. The other independent latent variable, ξ2 = high school academic performance, was identified by four observed variables: x4 = years of high school study in mathematics, x5 = years of high school study in physical science, x6 = years of high school study in biological science, and x7 = high school GPA. To illustrate students’ self-concept, two endogenous (dependent latent) variables were measured. Students’ self-concept, η1 = self-concept, was measured by two observed variables: y1 = academic self-concept and y2 = social self-concept. The other endogenous variable was constructed on the basis of the peer influential factors identified by Astin (1993) that were specifically relevant to students in science, mathematics, and engineering. The variable, η2 =attitude toward science, was measured by two observed variables: y3 = scientific orientation and y4 = materialism and status. The descriptions of four latent variables and observed variables that constructed these latent variables are the following: Background Characteristics. Three items: (a) parents’ income, (b) father’s education, and (c) mother’s education were used to assess students’ background characteristics, identified as “inputs” in the modified I-E-O model. High School Academic Performance. Four items were used to measure high school academic performance, which was conceptualized as the “environment” in the model. The items are (a) years of high school study in mathematics, (b) years of high school study in physical science, (c) years of high school study in biological science, and (d) high school GPA. Attitude toward Science. A latent factor had two measures: (a) scientific orientation (i.e., “make a theoretical contribution to science”) and (b) materialism and status (i.e., “being very well off financially and being successful in my own business). Attitude toward science was identified as “outcome” variable in the model. Self-Concept. Identified as the second “outcome” in the model, self-concept includes nine items that were separated into two parcels: academic self-concept and social selfconcept. Representative items of the academic self-concept included, academic and mathematical abilities. Items such as “public speaking” and “drive to achieve” were included in the social self-concept.

The model depicts the effects of three latent variables: (1) background characteristics; (2) high school academic performance; and (3) attitude toward science on the dependent latent variable, self-concept. Analyses To address the first research question, descriptive statistics were conducted to illustrate the target population pertaining to students’ background characteristics, selection of probable majors, and high school academic performance described above. The background characteristics variables and the variable that depicts students’ probable majors were crosstabulated with respondent gender to provide a portrait of gender differences in community college students in science, mathematics, and engineering. Furthermore, students’ high




217

school academic performance variables were cross-tabulated with the gender variable to examine the gender differences in academic participation in science and mathematics during high school. The structural equation model developed in this study was analyzed by using LISREL Version 8.54 (Jöreskog & Sörbom, 2002). The model was developed to investigate the selfconcept of students in science, mathematics, and engineering in community colleges. The hypothesized structural model represents a twofold approach: measurement (confirmatory factor analysis) and latent variable structural model. The model depicts the relationship between four latent variables: ξ1 = background characteristics, ξ2 = high school academic performance, η1 = students’ self-concept, and η2 = attitude toward science. As a first step, measurement models for both independent and dependent latent variables were examined by conducting confirmatory factor analysis. For the Figure 1 model, 11 measurement equations were developed, one for each observed variable, as follows: Parents’ income Father’s education Mother’s education

= function of background characteristics + error = function of background characteristics + error = function of background characteristics + error

High High High High

= = = =

school school school school

mathematics physical science biological science GPA

function function function function

of of of of

high high high high

school school school school

academic academic academic academic

performance performance performance performance

Academic self-concept Social self-concept

= function of self-concept + error = function of self-concept + error

Scientific orientation Materialism and status

= function of attitude toward science + error = function of attitude toward science + error

These equations were written for the x observed variables as x1 = x2 = x3 = x4 = x5 = x6 = x7 =

1 λx21 λx31 1 λx52 λx62 λx72

ξ1 ξ1 ξ1 ξ2 ξ2 ξ2 ξ2

+ + + + + + +

δ1 δ2 δ3 δ4 δ5 δ6 δ7

and for the y observed variables as y1 y2 y3 y4

= = = =

1 λy21 1 λy42

η1 + η1 + η2 + η2 +

ε1 ε2 ε3 ε4

The complete matrix equations for the x observed variables were written as

Volume 11, Issue 3


+ + + +

error error error error

218


⎛ 1 ⎛ x1 ⎞ ⎜ λx ⎜x ⎟ ⎜ 21 ⎜ 2⎟ ⎜ λx31 ⎜ x3 ⎟ ⎜ ⎜ ⎟ ⎜ x4 ⎟ = ⎜ 0 ⎜ 0 ⎜ x5 ⎟ ⎜ ⎜ ⎟ ⎜ 0 ⎜ x6 ⎟ ⎜ 0 ⎜x ⎟ ⎝ ⎝ 7⎠

0 ⎞ ⎛ δ1 ⎞ ⎟ ⎜δ ⎟ 0 ⎟ ⎜ 2⎟ ⎜ δ3 ⎟ 0 ⎟ ⎟ [ξ1 ] ⎜ ⎟ 1 ⎟ + ⎜ δ4 ⎟ [ ξ2 ] ⎜ δ5 ⎟ λx52 ⎟ ⎟ ⎜ ⎟ λx62 ⎟ ⎜ δ6 ⎟ ⎟ ⎜δ ⎟ λx72 ⎠ ⎝ 7⎠

and for the y observed variables, the matrix equation was indicated as

⎛ y1 ⎞ ⎜y ⎟ ⎜ 2⎟ ⎜ y3 ⎟ ⎜ ⎟ ⎝ y4 ⎠

⎛ 1 ⎜ λy ⎜ 21 ⎜ 0 ⎜ ⎝ 0

⎞ ⎛ ε1 ⎞ ⎟ η ⎜ ⎟ ⎟ ⎛ 1 ⎞ + ⎜ ε2 ⎟ ⎟ ⎜⎝ η2 ⎟⎠ ⎜ ε3 ⎟ ⎟ ⎜ ⎟ λy42 ⎠ ⎝ ε4 ⎠ 0 0 1

The next step was to develop two structural equations, one for students’ self-concept and the other for attitude toward science, as follows: Students’ self-concept

= background characteristics + high school academic performance + attitude toward science + error Attitude toward science = background characteristics + high school academic performance + error

In matrix notation, these equations were written as η1 = β12 η2 + γ11 ξ1 + γ12 ξ2 + ζ1, and η2 = γ21 ξ1 + γ22 ξ2 + ζ2. In matrix form, they are represented as ⎛0 ⎛ η1 ⎞ ⎜η ⎟ = ⎜0 ⎝ ⎝ 2⎠

β12 ⎞ ⎛ η1 ⎞ ⎛ γ 11 γ 12 ⎞ ⎛ ξ1 ⎞ ⎛ ζ1 ⎞ + ⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎟ ⎜ ⎟ 0 ⎠ ⎝ η2 ⎠ ⎝ γ 21 γ 22 ⎠ ⎝ ξ2 ⎠ ⎝ ζ 2 ⎠

In these matrices, the factor loadings are given as γi and the prediction errors as ζi. The plausibility of the hypothetical structural equation model was assessed using several fit criteria: the comparative fit index (CFI) of more than 0.90 (Bentler, 1990) and the rootmean-square error of approximation (RMSEA) of 0.05 or less (Steiger & Lind, 1980). In terms of determining the sample size, the range of 250 to 500 or greater and a ratio of at least 10 subjects per variable (Schumacker & Lomax, 1996, 2004) were used as guidelines. For each analysis, a sample size was determined as 500 for the overall group and 250 for the female group. First, the proposed structural equation model was examined for the overall




219

sample of 500 students. Once a fit model was determined for the overall sample, the sample of female students was tested for a model fit.

RESULTS Descriptive Analysis The target population of first-time, full-time community college students who responded to the 1996 CIRP Freshman Survey represented 1,599 students. With regard to gender distribution, this target population included approximately 70% male and 30% female students. Table 1 illustrates a cross-tabulation analysis of age and racial background distributions of students by gender. Overall, approximately 86% of students were of traditional college-going age (less than 24-years old). Although similar percentages of male and female students represent the traditional college-going age (86.2% and 84.2%, respectively), female students were slightly younger than their male counterparts. Specifically, at the traditional high school graduation age of 18, 44.1% of female students and 35.9% of male students were accounted for in this age distribution. Observing the distributions of older students, 3% of male students and 2.5% of female students were 40 years or older in this population. In terms of students’ ethnicity, more than 25% were minority students. In terms of gender differences in the distribution of minority students, African-American female students accounted for the largest minority representation (8.5%). The percentage of Asian-American/Asian students Table 1. Frequency Distributions of Community College Students’ Demographics Frequencies

Percentages

Variables

Male

Female

Male

Female

Age (on December 31, 1996) 16 or younger 17 18 19 20 21 to 24 25 to 29 39 to 39 40 to 54 55 or older

8 19 407 313 76 154 71 53 29 4

3 7 196 102 25 41 28 31 10 1

0.7 1.7 35.9 27.6 6.7 13.6 6.3 4.7 2.6 0.4

0.7 1.6 44.1 23.0 5.6 9.2 6.3 7.0 2.3 0.2

Racial/Ethnic Background African American/Black American Indian Asian American/Asian Mexican American/Chicano Puerto Rican Other Latino Other White/Caucasian

75 39 35 52 17 36 43 883

40 8 24 12 6 17 16 349

6.4 3.3 3.0 4.4 1.4 3.1 3.6 74.8

8.5 1.7 5.1 2.5 1.3 3.6 3.4 73.9

Volume 11, Issue 3


220


(5.1%) comprised the next highest minority representation in the female student group. However, when a sum of Mexican-American/Chicano (2.5%), Puerto Rican (1.3%), and Other Latino (3.6%) students were counted as a total Hispanic female population, 7.4% of students were represented in this group. Interestingly, for male students, the total count of Hispanic (4.4% Mexican-American/Chicano, 1.4% Puerto Rican, and 3.6% Other Latino), which accounts for 8.9% of the male population, exceeded the representation of AfricanAmerican students (6.4%). In sum, with regard to the distribution of students’ race and ethnicity, the 1996 CIRP data report that 74.6% of students were White (non-Hispanic), 8.5% were Hispanic, 7.0% were African American, 3.6% were Asian American/Asian, and only 2.8% were American Indian. With regard to the probable majors of the target population, 53.1% of students chose engineering, 35.1% and 11.8% chose biological science and physical science, respectively (see Table 2). Considering the gender differences in choosing majors, almost 67% of male students chose majors in engineering, and the most represented majors were electrical or electronic engineering (23.2%) and mechanical engineering (16.7%). Conversely, approximately 66% of female students selected majors in biological science. Among female students in biological science, more than one third of students chose biology as their probable major. Although the numbers were relatively small, physical science majors were equally represented by male (10.6%) and female (14.8%) students. In reviewing the distribution across high school academic performance of the target population, several statistical indications show that female students performed better overall. For instance, Table 3 indicates that more than 70% of female students had high school GPAs of “B” or better, whereas 57.1% of male students had such grades. The distribution also revealed that the mode was at grade “B” (26.2%) for female students and at grade “D” (29.3%) for their male counterparts. Furthermore, the distributions for both genders indicated a peak at grade “B”; female distribution was skewed from the higher grade to the lower grade, whereas male distribution illustrated the opposite skewness. In terms of students’ high school years of study in mathematics, physical science, and biological science, the gender differences were evident only in biological science. In the distributions of years of study in mathematics, 87% of the male and 88% of the female students had 3 years or more of study in mathematics. Furthermore, the percentage of male students responding as having studied 3 years or more in physical science was 24.5%, whereas the percentage of female responses was 22.9%. In biological science study, however, 20.9% of the female students had 3 years or more of study, and only 10.4% of male students had such a length of study in this field. Model Analysis Results The structural model was designed to examine the effects of three latent variables: (1) background characteristics, (2) high school academic performance, and (3) attitude toward science on the dependent latent variable, self-concept. The previous investigation of the impact of college experiences on math self-concept among 4-year college students indicates that precollege experiences have stronger influences on students’ ultimate level of math confidence than the influences derived from college experiences (Sax, 1994a). As for female students, Sax (1994a) found that precollege experiences of students, which include initial interests in science, higher high school grades, and greater math and science preparation in high school, appeared as strong predictors of students’ math self-concept.




221

Table 2. Frequency Distributions of Community College Students’ Probable Major Frequencies Variables Student’s Probable Major Biological Science Biology Biochemistry or Biophysics Botany Environmental Science Marine (life) Science Microbiology or Bacteriology Zoology Other Biological Science Engineering Aeronautical or Astronautical Engineering Civil Engineering Chemical Engineering Electrical or Electronic Engineering Industrial Engineering Mechanical Engineering Other Engineering Physical Science Astronomy Atmospheric Science Chemistry Earth Science Marine Science Mathematics Physics Other Physical Science Total

Percentages

Male

Female

Male

Female

76 20 9 81 18 8 19 26 257

96 17 4 43 40 16 44 44 304

6.7 1.8 0.8 7.1 1.6 0.7 1.7 2.3 22.6

20.9 3.7 0.9 9.3 8.7 3.5 9.6 9.6 66.1

29 81 28 264 35 190 134 761

8 10 13 18 3 11 25 88

2.5 7.1 2.5 23.2 3.1 16.7 11.8 66.8

1.7 2.2 2.8 3.9 0.7 2.4 5.4 19.1

9 9 22 15 14 21 12 19 121

3 4 16 4 12 13 5 11 68

0.8 0.8 1.9 1.3 1.2 1.8 1.1 1.7 10.6

0.7 0.9 3.5 0.9 2.6 2.8 1.1 2.4 14.8

1,139

460

100.0

100.0

First, the model was tested for the sample from the target population. The model fit indices indicated that the proposed model did not fit the data and required several model modifications. The modified model produced the better fit indices (χ2 = 46.57, df = 34, CFI = 0.986, RMSEA = 0.027, R2 = 0.379). The structural parameter estimates indicated that direct effects of high school academic performance (γ = 1.05) and attitude toward science (γ = 0.15) were statistically significant and positively predictive of self-concept. Furthermore, direct effect of background characteristics on high school academic performance was also significant (γ = 0.18). As illustrated in Table 4, the total effect of background characteristics on self-concept (background characteristics Æ self-concept, background characteristics Æ high school academic performance Æ self-concept, background characteristics Æ attitude toward science Æ self-concept, and background charac-

Volume 11, Issue 3


222


teristics Æ high school academic performance Æ attitude toward science Æ self-concept) was significant (γ = 0.35). The proposed model was also tested for the female sample. With several modifications, the modified model showed a good fit (χ2 = 31.36, df = 34, CFI = 1.00, RMSEA = 0.00, R2 = 0.312). Direct effects among the latent variables, the effects of high school

Table 3. Frequency Distributions of Students’ High School Academic Performances Frequencies

Percentages

Variables

Male

Female

Male

Female

High School GPA D C C+ B– B B+ A– A or A+ Total

17 123 180 170 274 177 119 58 1,101

6 21 57 55 116 84 61 49 443

29.3 11.2 16.3 15.4 24.9 16.1 10.8 5.3 100.0

12.2 4.7 12.9 12.4 26.2 19.0 13.8 11.1 100.0

High School Years of Study Mathematics None One half One Two Three Four Five or more Total

1 9 23 114 265 643 72 1,127

3 1 7 43 138 239 21 452

0.1 0.8 2.0 10.1 23.5 57.1 6.4 100.0

0.7 0.2 1.5 9.5 30.5 52.9 4.6 100.0

Physical Science None One half One Two Three Four Five or more Total

68 37 437 288 159 98 12 1,099

37 19 167 118 52 43 6 442

6.2 3.4 39.8 26.2 14.5 8.9 1.1 100.0

8.4 4.3 37.8 26.7 11.8 9.7 1.4 100.0

Biological Science None One half One Two Three Four Five or more Total

54 42 588 303 64 45 6 1,102

24 12 176 140 51 34 8 445

4.9 3.8 53.4 27.5 5.8 4.1 0.5 100.0

5.4 2.7 39.6 31.5 11.5 7.6 1.8 100.0




223

Table 4. Decomposition of Effects: Predicting Self-Concept Background Constructs

HS academic performance

All

Female

HS academic performance

0.18*a —b 0.18*c

0.23 — 0.23

Attitude toward science

0.09 0.02 0.11

Self-concept

0.15 0.20 0.35*

0.08 –0.02 0.06 0.21 0.20 0.40*

All

Female

0.10 — 0.10

–0.07 — –0.07

1.05* 0.02 1.07*

0.83* –0.01 0.82*

Attitude toward science All

Female

0.15* — 0.15*

0.16* — 0.16*

Note. Hypothesized effects are the direct effects. *t value is significant at 0.05. a Direct effect. b Indirect effect. c Total effect.

academic performance (γ = 0.83) and attitude toward science (γ = 0.16), were statistically significant for predicting self-concept. Unlike the sample from the target population (males and females), the direct effect of background characteristics on high school academic performance was not significant. Considering that indirect effects exist between background characteristics and self-concept, the total effect of background characteristics on selfconcept (γ = 0.40) and the total effect of high school academic performance on self-concept (γ = 0.82) were statistically significant.

CONCLUSIONS The results from the descriptive and model analyses of the data from the 1996 CIRP Freshman Survey revealed valuable information about community college students who aspired to major in science, mathematics, and engineering. Demographically, although the majority of this target population was male and traditional college-age students who are Caucasians, the gender differences were found in age and racial/ethnic distributions. For instance, among this target population, male students were slightly older than female students. The distributions of racial and ethnic background of students were found as consistent with the literature, which portrays the underrepresentation of minority students in science, mathematics, and engineering majors (National Science Foundation, 2000). In terms of the gender difference in students’ racial and ethnic background, African-American students (8.5%) had the highest representation in the female minority group, whereas the number of Hispanic (Mexican-American/Chicano, Puerto Rican, and Other Latino) students (8.9%) accounted for the largest male minority representation. It is noteworthy that regardless of students’ ethnicity, the gender discrepancy is evident from the descriptive analysis results. As America’s more than 1,200 community colleges continue to provide educational opportunities for many female students, community colleges deserve much

Volume 11, Issue 3


224


attention from policymakers, scholars, and the general public to promote an increase in representation of female students in science, mathematics, and engineering. The results from the structural model analysis in this study provided three important findings. First, concerning the conceptual framework that guided the hypothetical structural model, it is clear that the learning “environment” appears to be a contributing factor for predicting the “outcome” variable, self-concept. More specifically, as defined as the measurement for the learning “environment,” students’ high school academic performance was a positive contributing factor of predicting students’ self-concept in the model. This finding was observed in the samples from the target population and the female group. Furthermore, this finding was congruent with the conclusion of previous studies (Astin, 1993; Pascarella et al., 1987) that high school experiences (i.e., high school GPA and high school coursework) were associated with student’s self-concept (i.e., academic and mathematical ability). This finding suggests the importance of educational experiences, specifically academic preparation during high school, for students who aspire to major in science, mathematics, and engineering at community colleges. In examining the influence of background characteristics, “input latent variable,” on predicting students’ self-concept, results from this study did not yield a significant contribution of the background characteristics. However, when the indirect effects (background characteristics Æ high school academic performance Æ self-concept, background characteristics Æ attitude toward science Æ self-concept, and background characteristics Æ high school academic performance Æ attitude toward science Æ self-concept) were taken into account, the total effect of background characteristics showed a significant contribution to predicting students’ self-concept. It is notable that the total effect of background characteristics was stronger for the female group. It is assumed that this “input latent variable” plays a complex and important role because it indirectly influences predicting students’ attitude toward science and self-concept. Thus, it might be beneficial to include additional variables or concepts in this “input latent variable” to better understand the complexity of its influence on predicting self-concept or attitude toward science, specifically among female community college students in science, mathematics, and engineering. Second, a critical finding from the model analysis resulted from another conceptual framework used in this study: the notion of the differential coursework hypothesis. Although female students in this study were more academically prepared (higher GPA and more courses taken in biological sciences) than male students, the influence of high school academic performance on predicting self-concept was less than that of overall students (males and females). The findings suggest that the converse effect of the differential coursework hypothesis with respect to predicting students’ self-concept was found in this study. In addition, it may be that there are other factors in addition to high school academic achievement that may influence the development of female students’ self-concept. Finally, as supported by the literature, the finding of the strong influence of attitude toward science on predicting self-concept among female students indicates the need for support programs and services for them to enhance their attitude toward science. For instance, female students, who are marginalized in science, mathematics, and engineering, can benefit from sociopsychological interventions provided by high school and community college experiences (academic, social, etc.) to develop their self-concept over time. There are several limitations to this study. The sample of this study was a cohort of 1996 first-time, full-time community college students. Although this data set includes students who started college almost 10 years ago, the goal of this research was to build on




225

the 1990 study conducted by Nora and Rendón. Furthermore, after the initial data screening of the CIRP Freshman Survey data of 1999 and 2000, the researchers concluded that the 1996 data set offers the most comprehensive variables with regard to students’ high school academic performances. There was a lack of representation of female students in the sample. As the findings from the descriptive analysis indicated, the majority of the students represented in the data set is Caucasian males. Initially, the researchers attempted to conduct a multiple group analysis to examine gender differences in predicting self-concept. Once the model fit analysis was conducted and finalized with the modifications for the target population (both male and female students), the modified model tested male and female groups separately. The male sample did not fit the data; therefore, the analysis and interpretation of the results were focused on the sample from the target population (includes both males and females) and the female sample. Beyond the limitations for statistical analysis, the lack of female representation in science, mathematics, and engineering forewarns policymakers that the gender disparity already exists at the community college level. The underrepresentation of female students also challenges community college practitioners and faculty to address the issue of retaining the pool of female students. On the basis of the results from this study, the following recommendations are presented and have implications for future research and policymakers to facilitate the promotion of increasing the representation of female students in science, mathematics, and engineering in the postsecondary education arena: • Introduce and examine new variables or constructs for the “input variable” (e.g., parents’ attitude toward science, mathematics, and engineering; parents’ occupation; and financial aid) to evaluate its influence on predicting students’ attitude toward science and self-concept. • Use transcript data to measure high school academic performances. Transcript data permit researchers to use science and mathematics course grades in addition to the high school Grade Point Average (GPA). • Explore institutional characteristics variables as additional “environment” factors in the model. Female students who aspire to study in science, mathematics, and engineering may benefit from the learning environment in community colleges. At a community college, the majority of students are female, and the learning environment is less competitive than in 4-year institutions. Learning environment factors can be measured by analyzing national statistics (e.g., data from the Integrated Postsecondary Education Data System). Another “environment” factor can be measured by examining students’ perceived institutional characteristics (e.g., learning experiences, satisfaction in teaching, and other cultural contexts of community colleges). • Employ a longitudinal design that examines how students develop and change over time. Specifically, the design that includes students’ high school as well as community college experiences and aspirations is critical in understanding students’ level of self-concept, as well as their attitude toward science, mathematics, and engineering. • Develop a longitudinal design that investigates community college students who transfer from community colleges to 4-year institutions to study science, mathematics, and engineering. A comparative study of two cohorts: (1) community college transfer students and (2) students who began their study as freshmen at a

Volume 11, Issue 3


226

Soko S. Starobin and Frankie Santos Laanan 4-year institution will allow researchers to examine the influences of students’ academic and social experiences at community colleges on their academic progress and social adjustment at 4-year institutions. Findings from the comparative study can provide vital information for educators and policymakers to develop a seamless pathway for community college students to obtain a baccalaureate degree in science, mathematics, and engineering. • Conduct studies that examine gender and racial/ethnic differences among students. The results could inform student support professional strategies to recruit, retain, and graduate women and minorities in science, mathematics, and engineering.

In conclusion, there is no doubt that community colleges play a significant role in providing access to postsecondary education for traditionally underrepresented student groups in science, mathematics, and engineering. This segment of American higher education continues to play a critical role in educating and training a highly skilled science, technology, and engineering workforce in the global marketplace. The future of increasing the representation of women and minorities pursuing baccalaureate degrees rests on the extent to which the transfer function of community colleges facilitates student movement from the 2- to 4year college or university. The interinstitutional collaboration, as well as transfer and articulation policies, will ultimately determine the educational pathway of community college students who aspire to study in science, mathematics, and engineering. Community colleges are the place in which students begin their educational pursuits, and it is also a viable sector that leads to the pathway to a science, mathematics, and engineering degree.

REFERENCES Astin, A. W. (1993). What matters in college? Four critical years revisited. San Francisco: Jossey-Bass, Inc. Bailey, R. C. (1971). Self-concept differences in low and high achieving students. Journal of Clinical Psychology, 27(2), 188-191. Bentler, P. M. (1990). Comparative fit indices in structural models. Psychological Bulletin, 107, 238-246. Bryne, B. M. (1984). The general/academic nomological network: A review of construct validation research. Review of Educational Research, 54(3), 427-456. Carlone, H. B. (2003). (Re)producing good science students: Girls’ participation in high school physics. Journal of Women and Minorities in Science and Engineering, 9(1), 17-34. Cohen, A. M., & Brawer, F. B. (2003). The American community college (4th ed.). San Francisco: Jossey-Bass. Drake, K., Clewell, B. C., & Sevo, R. (2002). Meeting the challenge: The impact of the National Science Foundation’s program for women and girls. Journal of Women and Minorities in Science and Engineering, 8(3&4), 285-303. Ethington, C. A., & Wolfle, L. M. (1984). Sex differences in a causal model of mathematics achievement. Journal for Research in Mathematics Education, 15, 361-377. Ethington, C. A, & Wolfle, L. M. (1986). A structural model of mathematics achievement for men and women. American Educational Research Journal, 23(1), 65-75. Fennema, E., & Sherman, J. (1977). Sex-related differences in mathematics achievement, spatial visualization and affective factors. American Educational Research Journal, 14, 51-71. Hansford, B. C., & Hattie, J. A. (1982). The relationship between self and achievement/performance measures. Review of Educational Research, 52(1), 123-142.




227

House, D. J. (1995). The predictive relationship between academic self-concept, achievement expectancies, and grade performance in college calculus. Journal of Social Psychology, 135(1), 111-112. Huang, P. M., & Brainard, S. G. (2001). Identifying determinants of academic self-confidence among science, math, engineering, and technology students. Journal of Women and Minorities in Science and Engineering, 7(4), 315-337. Hughes, W. J. (2000). Perceived gender interaction and course confidence among undergraduate science, mathematics, and technology majors. Journal of Women and Minorities in Science and Engineering, 6(2), 155-167. Jöreskog, K. G., & Sörbom, D. (2002). LISREL 8.54. Lincolnwood, IL: Scientific Software International Inc. Laanan, F. S. (2003). Degree aspirations of two-year college students. Community College Journal of Research and Practice, 27(6), 495-518. Lips, H. M. (1995). Predicting university women’s participation in mathematics and science: A causal model. Journal of Women and Minorities in Science and Engineering, 2(4), 193–206. National Center for Education Statistics. (2002). Digest of Education Statistics: 2001. Washington, DC: U.S. Department of Education. National Science Foundation. (2000). Women, minorities, and persons with disabilities in science and engineering: 2000 (Report No. NSF 00-327). Arlington, VA: Author. National Science Foundation Authorization Act of 2002, Pub. L. 107-368, §3, 116 Stat. 3035 (2002). Nora, A., & Horvath, F. (1990). Structural pattern differences in course enrollment rates among community college students. Research in Higher Education, 31(6), 539-554. Nora, A., & Rendón, L. (1990). Differences in mathematics and science preparation and participation among community college minority and non-minority students. Community College Review, 18(2), 29-40. Pallas, A. M., & Alexander, K.L. (1983). Sex differences in quantitative SAT performance: New evidence on the differential coursework hypothesis. American Educational Research Journal, 20, 165-182. Pascarella, E. T., Smart, J. C., Ethington, C. A., and Nettles, M. T. (1987). The influence of college on self-concept: A consideration of race and gender differences. American Educational Research Journal, 24(1), 49-77. Phillippe, K., & Patton, M. (2000). National profile of community colleges: Trends and statistics (3rd ed.). Washington, DC: Community College Press. Sax, L. J. (1994a). Mathematical self-concept: How college reinforces the gender gap. Research in Higher Education, 35(2), 141-166. Sax, L. J. (1994b). Predicting gender and major field differences in mathematical self-concept during college. Journal of Women and Minorities in Science and Engineering, 1(4), 291-307. Sax, L. J., Astin, A. W., Korn, W. S., & Mahoney, K. (1996). The American freshman: National norms for fall 1996. Los Angeles: Higher Education Research Institute, University of California, Los Angeles. Schumacker, R. E., & Lomax, R. G. (1996). A beginner’s guide to structural equation modeling. Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Schumacker, R. E., & Lomax, R. G. (2004). A beginner’s guide to structural equation modeling (2nd ed.). Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Sherman, J. (1980). Mathematics, spatial visualization, and related factors: Changes in girls and boys, grades 8-11. Journal of Educational Psychology, 72, 476-482. Sherman, J. (1982). Continuing in mathematics: A longitudinal study of the attitudes of high school girls. Psychology of Women Quarterly, 7(2), 132-140. Sherman, J. (1983). Factors predicting girls’ and boys’ enrollment in college preparatory mathematics. Psychology of Women Quarterly, 7(3), 272-281. Spears, J. D., Dyer, R. A., Franks, S. E., & Montelone, B. A. (2004). Building a network to support girls and women in science, technology, engineering, and mathematics. Journal of Women and Minorities in Science and Engineering, 10(2), 161-177.

Volume 11, Issue 3


228


Steiger, J. H., & Lind, C. (1980, May). Statistically based tests for the number of common factors. Paper presented at the annual meeting of the Psychometric Society, Iowa City, IA. Tsapogas, J. (2004a). The role of community colleges in the education of recent science and engineering graduates. InfoBrief NSF 04-315. Washington, DC: National Science Foundation, Directorate for Social, Behavioral, and Economic Sciences. Tsapogas, J. (2004b). More than one-fifth of all individuals employed in science and engineering occupations have less than a bachelor’s degree education. InfoBrief NSF 04-333. Washington, DC: National Science Foundation, Directorate for Social, Behavioral, and Economic Sciences. Wyer, M. (2003). Intending to stay: Images of scientists, attitudes toward women, and gender as influences on persistence among science and engineering majors. Journal of Women and Minorities in Science and Engineering, 9(1), 1-16.




229

APPENDIX Coding and Scaling of the Variables in the CIRP Survey Variables

Coding/Scale

Parents’ annual income

14-point scale: 1 = less than $6,000; 14 = $200,000 or more

Father’s educational attainment Mother’s educational attainment

8-point scale: 1 = grammar school or less; 8 = graduate degree

Years of high school study in mathematics Years of high school study in physical science Years of high school study in biological science Average high school grade

7-point scale: 1 = none; 7 = five or more 8-point scale: 1 = D; 8 = A or A+

Scientifically oriented (α = 0.67) Become authority in my own field Obtain recognition from colleagues Make theoretical contribution to science

4-point scale: 1 = not important; 4 = essential

Materialism/status oriented (α = 0.58) Have administrative responsibility Be very well off financially Be successful in own business

4-point scale: 1 = not important; 4 = essential

Academic self-concept (α = 0.64) Academic ability Mathematical ability Self-confidence (intellectual) Writing ability

5-point scale: 1 = lowest; 5 = highest

Social self-concept (α = 0.68) Drive to achieve Leadership ability Public speaking ability Self-confidence (social) To be elected to an academic honor society

5-point scale: 1 = lowest; 5 = highest

Volume 11, Issue 3

